Silvaco TCAD Athena User Manual

Athena
User’s Manual
Silvaco, Inc.
4701 Patrick Henry Drive, Bldg. 2 May 15, 2014
Santa Clara, CA 95054
Phone: (408) 567-1000
Web: www.silvaco.com
Notice
The information contained in this document is subject to change without notice.

Silvaco, Inc. MAKES NO WARRANTY OF ANY KIND WITH REGARD TO THIS
MATERIAL, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTY
OF FITNESS FOR A PARTICULAR PURPOSE.
Silvaco, Inc. shall not be held liable for errors contained herein or for incidental or
consequential damages in connection with the furnishing, performance, or use of this
material.
This document contains proprietary information, which is protected by copyright laws of the
United States. All rights are reserved. No part of this document may be photocopied,
reproduced, or translated into another language without the prior written consent of Silvaco
Inc.
AccuCell, AccuCore, Athena, Athena 1D, Atlas, Blaze, C-Interpreter, Catalyst AD, Catalyst
DA, Clarity RLC, Clever, Clever Interconnect, Custom IC CAD, DeckBuild, DevEdit,
DevEdit 3D, Device 3D, DRC Assist, Elite, Exact, Expert, Expert C++, Expert 200,
ExpertViews, Ferro, Gateway, Gateway 200, Giga, Giga 3D, Guardian, Guardian DRC,
Guardian LVS, Guardian NET, Harmony, Hipex, Hipex C, Hipex NET, Hipex RC,
HyperFault, Interconnect Modeling, IWorkBench, Laser, LED, LED 3D, Lisa, Luminous,
Luminous 3D, Magnetic, Magnetic 3D, MaskViews, MC Etch & Depo, MC Device, MC
Implant, Mercury, MixedMode, MixedMode XL, MultiCore, Noise, OLED, Optolith,
Organic Display, Organic Solar, OTFT, Quantum, Quantum 3D, Quest, RealTime DRC, REM
2D, REM 3D, SEdit, SMovie, S-Pisces, SSuprem 3, SSuprem 4, SDDL, SFLM, SIPC, SiC,
Silvaco, Silvaco Management Console, SMAN, Silvaco Relational Database, Silos,
Simulation Standard, SmartSpice, SmartSpice 200, SmartSpice API, SmartSpice Debugger,
SmartSpice Embedded, SmartSpice Interpreter, SmartSpice Optimizer, SmartSpice RadHard,
SmartSpice Reliability, SmartSpice Rubberband, SmartSpice RF, SmartView, SolverLib,
Spayn, SpiceServer, Spider, Stellar, TCAD Driven CAD, TCAD Omni, TCAD Omni Utility,
TCAD & EDA Omni Utility, TFT, TFT 3D, Thermal 3D, TonyPlot, TonyPlot 3D, TurboLint,
Universal Token, Universal Utility Token, Utmost III, Utmost III Bipolar, Utmost III Diode,
Utmost III GaAs, Utmost III HBT, Utmost III JFET, Utmost III MOS, Utmost III MultiCore,
Utmost III SOI, Utmost III TFT, Utmost III VBIC, Utmost IV, Utmost IV Acquisition
Module, Utmost IV Model Check Module, Utmost IV Optimization Module, Utmost IV
Script Module, VCSEL, Verilog-A, Victory, Victory Cell, Victory Device, Victory Device
Single Event Effects, Victory Process, Victory Process Advanced Diffusion & Oxidation,
Victory Process Monte Carlo Implant, Victory Process Physical Etch & Deposit, Victory
Stress, Virtual Wafer Fab, VWF, VWF Automation Tools, VWF Interactive Tools, and Vyper
are trademarks of Silvaco, Inc.
All other trademarks mentioned in this manual are the property of their respective owners.
Copyright © 1984 - 2014, Silvaco, Inc.
2 Athena User’s Manual

How to Read this Manual
Style Conventions
Font Style/Convention Description Example
• This represents a list of items or • Bullet A

terms. • Bullet B
• Bullet C
1. This represents a set of directions To open a door:
2. to perform an action. 1. Unlock the door by inserting
3. the key into keyhole.
2. Turn key counter-clockwise.
3. Pull out the key from the
keyhole.
4. Grab the doorknob and turn
clockwise and pull.
 This represents a sequence of FileOpen
menu options and GUI buttons to
perform an action.
Courier This represents the commands, HAPPY BIRTHDAY
parameters, and variables syntax.
Times Roman Bold This represents the menu options File
and buttons in the GUI.
New Century Schoolbook This represents the variables of x+y=1
Italics equations.
This represents the additional
Note: important information. Note: Make sure you save often when
working on a manual.

Table of Contents
Chapter 1 
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.1 Athena Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.1.1 Using This Manual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.1.2 Technical Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.2 Athena Features and Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.2.1 Using Athena With Other Silvaco Software . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.2.2 The Value Of Physically-Based Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Chapter 2 
Tutorial. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1 Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.1.1 Running Athena Under DeckBuild . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.1.2 Loading And Running Athena Standard Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2 Operation Modes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.1 Interactive Mode With DeckBuild . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.2 Batch Mode With DeckBuild. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2.3 No Windows Batch Mode With DeckBuild . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.4 Running Athena inside DeckBuild . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.2.5 Running Athena In Standalone Mode (Without DeckBuild) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2.6 Running Athena 1D (One-Dimensional Version of Athena) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Creating a Device Structure Using Athena. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.1 Procedure Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.2 Athena Input/Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.3 Creating An Initial Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.4 Choosing Models In SSuprem4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.4.1 Implantation, Oxidation, RTA, Diffusion and Epitaxy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.4.2 The Reason for Multiple Models for Each Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.4.3 Choosing an Appropriate Model Using the Method Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
2.4.4 Changing the Method Statement During the Process Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.4.5 Modelling the Correct Substrate Depth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.4.6 Simulating Rapid Thermal Anneals (RTA) Notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.4.7 Simulating Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
2.4.8 Simulating the Epitaxy Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
2.5 Calibrating Athena for a Typical MOSFET Flow. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
2.5.1 Input Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
2.5.2 Tuning Oxidation Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
2.5.3 Tuning Implantation Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
2.5.4 Tuning Diffusion Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
2.5.5 Related Issues on using the Device Simulator Atlas for MOS Process Tuning . . . . . . . . . . . . . . . . . . . 72
2.6 Calibrating Athena for a Typical Bipolar Process Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
2.6.1 Tuning Base and Collector Currents – All Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.6.2 Tuning the Base Current – All Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
2.6.3 Tuning the Collector Current – All Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
2.6.4 The Base Current Profile – Medium Injection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

Table of Contents
2.6.5 The Base Current Profile – Low Injection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

2.7 Using Athena for Simulating SiGe Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.7.1 METHOD Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.7.2 MATERIAL Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.7.3 DEPOSIT Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
2.7.4 DIFFUSE Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
2.8 Using Advanced Features of Athena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
2.8.1 Structure Manipulation Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
2.8.2 Deposition and Wet/Dry Etching using the Physical Models in Athena/Elite . . . . . . . . . . . . . . . . . . . . . 85
2.8.3 MaskViews Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
2.9 Using Athena/Optolith . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
2.9.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
2.9.2 Creating A Mask. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
2.9.3 Illumination System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
2.9.4 The Projection System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
2.9.5 Imaging Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
2.9.6 Defining Material Properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
2.9.7 Structure Exposure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
2.10 Adaptive Meshing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
2.10.1 Introduction to Mesh Adaption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
2.10.2 Interface Mesh Control. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Chapter 3 
SSuprem4 Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
3.1 Diffusion Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
3.1.1 Mathematical Description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
3.1.2 The Fermi Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
3.1.3 Impurity Segregation Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
3.1.4 The Two Dimensional Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
3.1.5 The Fully Coupled Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
3.1.6 Electrical Deactivation and Clustering Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
3.1.7 Grain-based Polysilicon Diffusion Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
3.2 Advanced Diffusion Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
3.2.1 Classical Model of Dopant Diffusion (CDD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
3.2.2 Solid Solubility Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
3.2.3 Interstitials Clusters Model (IC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
3.2.4 Vacancy Cluster Model (VC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
3.2.5 Electrical Deactivation and Clustering Models (DDC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
3.2.6 Typical Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
3.3 Oxidation Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
3.3.1 Numerical Oxidation Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
3.3.2 Compress Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
3.3.3 Viscous Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
3.3.4 Linear Rate Constant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
3.3.5 Parabolic Rate Constant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
3.3.6 Mixed Ambient Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
3.3.7 Analytical Oxidation Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
3.3.8 Recommendations for Successful Oxidation Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
3.4 Silicidation Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193

Table of Contents
3.5 Ion Implantation Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

3.5.1 Analytic Implant Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195
3.5.2 Multi-Layer Implants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
3.5.3 Creating Two-Dimensional Implant Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
3.5.4 Monte Carlo Implants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
3.5.5 Ion Implantation Damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
3.5.6 Stopping Powers in Amorphous Materials and Range Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
3.6 Deposition Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
3.6.1 Deposition of Doped Layers. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
3.6.2 Grid Control During Deposit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
3.6.3 Epitaxy Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
3.7 Etching Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
3.8 Compound Semiconductor Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
3.8.1 Diffusion Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
3.8.2 Implantation Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
3.9 SiGe/SiGeC Simulation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
3.9.1 Deposition of SiGe/SiGeC Epitaxial Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
3.9.2 Boron Diffusion in SiGe/SiGeC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
3.9.3 Boron Transient Diffusion Suppression by Carbon Incorporation Models . . . . . . . . . . . . . . . . . . . . . . 229
3.10 Stress Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
3.10.1 Dopant-Induced Intrinsic Stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
3.10.2 Stress History. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
3.10.3 Stress-Induced Diffusion Enhancement/Retardation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Chapter 4 
Elite Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
4.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
4.2 String Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
4.3 Deposition Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
4.3.1 Conformal Deposition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
4.3.2 CVD Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
4.3.3 Unidirectional Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
4.3.4 Dual Directional Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
4.3.5 Hemispheric Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
4.3.6 Planetary Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
4.3.7 Conical Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
4.3.8 Monte Carlo Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
4.3.9 Custom Deposition Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
4.4 Etch Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
4.4.1 Isotropic Etch Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
4.4.2 RIE Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
4.4.3 Dopant Enhanced Etching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
4.4.4 Plasma Etch Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
4.4.5 Monte Carlo Etching Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
4.5 Reflow Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
4.6 Chemical Mechanical Polish (CMP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
4.6.1 Hard Polish Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
4.6.2 Soft Polish Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258

Table of Contents
Chapter 5 
Optolith Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 260
5.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
5.2 The Imaging Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
5.3 Optical System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
5.3.1 Discretization Errors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
5.3.2 Mesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
5.3.3 Computation Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
5.4 The Exposure Module. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
5.5 Photoresist Bake Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
5.6 The Development Module . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
5.6.1 Dill’s Development Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
5.6.2 Kim’s Development Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
5.6.3 Mack’s Development Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
5.6.4 Trefonas’ Development Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
5.6.5 Hirai’s Development Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
5.7 Proximity Printing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
5.7.1 General Description of Proximity Lithography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
5.7.2 Theory of Proximity Printing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278
5.7.3 Simulation Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Chapter 6 
Statements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280
6.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
6.1.1 Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281
6.1.2 Continuation Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
6.1.3 Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
6.1.4 General Syntax Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
6.1.5 Command Line Parsing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
6.2 Athena Statements List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
6.2.1 Structure and Grid Initialization Statements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
6.2.2 Structure and Mesh Manipulation Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
6.2.3 Simulation Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
6.2.4 Model Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 286
6.2.5 Special DECKBUILD Statements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
6.2.6 Post-processing Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
6.2.7 Execution Control Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
6.2.8 Obsolete Statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
6.2.9 Standard and User-Defined Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
6.2.10 Standard Impurities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
6.3 ABERRATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
6.4 ADAPT.MESH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
6.5 ADAPT.PAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
6.6 BAKE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
6.7 BASE.MESH. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
6.8 BASE.PAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297
6.9 BOUNDARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
6.10 CLUSTER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
6.11 COMMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300

Table of Contents
6.12 CPULOG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

6.13 DEPOSIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 302
6.14 DEVELOP. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
6.15 DIFFUSE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
6.16 DISLOC.LOOP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
6.17 ELECTRODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
6.18 EPITAXY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
6.19 ETCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
6.20 EXPOSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
6.21 EXTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
6.22 FOREACH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
6.23 GO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
6.24 HELP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
6.25 ILLUM.FILTER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
6.26 ILLUMINATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
6.27 IMAGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
6.28 IMPLANT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
6.29 IMPURITY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334
6.30 INITIALIZE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
6.31 INTERSTITIAL and VACANCY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 341
6.32 LAYOUT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
6.33 LINE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
6.34 MASK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348
6.35 MATERIAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349
6.36 METHOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
6.37 MOMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
6.38 OPTICAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
6.39 OPTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
6.40 OXIDE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
6.41 POLISH. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
6.42 PRINT.1D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368
6.43 PRINTF . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
6.44 PROFILE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
6.45 PROJECTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 372
6.46 PUPIL.FILTER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
6.47 QUIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
6.48 RATE.DEPO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
6.49 RATE.DEVELOP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
6.50 RATE.DOPE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379
6.51 RATE.ETCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
6.52 RATE.POLISH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 384
6.53 REGION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
6.54 RELAX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
6.55 SELECT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 388
6.56 SET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 390
6.57 SETMODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
6.58 SILICIDE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392
6.59 SOURCE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 393
6.60 STRESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
6.61 STRETCH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395

Table of Contents
6.62 STRIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397

6.63 STRUCTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
6.64 SYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
6.65 TONYPLOT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
6.66 TRAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 402
6.67 UNSETMODE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403
Appendix A 
C-Interpreter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 404
A.1 C-Interpreter Overview. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
A.2 Example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
Appendix B 
Default Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407
B.1 Oxidation Rate Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
B.1.1 Dry Ambient For <111> Orientation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
B.1.2 Wet Ambient for <111> Orientation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
B.1.3 Orientation Factors For Linear Coefficients (both Ambients). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
B.1.4 Pressure Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
B.1.5 Chlorine Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 409
B.1.6 Doping Dependence Of Oxidation Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
B.1.7 Coefficients for the Analytical Guillemot Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
B.1.8 Numerical Oxidation Coefficients. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
B.1.9 Stress-dependent Growth Model Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 410
B.1.10 Linear Coefficients Of Thermal Expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
B.1.11 Volume Expansion Ratio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
B.2 Impurity Diffusion Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412
B.3 Impurity Segregation Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
B.4 Interface Transport Coefficients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
B.5 Solid Solubility In Silicon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
B.6 Point Defect Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414
B.7 Defect Interface Recombination Parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
B.8 Defect Growth Injection Interface Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
B.9 Material Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
B.9.1 Diffusion Related . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
B.9.2 Stress and Oxidation Related . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 416
Appendix C 
TSUPREM-4 and TSUPREM-3 Compatibility Features . . . . . . . . . . . . . . . . . . . . . . . . 417
C.1 General Syntax Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
C.2 Execution Control Capabilities Provided by DeckBuild . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
C.2.1 DEFINE Statement and Substitutions Capability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
C.2.2 IF/ELSEIF/ELSE/IF.END Capability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
C.2.3 LOOP/L.END/ASSIGN/L.MODIFY Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
C.3 MESH Statement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420
C.4 Using MASK statement with the parameter IN.FILE and XLINES for Automatic 
grid generation in the horizontal direction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
C.5 Using mask information with the EXPOSE MASK=<maskname> statement.. . . . . . . . . . . . . . . . . . . . 422
C.6 Aliases and substitutions for some statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424

Table of Contents
C.7 Changes in the INITIALIZE statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424

C.8 Changes in the DEPOSIT statement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
C.9 Changes in the DIFFUSE statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
C.10 Changes in the ETCH statement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
C.11 Changes in the STRUCUTURE (SAVEFILE) statement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
C.12 Changes in the IMPLANT statement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
C.13 Changes in the ELECTRODE statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
C.14 Changes in the METHOD statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426
C.15 Changes in the MATERIAL statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426

Chapter 1
Introduction
Athena Overview Introduction
1.1 Athena Overview

Athena is a simulator that provides general capabilities for numerical, physically-based, two-
dimensional simulation of semiconductor processing. Athena has a modular architecture that
the following licensable tools and extensions:
• Athena: This tool performs structure initialization and manipulation, and provides basic
deposition and etch facilities
• SSuprem4: This tool is used in the design, analysis, and optimization of silicon
semiconductor structures. It simulates silicon processing steps such as ion implantation,
diffusion and oxidation.
• Elite: This tool is a general purpose 2D topography simulator that accurately describes a
wide range of deposition, etch and reflow processes used in modern IC technologies.
• Optolith: This tool performs general optical lithography simulation including 2D aerial
imaging, non-planar photoresist exposure, and post exposure bake and development.
1.1.1 Using This Manual

This chapter is an overview of Athena. For new users, read Chapter 2 “Tutorial”, especially
the sections that describe the simulator or modules that you have licensed. This chapter will
give you a basic understanding of what Athena can do and how it’s used. The remaining
chapters will give you a detailed understanding of Athena’s capabilities and how to access
them.
Athena is supplied with a number of example problem descriptions. You can access them
through DeckBuild as described in the DeckBuild User’s Manual. These examples
demonstrate the capabilities of Athena. The input files provided as part of these examples can
provide an excellent starting point for developing your own Athena input files.
1.1.2 Technical Support

If you have difficulties or questions relating using Athena, e-mail Silvaco Support at
support@silvaco.com.
When you send us an e-mail message, please:
1. Explain the problem or question in detail.
2. Include any input files that you have created.
3. Provide us with the version number of Athena and the version numbers of the VWF
Interactive Tools that you are using.
4. Include your business telephone number and fax number.
Silvaco support will contact you promptly and resolve your problem as quickly as possible.
User feedback helps further develop Athena. Please send your comments on the programs,
suggestions for improvements, and additional feature requests to support@silvaco.com.

Athena Features and Capabilities Introduction
1.2 Athena Features and Capabilities

Table 1-1 shows the features and capabilities of Athena.
Table 1-1 Athena Features and Capabilities
Features Capabilities
Bake • Time and temperature bake specification.

• Models photoresist material flow.
• Models photo-active compound diffusion.
C-Intepreter • Allows user-defined models for implant damage, Monte Carlo
plasma etching and diffusion in SiGeC.
CMP • Models Chemical Mechanical Polishing.
• Hard and soft models or a combination of both.
• Includes isotropical etch component.
Deposition • Conformal deposition model.
• Hemispherical, planetary, and conical metallization models.
• Unidirectional or dual directional deposition models.
• CVD model.
• Surface diffusion/migration effects.
• Ballistic deposition models including atomistic positioning
effects.
• User-definable models.
• Default deposition machine definitions.
Development • Five different photoresist development models.
Diffusion • Impurity diffusion in general 2D structures including diffusion in
all material layers.
• Fully coupled point defect diffusion model.
• Oxidation enhanced/retarded diffusion effects.
• Rapid thermal annealing.
• Models simultaneous material reflow and impurity diffusion.
• Impurity diffusion in polysilicon accounting for grain and grain
boundary components.
Epitaxy • 2D epitaxy simulation including auto-doping.

Etch • Extensive geometric etch capability.

• Wet etching with isotropic profile advance.
• RIE model that combines isotropic and directional etch
components.
• Microloading effects.
• Angle dependence of etchant source.
• Default etch machine definitions.
• Monte Carlo plasma etching.
• Dopant enhanced etching.
Exposure • Model is based on the Beam Propagation Method simulating
reflections and diffraction effects in non-planar structures with
capability to take into account local modification of material
optical properties the absorbed dose.
• Defocus and large numerical aperture effects.
Imaging • Two dimensional, large numerical aperture, aerial image
formation.
• Up to 9th order imaging system aberrations.
• Extensive source and pupil plane filtering for enhanced aerial
images.
• Full phase shift and transmittance variation mask capabilities.

Implantation • Experimentally verified Pearson and dual Pearson analytical

models.
• Extended low energy and high energy implant parameter tables.
• Binary Collision Approximation Monte Carlo calculations for
crystalline and amorphous materials.
• Universal tilt and rotation capability for both analytic and Monte
Carlo calculations.
Oxidation • Compressible and viscous stress dependent models.
• Separate rate coefficients for silicon and polysilicon materials.
• HCL and pressure-enhanced oxidation models.
• Impurity concentration dependent effects.
• Ability to simulate the oxidation of structures with deep trenches,
undercuts, and ONO layers.
• Accurate models for the simultaneous oxidation and lifting of
polysilicon regions.
Silicidation • Models for titanium, tungsten, cobalt, and platinum silicides.
• Experimentally verified growth rates.
• Reactions and boundary motion on silicide/metal and silicide/
silicon interfaces.
• Accurate material consumption model.
1.2.1 Using Athena With Other Silvaco Software

Athena is normally used in conjunction with the VWF Interactive Tools. These tools include
DeckBuild, TonyPlot, DevEdit, MaskViews and Optimizer. DeckBuild provides an
interactive run time environment. TonyPlot supplies scientific visualization capabilities.
DevEdit is an interactive tool for structure and mesh specification and refinement, and
MaskViews is an IC Layout Editor. The Optimizer supports black box optimization across
multiple simulators. For more information about VWF Interactive Tools, see their respective
user manuals.
Athena is also frequently used in conjunction with the Atlas device simulator. Athena predicts
the physical structures that result from processing. These physical structures are used as input
by Atlas, which then predicts the electrical characteristics associated with specified bias
conditions. Using Athena and Atlas makes it easy to determine the impact of process
parameters on device characteristics.
Athena can also be used as one of the core simulators within Virtual Wafer Fab (VWF). VWF
makes it convenient to perform highly automated simulation-based experimentation. VWF is
used in a way that closely resembles experimental research and development procedures.
Therefore, it links simulation closely to technology development, resulting in greatly
increased benefits from simulation use.
For more information about VWF, see the VWF User’s Manual.

1.2.2 The Value Of Physically-Based Simulation

Physically-based process simulators predict the structures that result from specified process
sequences. This is done by solving systems of equations that describe the physics and
chemistry of semiconductor processes. Detail analysis of various aspects of process
simulation can be found in [1] and [2].
Physically-based simulation provides three major advantages: it is predictive, it provides
insight, and it captures theoretical knowledge in a way that makes this knowledge available to
non-experts.
Physically-based simulation is different from empirical modeling. The goal of empirical
modeling is to obtain analytic formulae that approximate existing data with accuracy and
minimum complexity. Empirical models provide efficient approximation and interpolation.
Empirical models, however, doesn’t provide insight, predictive capabilities, or capture
theoretical knowledge. Physically-based simulation is an alternative to experiments as a
source of data. Empirical modeling can provide compact representations of data from either
source.
Physically-based simulation has become very important for two reasons. One, it’s almost
always much quicker and cheaper than performing experiments. Two, it provides information
that is difficult or impossible to measure.
Physically-based simulation has two drawbacks: you must incorporated are that all the
relevant physics and chemistry into a simulator and numerical procedures, and you must be
implemented to solve the associated equations. But these tasks have been taken care of for
Athena users.
Physically-based process simulation tools users must specify the problem to be simulated.
Athena users specify the problem by defining the following:
• The initial geometry of the structure to be simulated.
• The sequence of process steps (e.g., implantation, etching, diffusion, exposure) that are to
be simulated.
• The physical models to be used.
The subsequent chapters of this manual describe how to perform these steps.

Chapter 2
Tutorial
Getting Started Tutorial
2.1 Getting Started

This chapter is to help you start using Athena by providing a step-by-step tutorial centered on
a typical process simulation sequence. It explains how Athena uses the Interactive tools (i.e.,
DeckBuild, TonyPlot, MaskViews, DevEdit, and Optimizer). These tools make Athena easier
to use and they provide visualization and interface capabilities with other Silvaco tools. This
tutorial assumes that you are familiar with the basic features of the Interactive Tools. For
more information about these tools, see their respective user manuals.
This chapter begins by explaining how to start Athena and continues with tutorials on how to
use the program. We recommend that you read Section 2.3 “Creating a Device Structure
Using Athena” before you move on to the section appropriate to the particular tool you will
be using. This section explains how to start Athena, how to load and run standard examples,
and how to use the Athena online help facility.
The following explanations assume that Athena has been properly installed. See the Silvaco
Installation, Migration and Troubleshooting Guide if you encounter installation difficulties.
2.1.1 Running Athena Under DeckBuild

DeckBuild is an interactive graphic environment that is used for the following purposes:
• Generating input files for process or device simulation or both,
• Running simulations interactively,
• Interfacing between different simulators,
• Invoking other Interactive tools.
To start Athena under DeckBuild in interactive mode, enter the following UNIX command:
deckbuild -an
Note: To start the one-dimensional version of Athena ( Athena 1D) under DeckBuild, use the following command: 
deckbuild -an1.
After a short delay, the Main DeckBuild Window (See Figure 2-1) will appear. The lower text
window of this window will contain the Athena logo and version number, a list of available
modules, and a command prompt. Athena is now ready to run. To become familiar with the
mechanics of running Athena under DeckBuild, load and run some of the Athena standard
examples.
The method described here is the recommended procedure for starting the program. There are
other methods and modes of running Athena. Section 2.2 “Operation Modes” or Section 2.8
“Using Advanced Features of Athena” will describe these methods and modes.

Figure 2-1 Main DeckBuild Window

2.1.2 Loading And Running Athena Standard Examples

DeckBuild makes it possible to load and run a number of example simulation input files. To
access the Athena examples, select Main ControlExamples and the DeckBuild Examples
Window will appear (See Figure 2-2).
Groups of DeckBuild examples are listed in the Section menu and are grouped according to
the simulator or simulation topic the example demonstrates. The Sub-section menu lists
individual example input files. To run examples, select one of the sections (e.g.,
ATHENA_IMPLANT) in the Section menu. This will open a list of input file names. Short
descriptions of the examples will appear in the Examples Window.
Select one of the input files using the Sub-section menu or by double-clicking on the input
file name and a description of the selected input file will appear.
Press the Load Example button to load the selected input file into the DeckBuild Text
Subwindow (bottom panel of the window). The input file, along with other files associated
with the input file (structure files, setfiles for TonyPlot, and layout files for MaskViews), will
be copied into your current directory.
Note: To access the examples in PC, select HelpExamples.
Once the input file is in the DeckBuild Text Subwindow (bottom panel of the window), press
the Run button in the Main DeckBuild window or follow the special instructions in the
DeckBuild Examples Window to run the input file.
Most of the Athena examples contain preset calls to the graphical postprocessing tool
TonyPlot. One or more plots will appear while the selected example is running.
If you are not familiar with DeckBuild, use a simple example to learn the basic DeckBuild
features and capabilities. For more information, see the DeckBuild User’s Manual. This will
assist you in working through the rest of the tutorial.

Figure 2-2 DeckBuild Examples Window

Online Help
You can find information on Athena statements and syntax using the online help facility. You
can access this facility in interactive mode or through DeckBuild. Typing help at the
ATHENA> prompt either in Interactive Mode or in the DeckBuild Text Subwindow will
display a list of valid Athena statements. The syntax for the help command is shown below.
HELP <command name>
This command will give you additional information about parameter names, types, initial
values, and a description of the parameters for the specified command.
To obtain more information on Athena default parameters (stored in a special file called
athenamod), select CommandModels... in the Main DeckBuild Window. This opens
athenamod in a text editing window, making it possible to read the file or copy and paste
statements from the file into a DeckBuild Text Subwindow.
Select CommandNotes... to open a special information file that includes the current
release notes and a release history.
For more information about Athena syntax, statements, parameters, and their default values,
see Chapter 6 “Statements”.

Operation Modes Tutorial
2.2 Operation Modes

Athena is normally used through the DeckBuild run-time environment, which supports both
interactive and batch mode operation. We recommend that you always use Athena through
DeckBuild. In this section, we present the basic information you need to run Athena in the
DeckBuild environment. The DeckBuild User’s Manual provides a more detailed description
of the features and capabilities of DeckBuild.
2.2.1 Interactive Mode With DeckBuild

To start Athena under DeckBuild, type:
deckbuild -an
at the UNIX system command prompt. The command line option -an instructs DeckBuild to
start Athena as the default simulator.
To start with an existing input file, type:
deckbuild -an <input filename>
The run-time output shows the execution of each Athena command and includes error
messages, warnings, extracted parameters and other important output for evaluating each
Athena run. When you run Athena interactively, the run-time output is sent to the DeckBuild
Text Subwindow of the DeckBuild Application Window and you can save it as needed. You
don’t need to save the run-time output explicitly. The following command line, however,
specifies the file name that will be used for storing the run-time output:
deckbuild -an <input filename> -outfile <output filename>
In this case, the run-time output is sent to the output file and to the output section of the
DeckBuild Window.
2.2.2 Batch Mode With DeckBuild

To use DeckBuild in a non-interactive or batch mode, add the -run parameter to the
command that invokes DeckBuild. A prepared command file is required for running in batch
mode. We advise you to save the run-time output to a file because error messages in the run-
time output will be lost when the batch job completes.
deckbuild -run -an <input filename> -outfile <output filename>
Using this command requires a local X-Windows system to be running. The job runs inside a
DeckBuild icon on the terminal and quits automatically when the Athena simulation is
complete. You can also run DeckBuild using a remote display. For example:
deckbuild -run -an <input file> -outfile <output file> -display
<hostname>:0.0

2.2.3 No Windows Batch Mode With DeckBuild

The -ascii parameter is required for completely non-X Windows operation of DeckBuild.
For example:
deckbuild -run -ascii -an <input filename> -outfile <output filename>
This command directs DeckBuild to run the Athena simulation without any display of the
DeckBuild window or icon. This is useful for remote execution without an X windows
emulator or for replacing UNIX-based Athena runs within framework programs.
When using batch mode, use the UNIX command suffix & to detach the job from the current
command shell. To run a remote Athena simulation under DeckBuild without display, and
then logout from the system, use the UNIX nohup command before the following DeckBuild
command line:
nohup deckbuild -run -ascii -an <input filename> -outfile <output
filename> &
2.2.4 Running Athena inside DeckBuild

Each Athena run inside DeckBuild should start with the following command line:
go athena
A single input file can contain several Athena runs each separated with a go athena line.
Input files within DeckBuild can also contain runs from other programs such as Atlas or
DevEdit along with the Athena runs.
Running a given version number of Athena
You can modify the go statement to provide parameters for the Athena run. To run version
5.8.0.R, the syntax is
go athena simflags="-V 5.8.0.R"
Starting Parallel Athena
The -P option is used to set the number of processors to use in a parallel Athena run (only the
MC Implant module is parallelized starting Athena release 5.16.0.R). If the number of
processors specified by -P is greater than the number of processors available, it is
automatically reduced to this cap number. If -P parameter is not specified, Athena will run on
all available processors automatically. To run Athena on 4 processors, use the command:
go athena simflags=" -V 5.16.0.R -P 4"
Running Athena with a user-specified default parameter file
Athena supports the use of multiple default parameter files. These files have the default root
filename athenamod. To start Athena with athenamod.97, the syntax is
go athena simflags="-modfile 97"

2.2.5 Running Athena In Standalone Mode (Without DeckBuild)

You can run Athena outside the DeckBuild environment, but we don’t recommended it. If you
don’t want the overhead of the DeckBuild Window, use the No Windows Mode. Many
important features such as variable substitution, automatic interfacing to device simulation,
and parameter extraction are unavailable outside the DeckBuild environment. To run Athena
directly under UNIX, use the following command line:
athena <input filename>
To save the run-time output to a file, don’t use the UNIX redirect command (>). Instead,
specify the name of the output file:
athena <input filename> -logfile <output filename>
Note: Some of the standard examples supplied with Athena will not run correctly outside of DeckBuild.
2.2.6 Running Athena 1D (One-Dimensional Version of Athena)

If you have license for Athena 1D, you can run it exactly as described above by using
athena1d instead of athena and -an1 parameter instead of -an parameter.

Creating a Device Structure Using Athena Tutorial
2.3 Creating a Device Structure Using Athena

2.3.1 Procedure Overview
Athena is designed as a process simulation framework. The framework includes simulator
independent operations and simulator specific functions that simulate different process steps
(e.g., implant, RIE, or photoresist exposure). This section describes Athena input/output and
the following basic operations for creating an input file:
• Developing a good simulation grid
• Performing conformal deposition
• Performing geometric etches
• Structure manipulation
• Saving and loading structure information
• Interfacing with device simulators
• Using different Interactive tools
These operations are relevant to all individual Athena process simulators. This part of the
tutorial should help you if you’re new to each of the process simulators.
The three sections of the tutorial: SSuprem4, Elite, and OPTOLITH are devoted to individual
simulators and should be read if you’re going to use those simulators.
2.3.2 Athena Input/Output

Before proceeding to the Athena operation, we will discuss how to provide Athena with input
information and the forms of output information available from Athena.
Input Information
The bulk of input information for Athena is usually provided in the form of input files. An
input file is a text file that can be prepared by using DeckBuild (which will be described
throughout the rest of the tutorial) or any ASCII text editor (such as vi on any UNIX system,
or textedit on a SUN system). The individual lines of the text file are called statements. Each
statement consists of a statement name and a set of parameters that specify a certain step of a
process simulation or model coefficients used during subsequent simulation steps. See
Section 6.1 “Overview” for details on statement syntax.
The remainder of this tutorial will introduce you to the task of creating good input files.
Since Athena uses a great deal of default information, much of the default information is
stored in several non-user-specified files. These files are as follows:
• The athenamod file includes default parameters of physical models, diffusion and oxidation
coefficients, default parameters of numerical methods, characteristics of predefined deposition
and etching machines, and optical parameters of materials for lithography simulation.
• The std_table and several svdp**** files in the implant_tables directory contain ion implanta-
tion look-up tables.
• Several files with suffix .mod in the pls and models directories contain parameters for
advanced diffusion models (Section 3.2 “Advanced Diffusion Models”).
• The athenares file includes resistivity vs. doping concentration data.

It’s important to be aware that information from the athenamod file is loaded into Athena
each time it starts. You can override any of the athenamod default parameters by specifying
an alternative parameter in an input file or by specifying the entire models file using the -
modfile option.
Output Information
All run-time output generated by Athena will appear in the DeckBuild Text Subwindow when
running DeckBuild, or in the current window (or specified output file) when running Athena
standalone. Run-time output can be grouped into two categories: Standard Output and
Standard Error Output.
Standard Output consists of the output from the PRINT.1D statements or the EXTRACT
statement of DeckBuild or both, and the normal information messages generated by Athena.
The number of messages generated depends on the output mode chosen in the OPTION
statement. The QUIET mode is the default. Minimum output is generated in this case: all
statements are echoed and the status of a time consuming simulation is reported. The NORMAL
option produces some additional output information, including information about the current
grid (e.g., number of nodes or triangles). VERBOSE and DEBUG modes are useful for
debugging but these options produce too much output for any other purpose.
Standard Error Output consists of the warning and error messages describing syntax errors,
file operation errors, system errors, and internal inconsistencies.
Standard Structure File Format
The main channel of information exchange between Athena and other simulators and tools is
the Standard Structure File (SSF) format. SSF is a universal file format used by a number of
Silvaco simulation programs. The STRUCTURE statement of Athena creates a Standard
Structure File, which contains mesh and solution information, model information, and other
related parameters.
The saved Standard Structure File can be used by the following:
• Athena to re-initialize the structure and continue process simulation.
• Atlas or other device simulators to perform electrical analysis of the structure produced
by Athena.
• TonyPlot to graphically display a solution created by Athena.
• DevEdit to generate an updated mesh and export the mesh and doping back to Athena or
any other simulator.
For more information on structure files, see “Saving a Structure File for Plotting or
Initializing an Athena Input file for Further Processing” on 50.

2.3.3 Creating An Initial Structure

This section will describe how to use DeckBuild’s Commands menu to create a typical
Athena input file. The goal of this section is not to design a real process sequence, but to
demonstrate the use of specific Athena statements and parameters, as well as some DeckBuild
features, to create a realistic input file. You can find many realistic process input files among
the examples and use them as a starting point in your process simulation.
Defining Initial Rectangular Grid
Once DeckBuild is running and the current simulator is set to Athena, open and pin the
Commands menu as shown in Figure 2-3. Then, select Mesh Define.... and the Athena Mesh
Define Menu will appear. We recommend that you pin this popup because it will be used
often in designing an initial mesh.
Figure 2-3 Commands Menu

Now, you can specify the initial rectangular grid. The correct specification of a grid is critical
in process simulation. The number of nodes in the grid Np has a direct influence on simulation
accuracy and time. A finer grid should exist in those areas of the simulation structure where
ion implantation will occur, where p-n junction will be formed, or where optical illumination
will change photoactive component concentration. The number of arithmetic operations
necessary to achieve a solution for processes simulated, using the finite element analysis
method could be estimated as (Np), where  is of order 1.5 - 2.0.
Therefore, to maintain the simulation time within reasonable bounds, the fine grid should not
be allowed to spill over into unnecessary regions. The maximum number of grid nodes is
20,000 for Athena simulations, but most practical simulations use far fewer nodes than this
limit.
To create a simple uniform grid in a rectangular 1 m by 1 m simulation area, click on the
Location field and enter a value of 0.0. Then, click on the Spacing field and enter a value of
0.10. Then, click on the Insert button and the line parameters will appear in the scrolling list.
Note: Athena coordinate system has positive x axis pointed to the right along the structure surface and positive y
axis pointed down to the depth of the structure.
In the same way, set the location of a second X line to 1.0 with a spacing of 0.1. You can
either set the values by dragging a slider or by entering a number directly.
Now, select the Y direction and set the lines with the same values as the X direction. You can
now add the comments at the Comment line. The Athena Mesh Define menu should appear
as shown in Figure 2-4.
Figure 2-4 Athena Mesh Define Menu

You can now write the menu-prepared mesh information into the input file. But first, preview
the rectangular grid by selecting the View... button and the View Grid window (Figure 2-5)
will appear. Notice that vertical and horizontal grid lines are distributed uniformly, and the
121 points and the 200 triangles will be generated.

Figure 2-5 View Grid Window

A uniform grid such as the one shown in Figure 2-5 is inefficient for performing complex
simulations. Therefore, the grid must be improved. First, make a better grid in the y-direction.
Usually, it’s necessary to get better resolution for the depth profile after the ion implantation
step. When adaptive gridding capability isn’t used, apply preliminary knowledge of the
process you are going to simulate.
Suppose you want to perform a 60 keV boron implant so that the implant peak would be
around 0.2 m. It is reasonable to make a finer grid at this depth. To achieve this, simply add
one more Y-line by setting the Location to 0.2 and the Spacing to 0.02. The new
rectangular grid (Figure 2-6) will now appear. Notice the number of points and triangles have
increased to 231 and 400 respectively.
Figure 2-6 New Rectangular Grid

The minimum spacing in the Y-direction is at 0.2 m, and the spacing gradually increases
toward the bottom and the top of the structure. Since the spacing at y=0 is still 0.1, only 3
grid lines lie between 0 and 0.2 m. You may want to make a finer grid at the top of the
structure. To do this, select the top line of the Y Location scrolling list, change the spacing to
0.03, and press the Insert button. The selected line will be replaced by Y LOC=0.00
SPAC=0.03. If you then press the View... button, there will be 8 grid lines between y=0 and
y=0.2 (Figure 2-7).
Figure 2-7 Inserting New Grid Lines

To improve the initial grid in the x-direction, consider two things. First, make sure that a good
2D profile resolution is specified under the mask edges. Second, make sure the vertical grid
lines are placed along future mask edges. To build half of a 0.6 m MOS structure with the
center of the gate at x=0, there must be an additional X line at x=0.3 and spacing at this line
must be small enough to obtain good lateral resolution of source/drain implants. To add these
items, return to the X direction specification in the Mesh Define menu and insert an
additional X line at x=0.3 with spacing = 0.02.
After this final insertion and adding any desired Comment information, the Mesh Define
menu should appear as shown in Figure 2-8. The grid will have 525 points and 960 triangles
(see Figure 2-9).

Figure 2-8 Athena Mesh Define Menu
Figure 2-9 Redefined Grid

Finally, write the Mesh Define information to the file by pressing the Write button. A set of
lines like these will appear:
GO ATHENA
# NON-UNIFORM GRID
LINE X LOC=0.00 SPAC=0.1
LINE X LOC=1 SPAC=0.1
LINE Y LOC=0.00 SPAC=0.03
LINE Y LOC=0.2 SPAC=0.02
LINE Y LOC=1 SPAC=0.1

The first line (GO ATHENA) is called an autointerface statement and tells DeckBuild that
the following file should be run by Athena.
Defining the Initial Substrate
The LINE statements specified by the Mesh Define menu set only the rectangular base for the
Athena simulation structure. The next step is the initialization of the substrate region with its
points, nodes, triangles, background doping, substrate orientation, and some additional
parameters. To initialize the simulation structure, select ATHENA Command MenuMesh
Initialize... and the Mesh Initialize Menu will appear (see Figure 2-10) .
Figure 2-10 Mesh Initialize Menu

Background doping can be set by clicking on the desired impurity box (e.g., Boron). The
background impurity concentration specification will then become active. If the None box is
checked, the concentration information will become inactive and will appear grayed out from
the rest of the menu. Select the desired concentration using the slider (e.g., 3.0) and select an
exponent from the Exp: menu (e.g., 14). This will give a background concentration of 3.0e14
atom/cm3. You can set background concentration using the By Resistivity specification in
Ohm•cm. For this tutorial, check the 2D box in the Dimensionality field. This will run the
simulation in a two-dimensional calculation. The other items in this menu will be discussed in
Section 2.8 “Using Advanced Features of Athena”.

Note: Two-dimensional mode is used in this tutorial to demonstrate 2D grid generation and manipulation. In most
cases, however, it is unnecessary to change the Auto default in the Dimensionality item of the Mesh
Initialize menu. Athena will begin in 1D and will automatically switch to 2D mode at the first statement,
which disrupts the lateral uniformity of the device structure. This generally results in massive savings of
computation time.
You can now write the mesh initialization information into the file by pressing the Write
button. The following two lines will appear in the DeckBuild Text Subwindow:
# INITIAL SILICON STRUCTURE
INIT SILICON C.BORON=3.0E14 ORIENTATION=100 TWO.D
Now, run Athena to obtain the initial structure. Press the Run button on the DeckBuild
control. The following output will appear in the simulator subwindow:
ATHENA> # NON-UNIFORM GRID
ATHENA> LINE X LOC=0.00 SPAC=0.10
ATHENA> LINE X LOC=1 SPAC=0.1
ATHENA> #
ATHENA> LINE Y LOC=0.00 SPAC=0.03
ATHENA> LINE Y LOC=1 SPAC=0.1
ATHENA> # INITIAL SILICON STRUCTURE
ATHENA> INIT SILICON C.BORON=3.0E14 ORIENTATION=100 TWO.D
ATHENA> STRUCT OUTFILE=.history01.str
The line STRUCT OUTFILE=.history01.str is automatically produced by DeckBuild
through the history function. This function provides an important service when debugging
new files, performing “what if” simulations, and visualizing the structure at different steps of
simulation. This feature will be used throughout the tutorial. Use any of the three methods to
visualize the initial structure:
1. Click on the Tools menu button. DeckBuild will then automatically save a temporary
standard structure file and invoke TonyPlot with this file.
2. Click on the Main Control button and the DeckBuild Main Control popup will appear.
Then, click on the Plot Current Structure button. DeckBuild will then automatically
save a temporary standard structure file and invoke TonyPlot with this file.
3. Select (highlight) the name of a structure file (.history01.str in this case) and click on
the Tools or Plot Current Structure. DeckBuild will then start TonyPlot with the
selected structure file.
After a short delay, TonyPlot will appear. It will have only regional and material information.
Click on the Plot menu button and the Display (2D Mesh) popup will appear. Select only the
two left icons: Mesh and Edges and the Initial Triangular Grid (Figure 2-11) will appear in
TonyPlot.

Figure 2-11 Initial Triangular Grid

The grid in Athena consists of points connected to form a number of triangles. Each point has
one or more nodes associated with it. A point within a material region has one node, while a
point which belongs to several regions has several nodes. A node represents the solution (e.g.,
doping concentration) in a particular material region at the point. For example, a given node
may represent solution values in silicon at a point with coordinates (0.0, 0.0); an entirely
different node may represent solution values in oxide at the same point (0.0, 0.0).
So, the previous INIT statement creates the <100> silicon region of 1.0 m x 1.0 m size,
which is uniformly doped with boron concentration of 3e14 atom/cm3. This simulation
structure is ready for any process step (e.g., implant, diffusion, Reactive Ion Etching). Before
discussing the simulation of physical processing using SSuprem4, Elite or OPTOLITH
modules, it’s important to discuss structure manipulation statements that can precede or
alternate with physical process steps.

Simple Film Depositions

Conformal deposition can be used to generate multi-layered structures. Conformal deposition
is the simplest deposit model and can be used in all cases when the exact shape of the
deposited layer is not critical. Conformal deposition can also be used in place of oxidation of
planar or quasi-planar semiconductor regions when doping redistribution during the oxidation
process is negligible.
To set the conformal deposition step, select the menu items ProcessDepositDeposit...
from the Commands menu in DeckBuild and the Athena Deposit Menu (Figure 2-12) will
appear.
Figure 2-12 Athena Deposit Menu
As shown, Conformal Deposition is the default. If it is known that the oxide layer thickness
grown in a process is 200 Angstroms, you can substitute this with conformal oxide
deposition. Select Oxide from the Material menu and set its thickness to 0.02 m. It is always
useful to set several grid layers in a deposited layer. In this case, at least two grid layers are
needed to simulate impurity transport through the oxide layer. In some other cases (e.g.,
photoresist deposition over a non-planar structure), a sufficiently fine grid is needed to
accurately simulate processes within the deposited layer. There are also situations (e.g.,
spacer formation) when several grid layers in a deposited material region are needed to
properly represent the geometrical shape of the region.

The grid in the deposited layer is controlled by Grid Specification parameters in the Athena
Deposit Menu. Set the Total number of grid layers to 2, add a Comment, and click on the
Write button. The following lines will then appear in the DeckBuild Text Subwindow:
# GATE OXIDE DEPOSITION
DEPOSIT OXIDE THICK=0.02 DIVISIONS=2
The next step will be to deposit a phosphorus doped polysilicon layer of 0.5m thickness.
Select the material Polysilicon, and set the thickness to 0.5. To add doping, select the
Impurities box. The Impurity Concentration section will be immediately added to the
Athena Deposit Menu (See Figure 2-13).

Figure 2-13 Impurity Section of the Athena Deposit Menu

Click on the Phosphorus checkbox and set the doping level (e.g., 5.01019) using the slider
and the Exp menu. You can set a non-uniform grid in the deposited layer by changing the
Nominal grid spacing and the Grid spacing location parameters. To create a finer grid at the
polysilicon surface, set the total number of grid layers to 10, the Nominal grid spacing to
0.02 m and the Grid spacing location to 0.0 (at the surface). Then, click on the Write
button and the following deposition statement will be written in the input file as:
DEPOSIT POLY THICK=0.5 C.PHOSPHOR=5.0E19 DIVISIONS=10 \
DY=0.02 YDY=0.0 MIN.SPACING=0.001
Use the Cont button to continue the Athena simulation. This will create the three layer
structure shown in the left plot of Figure 2-14. The MIN.SPACING parameter preserves the
horizontal mesh spacing for high aspect ratio grids. Athena tries to reduce high aspect ratio
grids and MIN.SPACING stops this. To get a finer grid not at the polysilicon surface but in the
middle of polysilicon layer, change YDY to 0.2. This puts on a finer grid at a distance of
0.2m from the surface of the structure. You can do this by positioning the cursor in the input
file and backspacing over existing text, or entering new text. For example:
DEPOSIT POLY THICK=0.5 C.PHOSPHOR=5.0E19 DIVISIONS=10 \
DY=0.02 YDY=0.2
It is possible to see the effect of changing the YDY. parameter within the polysilicon without
rerunning the whole input file. To do this, highlight the previous statement (DEPOSIT
OXIDE...), select Main ControlInit from History button, and press the Cont button. The
new history file can then be loaded into TonyPlot (see the right plot in Figure 2-14).

Figure 2-14 Grid Control for Deposition

Simple Geometrical Etches

The next step in this tutorial process simulation is to define the polysilicon gate definition.
(Implant and thermal steps will be discussed in Section 2.4 “Choosing Models In
SSuprem4”). To set a geometrical etch step, select ProcessEtchEtch... from the
Command menu of DeckBuild. The Athena Etch Menu (Figure 2-15) will appear.
Figure 2-15 Athena Etch Menu
The Geometrical etch is the default method. Other methods will be discussed in Section 2.8.2
“Deposition and Wet/Dry Etching using the Physical Models in Athena/Elite”. Select
Polysilicon from the Material menu. This example will use a polysilicon gate edge at x=0.3
and set the center of the gate at x=0.0 for the initial grid. Therefore, polysilicon should be
etched to the right from x=0.3. To do so, select Right from the Geometrical type, and set
the Etch location to 0.3. This will give the following statement:
# POLY DEFINITION
ETCH POLY RIGHT P1.X=0.3
The structure created by this ETCH statement is shown in the left hand plot of Figure 2-16.
You can obtain an arbitrary shape of geometrical etching by using the Any Shape button.
For example, to make a tilted etch, specify X and Y locations of four Arbitrary points as
shown in Figure 2-17.

The following four etch lines will be inserted into the input file:
# POLY DEFINITION
ETCH POLY START X=0.2 Y=-1
ETCH CONT X=0.4 Y=1
ETCH CONT X=2 Y=1
ETCH DONE X=2 Y=-1
Figure 2-16 Structure Created by Etch Statement

Figure 2-17 Arbitrary Etch
If this input file fragment is runned instead of the previous one (using the INIT statement
from the History capability), the structure after this etch sequence will appear as displayed in
the right hand plot in Figure 2-16. Athena etches all polysilicon material within the specified
polygon. The polygon etch can consist of any number of points. If you use the Insert button,
an additional point will appear after the currently selected point.
An additional option for geometrical etching is a dry etch with a specified thickness. This can
be used for spacer formation as follows: deposit an oxide of a specified thickness (e.g.,
0.2m) and then etch the same thickness again.
# CLEAN GATE OXIDE
ETCH OXIDE DRY THICK=0.03
# SPACER DEPOSITION
# SPACER ETCHING
ETCH OXIDE DRY THICK=0.23
The dry etching step etches the specified material in the region between the top (exposed)
boundary of the structure and a line obtained by translating the boundary line down in the Y
direction. The etch distance is specified by the THICK parameter. Figure 2-19 shows the
resulting spacer.

Reducing Grid Points in Non-Essential Areas using the Relax Parameter

The previous sections demonstrate that the quality of the grid is extremely important for
Athena simulation. The rectangular based grid generated by the INITIALIZE or DEPOSIT
statements may remain intact in those areas not involved in the process steps affecting the
grid (e.g., etching or oxidation). The Grid Relax capability allows the spacing to be increased
in such areas at any point during the simulation. This capability is useful for two reasons.
First, the initial small spacings are propagated throughout the structure. For example, the fine
grid in the X-direction shown in Figure 2-9 may be needed only in the upper portion of the
structure where doping occurs. Eliminating some grid lines and points in the lower portion of
the structure will not affect the accuracy of implant and diffusion simulation. Second, it is
always necessary to set a fine grid in the area where ion implantation takes place, but the fine
grid may be uneccessary after the profile is leveled-off during thermal steps. So, relaxation of
an initially fine grid may save simulation time during subsequent steps. Parameters for the
RELAX statement are set from the Athena Relax Menu (Figure 2-18).
Figure 2-18 Athena Relax Menu
To open this menu, select StructureRelax... in the DeckBuild Commands menu. You
can preform grid relaxation over the whole structure if you select Entire Grid, or within a
selected rectangular area, if you choose Selected and specify Xmin, Xmax, Ymin, and
Ymax in Location Select. Selecting a material from the Material menu specifies which
material region will be affected by the grid relax operation. The default is all materials within
the specified area. You can perform the grid line elimination either in one direction or in both
directions by selecting X direction, Y direction, or Both. The X direction, Relax, cannot
be performed for individual materials except for the substrate.
To understand how the relax function changes a grid, we will use the structure we have
obtained after spacer formation was specified (Figure 2-19). If we relax the entire grid in both
directions (Figure 2-20), the following lines will be inserted into the tutorial input file:
# RELAX EVERYWHERE
RELAX DIR.X=T DIR.Y=T

Figure 2-19 Spacer Formation using Dry Etch

The resultant grid is shown in the upper-right corner of Figure 2-20.
The total number of grid points is reduced from 708 to 388. When comparing with the grid
before relaxation (upper-left corner of Figure 2-20), note that the grid within the oxide spacer
and polygate has not changed. This is due to three factors:
• the relax algorithm works only with rectangular base grid,
• it never eliminates grid lines adjacent to a region boundary,
• the relaxed area should be at least five by five grid points.
Within silicon, each second horizontal line is eliminated. The lower part of each second
vertical line is also eliminated. This happens because the algorithm doesn’t allow the
formation of obtuse triangles.

Figure 2-20 Grids after Various Relax Operations

If you don’t want the grid to be relaxed above y=0.3, choose Selected and set all four
boundaries in the Relax box. This will give the following RELAX statement:
# RELAX LOWER HALF OF THE STRUCTURE
RELAX X.MIN=0.00 X.MAX=1.00 Y.MIN=0.3 Y.MAX=1.00 DIR.X=T DIR.Y=T
In this case, the number of grid points is 567. The grid above y=0.3 remains intact (see the
plot in lower-left corner of Figure 2-20) and the elimination in X and Y- directions happens
only below y=0.3.
To increase spacing only in the X-direction in the area below y=0.3, select the X direction
and leave the Area and Location Selections as before. This will give the following Relax
statement:
# RELAX LOWER HALF ONLY IN X-DIRECTION RELAX X.MIN=0.00 X.MAX=1.00
Y.MIN=0.3 Y.MAX=1.00 DIR.X=T DIR.Y=F

Note: The only difference is that instead of DIR.Y=T, the statement contains DIR.Y=F, which prevents elimination
in Y-direction. This gives 638 grid points and a different pattern of elimination (see the plot in the lower-right
corner of Figure 2-20).
You can also apply several consequent RELAX statements to achieve grid elimination in
different areas of the structure.
An important thing to remember about the RELAX capability is that it allows you to avoid
creating obtuse triangles and avoid relaxing directly on the material boundaries. This
sometimes results in no relaxation or grid relaxation in a subset of the desired area. The most
desirable method for complete control over gridding is by using DevEdit as described briefly
in this chapter and in the DevEdit User’s Manual.
Reflecting a Structure in the “Y” Plane using the Mirror Parameter
This tutorial process simulation has been building one half of a MOSFET-like structure. At
some point in the simulation, you will need to obtain the full structure. This must be done
before exporting the structure to a device simulator or setting electrode names. In general,
structure reflection should be performed when the structure ceases to be symmetrical (e.g., a
tilted implant, an asymmetrical etching, or a deposition takes place), or when a reflecting
boundary condition no longer applies to the side, which is going to be the center of the
structure.
This example will explain how to mirror the structure at its left boundary. To mirror the
structure, select StructureMirror in the Commands menu (Figure 2-21).
Figure 2-21 Athena Mirror Menu

Then, press the Write button to write the following statement to the input file:
STRUCT MIRROR LEFT
The resulting structure is shown in Figure 2-22.

Figure 2-22 Athena Reflect Capability

The left half of the structure is a complete mirror copy of the right part, including node
coordinates, doping values, and so on. Beware of rounding errors when mirroring. If the
boundary of reflection is not smooth to within 0.1 angstroms, some points will be duplicated.
Specification of Electrodes in Athena
The ultimate goal of an Athena simulation is usually to create a device structure (material
layers plus doping), which then can be used by a device simulator (usually Atlas) for
electrical characterization. Although Atlas is able to specify the locations of electrodes, in
many cases specifying electrodes must be done in Athena. For example, it is impossible to
specify an electrode location in Atlas when the electrode does not consist of straight
segments. Also, when specifying electrodes in Athena, it is useful to transfer electrode layer
information from layout to electrical tests in a device simulator (see the description of the
auto-electrode capability in the MaskViews User’s manual).
Athena can attribute an electrode to any metal, silicide, or polysilicon region. A special case
is the backside electrode, which can be placed at the bottom of the structure without having a
metal region there. If you deposit 0.1 m aluminum layer on the full structure after reflection
(Figure 2-22) using:

DEPOSIT ALUMIN THICK=0.1

and etched the following part of the layer between x=-0.8 and x=0.8, using the Any Shape
specification in the Athena Etch Menu (See Figure 2-15):
ETCH ALUMINUM START X=-0.8 Y=-20
ETCH CONT X=-0.8 Y=20
ETCH CONT X=0.8 Y=20
ETCH DONE X=0.8 Y=-20
you will now have the structure shown in Figure 2-23.
Figure 2-23 MOSFET Structure with Electrodes

You can now use the Athena Electrode menu (see Figure 2-24) by selecting
CommandsStructure Electrode.... To set an electrode at a specified position, select the
Specified Position button, type in the X Position (e.g., -0.9) and Name (for example,
source), and press the Write button. The following statement will appear in the input file:
ELECTRODE NAME=SOURCE X=-0.9

Similarly, specify the drain electrode:

ELECTRODE NAME=DRAIN X=0.9
The polysilicon gate electrode specification has the same format. For this structure it can be
done the same way as for source or drain:
ELECTRODE NAME=GATE X=0.0
If the polysilicon layer is not the topmost layer at x=0, the Y Position can be specified. In this
case, check the Y Position checkbox and type in a y coordinate within the polygate layer
(e.g., -0.2). If Y is not specified and the electrode is not on top, Athena will look for the
electrode in the underlying layers. If it fails, an error will be reported. To specify a backside
electrode, select Backside from the Electrode Type field and type in a name (see also
Figure 2-24) .
Figure 2-24 Athena Electrode Menu

The following backside electrode statement will appear in the input file:
ELECTRODE NAME=BACK BACKSIDE
If an electrode name is not specified, DeckBuild issues the error message: NO ELECTRODE
NAME SPECIFIED and the command is not written to the input file.
If an incorrect position for electrode is specified, for example:
ELECTRODE NAME=JUNK X=0.6
Athena will output the following warning message:
Cannot find the electrode for this structure.
Electrode statement ignored and ignores the statement.
Saving a Structure File for Plotting or Initializing an Athena Input file for Further Processing
As mentioned in the “Standard Structure File Format” on 27, the DeckBuild history function
saves structure files after each process step. In many cases, however, you need to save and
initialize structures independently. There are several reason why it’s needed to save and
initialize structures independently.
The first reason is because the stack for the history files is limited (50 by default). The second
reason is because it is usually undesirable to keep dozens of history files on disc (each of
which occupy hundreds of Kbytes) after the DeckBuild session ends. The third reason is
because users often want to save the structure information generated after key process steps
(e.g., final structure).

To save or load a structure, use the Athena File I/O Menu (See Figure 2-25) by selecting
CommandsFile I/O.... Specify a file name (the file extension.str is recommended for all
Athena structure files) and press the Save button. The following line will appear in the input
file:
STRUCT OUTFILE=TUTOR.STR
You can reload this file (tutor.str) back into Athena at any time during the current
DeckBuild session or during any subsequent session. To reload the structure file, press the
Load button on the Athena File I/O Menu. The following INIT statement will appear:
INIT INFILE=TUTOR.STR
Figure 2-25 Athena File I/O Menu
Note: Only the structure will be reloaded if Athena is restarted before this INIT statement. Any parameters or
coefficients that were set during previous simulations must be reset if they are needed. This structure file
can also be used by any device simulator or DevEdit.

Choosing Models In SSuprem4 Tutorial
2.4 Choosing Models In SSuprem4

2.4.1 Implantation, Oxidation, RTA, Diffusion and Epitaxy
This section describes how to simulate process steps (e.g., implantation, diffusion/oxidation,
epitaxy, and silicidation) specific to the SSuprem4 module of Athena. Also discussed, are the
SSuprem4 model statements, such as METHOD, OXIDE, MATERIAL, and IMPURITY.
For more information about SSuprem4, see Chapter 3 “SSuprem4 Models”. For more
information about the MODEL statements, see Chapter 6 “Statements”.
When simulating any process involving dopant or its diffusion or both, it is absolutely critical
for simulation accuracy to use the appropriate model. Process steps where correct choice of
models are vital include implantation, diffusion, rapid thermal annealing, oxidation, and
epitaxy. This section provides specific advice on which models should be used for each
process step.
2.4.2 The Reason for Multiple Models for Each Process

The key to simulating any dopant related process is to accurately account for damage in the
semiconductor. For example, in silicon processing, typical implantation doses can cause
sufficient damage to the substrate to enhance dopant diffusion rates by three orders of
magnitude or more, so choosing the wrong model in this instance will result in inaccurate
results.
Well known device anomalies such as the Reverse Short Channel Effect in MOS processing
or the emitter push effect in bipolar processing are wholly the result of such damage enhanced
diffusion.
Other processes that consume the semiconductor, such as oxidation and silicidation also inject
damage into the substrate. This must be accounted for if accurate dopant profiles are a
requirement.
This section aims to provide you with a set of rules outlined, indicating the correct model that
can be used most of the time without you having a detailed knowledge of the physics
involved. The usual rules of model selection apply here. The more complicated the model, the
greater the simulation time. There is always a compromise between simulation accuracy and
simulation time. The following sections describe when to use the hierarchy of models so that
the most complicated models are only used when you make a significant difference to the
result.

2.4.3 Choosing an Appropriate Model Using the Method Statement

The hierarchy of diffusion and damage models available is broadly related to the maximum
level of damage already in the semiconductor or the maximum level of damage that the next
process step is likely to introduce at any particular time during the process flow. The level of
damage in the semiconductor at any one time is not a static quantity but will depend on when
and how much damage was induced by a process step and how much annealing has occurred
in subsequent thermal steps. The range of models available to you can account for all of the
above effects and allows accurate simulation of dopant diffusion if appropriate models have
been chosen.
The choice of model or combination of models for any of the process steps described above is
defined in the METHOD statement. The METHOD statement serves a number of functions but in
the context of defining damage models the METHOD statement is used for two purposes.
The first purpose is to specify models for how damage is induced during processes such as
implantation or oxidation. The second purpose is to specify how that damage anneals and
diffuses in subsequent or concurrent thermal processes.
It’s important to realize that the METHOD statement must be placed above the line, specifying
the process step or steps to which it refers in the input file. Any number of method statements
can be used in an input file allowing you to change the models at will during the process flow
to optimize the speed and accuracy of the simulation. The models specified in the METHOD
statement will hold true for all processes that follow it until it’s updated by a subsequent
method statement.
Table 2-1 below indicates a recommended method statement for typical processes. It should
be realized that these statements are hierarchical, so there is no accuracy lost if a more
complicated model is used where a simpler one would suffice. The only downside here is a
longer simulation time. The table below starts off with the simplest of models and progresses
to the more complicated ones.
Table 2-1 Recommended Method Statements for Typical Processes

Method Statement Syntax Suitability of using this method syntax
method fermi Use only before undamaged silicon diffusions, where doping
concentrations are less than 1e20/cm3 and no oxidizing
ambient is present.
method two.dim Use before implant doses less than 1e13/cm2 and for
oxidations.
method full.cpl Use before implant doses greater than 1e13/cm2.
cluster.dam high.conc

2.4.4 Changing the Method Statement During the Process Flow

It has previously been stated that the disadvantage of using the most advanced and complex
models is the time involved during diffusion cycle simulation. Accordingly, there is an
incentive during complex process simulations to switch back to a simpler model during a
diffusion cycle when the majority of the damage created by a previous implant has been
annealed. We will show you when to switch to a simpler model.
If the process being modeled has involved implantation or oxidation at any stage, we advise
not to use the fermi model. An exception to this would be in some power devices with very
long diffusion times where the exact nature of surface damage would have little impact on the
final distribution of the dopant and simulation time is at a premium.
In reality, for most small geometry processes, the question of switching models becomes one
of when to add a new method statement that changes from:
METHOD FULL.CPL CLUSTER.DAM HIGH.CONC
to
METHOD TWO.DIM
after a high dose implant.
Switching guidelines
A simple guideline to follow when to switch method statements during a process flow is by
switching back to the TWO.DIM model if the anneal temperature is greater than 900° and the
device has been annealed for at least one minute, following an implant where the dose is
greater than 1e13/cm2.
For a more accurate guideline, see the “RTA Diffusion Modelling” on 145. Table 3-6 shows
the anneal temperature/time combinations required for 95% of the clusters formed during
high dose implants to dissolve. Modeling these dopant/defect clusters requires the fully
coupled (full.cpl) and cluster damage (cluster.dam) models. Only when these clusters
have dissolved can the two.dim model be used without significant loss of simulation
accuracy. As a general rule, we recommend that the method statement be changed to method
two.dim only after a diffusion time that is at least two or three times as long as the values
quoted in the table.
If you wish to be certain of when it’s safe to switch models, the recommended procedure is to
save a structure file at the point of interest, load the file into Tonyplot and perform a 1D
cutline. Plot the clusters and interstitials. If the cluster concentration is still visible, it’s too
early to switch models.
For power devices, where simulation time is at a premium, the same method already
described should be used. But instead of using the cluster concentration as a guide of when to
switch models, the interstitial concentration should be used as the guide as to when to switch
models one more time from the TWO.DIM model to the basic FERMI model. When the
interstitial concentration near the surface during a very long anneal has been reduced to only
marginally above the background level at the anneal temperature concerned, the method
statement can be switched to METHOD FERMI to greatly reduce the simulation time. The
interstitial background level will be the level deep in the substrate where little damage has
occurred.

2.4.5 Modelling the Correct Substrate Depth

An important and often overlooked aspect of the correct modeling of dopant diffusion is the
choice of substrate depth. It has been mentioned previously that the rate of dopant diffusion is
highly dependent on the level of damage in the substrate. Therefore, the accurate modeling of
dopant diffusion requires the accurate modeling of substrate damage, particularly the
movement of interstitials. In general, the interstitials created directly or indirectly by
implantation and oxidation tend to diffuse much greater distances than the dopant. The
substrate depth chosen for modeling purposes must therefore be deep enough to allow the
interstitial concentrations to return to background levels at the bottom of the simulated
substrate, even if no dopant diffusion occurs at this depth.
Figure 2-26 Interstitials can move far into the substrate even after a short 10min anneal
Figure 2-26 shows typical diffusion profiles of interstitials after a 1e15/cm3 20keV Boron
implant at various anneal times. After only a 10 minute anneal, the interstitials have diffused
8um into the substrate.
Interstitials, like dopant, require a concentration gradient in order for overall diffusion to take
place. For example, if the concentration gradient of interstitials is removed by having too
shallow a substrate depth, the concentration of interstitials will start to pile up because they
are no longer being removed through diffusion into the bulk of the substrate. If the level of
modeled interstitials becomes too high, the diffusion of dopant, even near the surface of the
substrate, will also be too high and the simulation will be inaccurate.

Figure 2-27 Effect on boron diffusion profile when too small a substrate depth is used in the simulation
Figure 2-27 shows the boron profiles for two identical anneals, the only difference is the
depth of the simulated substrate. You’ll see that a shallow modeled substrate always results in
more total diffusion, even though the substrate depth was greater than the total diffusion
depth in both cases.
Modeling a deep substrate doesn’t need to involve a huge number of extra mesh points, since
the mesh points can be placed quite far apart near the bottom of the substrate. All that is
required of the mesh points near the bottom of the substrate is that there be sufficient to model
the gradient of interstitials in this region. The number of additional mesh points can be further
reduced in the X-direction by the using several RELAX statements.
For normal small geometry MOSFET/Bipolar processing, a substrate depth of 20 m should
be more than adequate. This depth can be reduced by plotting the vertical interstitial profiles
at various points in the process to find the maximum depth of interstitial diffusion. There is
little to be gained by reducing the depth of simulation, however, if the combination of large
grid spacing is deep in the substrate and the RELAX statement is used appropriately.

Simulating Ion Implantation

Ion implantation is the main method used to introduce doping impurities into semiconductor
device structures. Adequate simulation of the ion implantation process is very important
because modern technologies employ small critical dimensions (CDs) and shallow doping
profiles, high doses, tilted implants and other advanced methods.
The IMPLANT statement can be set by using the Athena Implant Menu (Figure 2-28). To open
this menu, select Process Implant... in the Commands menu.
Figure 2-28 Athena Implant Menu

The following gives the minimum set of parameters that should be specified:
• Name of implant impurity (e.g., boron)
• Implant dose using the slider for the pre-exponential value (e.g., 4.0) and the Exp menu
for the exponent (e.g., 12)
• Implant energy in KeV (e.g., 60)
• Tilt angle in degrees (e.g., 7°)
• Rotation angle in degrees (e.g., 30°)
All other parameters can use their default values.
Press the Write button and the following statement will appear in the input file.
# CHANNEL IMPLANT
IMPLANT BORON DOSE=4.0E12 ENERGY=60 PEARSON TILT=7 ROTATION=30 \
CRYSTAL

All of the parameters in the statement above are self-explanatory except CRYSTALLINE. The
CRYSTALLINE parameter indicates that for all analytical models, the range statistics extracted
for a single silicon crystal will be applied (when available). If AMORPHOUS is selected, the
range parameters measured in pre-amorphized silicon will be used (when available). The
CRYSTALLINE parameter also has another meaning for the Monte Carlo or BCA implant
models. It invokes the Crystalline Material Model which takes channeling into account. Note
that the latter model is much slower (5 - 10 times) than the Amorphous Material Model. The
Crystalline Material Model is the default model for BCA or Monte Carlo simulation.
For a detailed description of ion implant model selection, see Section 3.5 “Ion Implantation
Models”.
You can specify tilt and rotation angles of the ion beam. Positive tilt angles correspond to the
ion beam coming from the top left. Specifying the rotation angle makes sense only for non-
zero tilt angles. Zero rotation means that the ion beam vector lies in the plane parallel to the
2D simulation plane. 90° rotation means that the ion beam vector lies in the plane
perpendicular to the simulation plane.
Selecting Continual rotation causes SSuprem4 to rotate the wafer, i.e., implantation will
be performed at 24 different rotation angles from 0 to 345°, in increments of 15°.
There are several damage models available in SSuprem4. These models allow you to estimate
distributions of various defects generated after ion implantation. For more details about the
damage models and their effect on subsequent diffusion, see Section 3.5.5 “Ion Implantation
Damage”.
When the Monte Carlo model is selected, you can specify several additional optional
parameters (See Figure 2-29). The first three parameters are related to the Damage model
(Point defects, {311}-clusters, and dislocation loops). The three others control Monte Carlo
calculation (initial random number, number of trajectories, and smoothing). See Table 2-2 for
a quick reference of Athena implant models.
Table 2-2 Athena Implant Model Reference

Process Model Assumption Recommendation
Implant SIMS Verified Dual Empirical See Table 3-7.

Pearson (SVDP) - Default
Single Pearson Analytic All other cases
Monte Carlo Statistical Multi-layer structures: angled
Monte or BCA implants into a structure where
many ions could be reflected
(trenches), when channeling is not
described by SVDP, high or very
low energy.

Table 2-2 Athena Implant Model Reference

Silicon Amorphous No channeling effect Most of implant profile is within

Type is included amorphous materials (oxide,
polysilicon, pre-amorphized
silicon); channeling is negligible
or not important.
Crystal - Default Channeling effect is When channeling effects are
included important: light ions (boron,
phosphorus), zero or close to 0°
tilt, implant through thin
amorphous layer into crystalline
substrate.

Figure 2-29 Athena Implant Window

Simulating Diffusion
Simulation of thermal process steps is a focal point of SSuprem4. The hierarchy of diffusion
and oxidation models is described in this chapter and in Sections 3.1 “Diffusion Models” and
3.3 “Oxidation Models”. This section will demonstrate how to set different parameters and
models of diffusion, oxidation and silicidation. The last process will take place only if at least
one refractory metal or silicide layer is present in the structure.
The parameters and models of a diffusion/oxidation step can be prepared from the Athena
Diffuse Menu. (Figure 2-30).
Figure 2-30 Athena Diffuse Menu

To open this menu, select ProcessDiffuse... in the DeckBuild Commands menu. The
Diffuse menu has four sections. Only the Time/Temperature and Ambient fields appear
initially. The Impurities and Models fields appear only when the corresponding check
boxes are selected.
The minimum set of diffusion step parameters is as follows:
• Time (e.g., 60 minutes)
• Temperature (e.g., 1100° Celsius)
• Gas pressure (1 atmosphere is default)
The following input file statements will appear:
# DRIVE-IN
DIFFUSE TIME=60 TEMP=1100 NITRO PRESS=1.00
If you choose the Ramped box and End Temperature or Temperature rate, a ramped
temperature thermal step is simulated. The temperature rate is a variable by default, but it can
be set to a specific constant temperature rate by selecting Constant in the Rate box. If the
End Temperature is set to 1000, the following lines appear:

# RAMPING DOWN
DIFFUSE TIME=60 TEMP=1100 T.FINAL=1000 NITRO PRESS=1.00
The same pull down menu used for inert diffusions is also used for oxidations described in the
“Simulating Oxidation” on 63. But, since there are special considerations for inert diffusions
which come under the category of Rapid Thermal Anneals (RTA), the special notes pertaining
to this specific set of conditions are described in the next section. These notes are very
important for accurate simulation of high temperature, short duration anneals. We recommend
that you read these notes before attempting to write the RTA section of the input file.
2.4.6 Simulating Rapid Thermal Anneals (RTA) Notes

The usual reason for employing a Rapid Thermal Anneal (RTA) in a process flow is to anneal
out damage in the substrate that has been caused by a previous process step, usually an
implant, while at the same time minimizing dopant diffusion. Dopant activation also occurs
during this process. These anneals are usually high in temperature and low in duration for
sound device physics reasons.
Once again, the key to accurate simulation of RTA lies in the accurate simulation of substrate
damage behavior. The role of interstitials in enhanced dopant diffusion has already been
explained in Section 2.4 “Choosing Models In SSuprem4” to become familiar with the role of
interstitials during process simulation.
The reason why an RTA usually employs high temperatures and short durations is because for
a given high dose implant, if an anneal duration is selected so that a fixed percentage of the
damage is annealed, the lower the anneal temperature, the more dopant diffusion occurs.
The above statement requires an explanation since intuitively, the opposite would seem more
likely. A descriptive explanation of what is happening can be informative if the two extremes
of anneal temperature are considered.
For the lowest anneal temperatures, the damage anneal rate is almost zero, so dopant diffusion
rates are enhanced by a factor of 1000C or more for the long time periods required to remove
the damage. This results in high total dopant diffusion.
For the highest temperature anneals, a significant percentage of damage removal occurs in a
fraction of a second. Almost zero damage enhanced diffusion or total diffusion occurs in this
instance, and the anneal time to remove the damage is very short. Extrapolating between
these extremes provides a qualitative explanation of what occurs for intermediate temperature
anneals.
Two important points have now been established:
1. For sound device physics reasons, most RTA processes consist of high temperature, short
duration anneals.
2. Damage-enhanced diffusion will only occur for a few seconds at typical RTA
temperatures.
For accurate simulation of RTA, the second point is most important and often wrongly
neglected. Suppose an RTA consists of a 10 second ramp up to 1000C, followed by a 20
second anneal and a 10 second cool down. From the second point, it is apparent that most of
the Total Dopant Diffusion would have taken place during the Ramp Up Phase of the RTA.
Therefore, always model the temperature ramp up accurately when simulating an RTA
process. In most cases, the ramp down can be neglected, since all the diffusion has already
taken place at the beginning when the silicon was still damaged.

2.4.7 Simulating Oxidation

It has already been stated that the pull down menu for simulating oxidations is the same as
that for simulating inert diffusions described in the “Simulating Diffusion” on 61. See this
section for advice on selecting the appropriate pull down menu from DeckBuild.
The default method for oxidation is Compress. In SSuprem4 examples there are a number of
examples which illustrate the use of different models for different processes and structures.
In our previous example described in the “Simulating Diffusion” on 61, if the next
temperature step is going to be at a constant temperature of 1000°C in dry O2 with 3% of
HCL in the ambient, select the Dry O2 box and set HCL% equal to 3 in the Ambient section
of the Diffuse menu. The following input file fragment will appear:
# GATE OXIDE
DIFFUSE TIME=60 TEMP=1000 DRYO2 PRESS=1.00 HCL.PC=3
If the ambient is a mixture consisting of more than one oxidant, the total oxidation rate will
depend on the combined effect of all species in the ambient. To specify the contents of the
ambient mixture, select t he Gas Flow button in the Ambient section and an additional
Athena Gas Flow Properties Menu (Figure 2-31) will appear.
Figure 2-31 Athena Gas Flow Properties menu

If the Gas Flow components are selected, as shown in Figure 2-31, the following statement
will be generated:
# GATE OXIDE
DIFFUSE TIME=60 TEMP=1000 F.H2O=5.3 F.HCL=0.06 F.O2=8.0 \
PRESS=1.00
One or several impurities can be present in the ambient. To set ambient in the Impurity
Concentration section of the Athena Diffuse Menu (See Figure 2-30), check the
corresponding checkboxes, and set the values using sliders and the Exp menus.
For example, by selecting the appropriate boxes and values, the following DIFFUSE statement
will be inserted into the input file:
# FIELD OXIDE
DIFFUSE TIME=100 TEMP=850 T.FINAL=1060 WETO2 PRESS=1.00 \
HCL.PC=0 C.ARSENIC=9.0E19 C.PHOSPHOR=4.0E20
Several other parameters not included on the menu are available in the DIFFUSE statement
(Section 6.15 “DIFFUSE”). The DUMP, DUMP.PREFIX, and NO.DIFF parameters can be
useful. DUMP and DUMP.PREFIX can be used to make a movie using TonyPlot. The NO.DIFF

parameter specifies that impurity redistribution will be neglected. This provides a good
approximation for low temperature processes, such as silicidation.
Several other model specification statements are important for diffusion processes. These are
as follows:
• IMPURITY, INTERSTITIAL, and other impurity and point defect statements, which
specify model parameters (e.g., diffusivity or segregation) of these species.
• The OXIDE statement, which specifies parameters for different oxidation models.
• The MATERIAL statement, which specifies some basic parameters for all materials.
• The SILICIDE statement, which specifies silicidation coefficients.
Table 2-3 shows the basic diffusion and oxidation models.
Table 2-3 Basic Diffusion and Oxidation Models

Diffuse Fermi- Default Defect in equilibrium For undamaged substrates in inert

ambients
two.dim Transient defect During oxidation, and after medium
diffusion dose implant (e.g., OED)
full.cpl Defect and impurity Post high dose implant & co-diffusion
binding energy model effects, but execution time is high
Oxidation Vertical Planar 1D oxidation only (should never be
used)
Compress- Default Non-planar with linear 2D oxidation (e.g. birds beak)
flow
Viscous Non-planar with non- 2D oxidation (e.g. birds beak with thick
Elastic linear flow Si3N4, however, execution time is higher.
For a detailed description of all diffusion and oxidation models, see Sections 3.1 “Diffusion
Models” and 3.3 “Oxidation Models”.

2.4.8 Simulating the Epitaxy Process

Athena/SSuprem4 can simulate a high temperature silicon epitaxial processes. The epitaxy
process is considered as a combination of deposit and diffuse processes. Therefore, processes
such as autodoping from a highly doped buried layer into a lightly doped epitaxial layer can
be simulated. Diffusion parameters for epitaxial silicon, however, are considered the same as
for single crystal silicon.
The epitaxy process is defined in the Athena Epitaxy Menu (Figure 2-32). To open this menu,
select ProcessEpitaxy in the Commands menu. The Athena Epitaxy Menu consists of
four sections:
• The Time/temperature section selects temperature step parameters in the same way as in
the DIFFUSE statement.
• The Thickness/rate section selects either the total thickness of the epitaxial layer, or the
deposit rate in microns/minute. In the latter case, the total thickness will be determined by
the rate and time.
• The Grid Specification section specifies the vertical grid structure within the grown
epitaxial layer. All grid parameters are equivalent to those of the Athena Deposit Menu
(See Figure 2-12).
• The Impurity Concentrations section specifies the growing epitaxial layer in the same
way as in the DIFFUSE statement.
All parameters in the last two groups are optional. If the parameters of an epitaxial step are set
exactly as shown in Figure 2-32, the following statement will appear in the input file:
# EPI-LAYER
EPITAXY TIME=30 TEMP=900 T.FINAL=1000 THICKNESS=2.0 DIVISIONS=20 \
DY=0.1 YDY=0.00 C.PHOSPHOR=1.4E19
Note: The diffusion during the epitaxy process will use the Diffusion Model Set in the most recent METHOD
statement. If you need another METHOD statement, include it before the EPITAXY statement.

Figure 2-32 Athena Epitaxy Menu

Calibrating Athena for a Typical MOSFET Flow Tutorial
2.5 Calibrating Athena for a Typical MOSFET Flow

This section of the manual provides information on which parameters should be tuned in the
input file to provide predictive simulations using a typical MOSFET process flow. We assume
you are now familiar with the mechanics of making an input file and using the correct
methods and models (see Section 2.4 “Choosing Models In SSuprem4”). For example,
incorrect use of the METHOD statement will invalidate the rest of the following section.
Calibrating an Athena input file for a typical MOS process flow involves using the device
simulator, Atlas, since electrical measurements from the MOSFETs in question often
represents the majority of the physical data available for calibration. This can be thought as a
paradox since Atlas would also have to be correctly calibrated. The reason that this doesn’t
present a problem is discussed below.
An important point to remember when using Technology Computer Aided Design (TCAD) is
that the most critical task is to accurately model the process flow.
Note: For accurate MOSFET simulation, you should invest 90% of the time in achieving an accurate process
simulation, while only investing 10% of the time in fine-tuning the device simulation.
The reason for this, especially for silicon technologies, is that the device physics, in general,
is understood. For silicon, not only is the physics well understood, it is also well
characterized, so most of the default values in Atlas will be correct. Therefore, the calibration
of an Athena process file does not involve the calibration of well known quantities such as
diffusion coefficients. Instead, the calibration involves variables that are process and
production line dependent. For example, the damage caused by an implant cannot be
determined exactly, since it is dose rate dependent and can be influenced by beam heating of
the substrate, which is dependent on the carousel rotation speed and the efficiency of the
cooling system.
Note: If the process has been correctly modeled, the device simulation will also be accurate if appropriate models
have been chosen.
If a simulated device exhibits electrical characteristics that are totally inaccurate, you may
have done something wrong in the process simulation. Do not make the mistake of changing
well known default values in the simulators to make a curve fit one set of results because this
will lead to poor predictive behavior. Try and find the cause of a discrepancy.

2.5.1 Input Information

It may seem obvious but must be emphasized that an accurate process flow is vital for
simulation accuracy, especially for Rapid Thermal Anneals (see Section 2.4.6 “Simulating
Rapid Thermal Anneals (RTA) Notes” for details). Other process information required is an
accurate cross-section of the oxide spacer. Modeling the spacer profile accurately ensures the
lateral damage distribution due to the subsequent source-drain implants is correctly modeled.
Turning to electrical data, the most important device electrical data is a plot of threshold
voltage versus gate length for the NMOS devices. Figure 2-33 shows typical plots of
threshold voltage versus gate length. In this figure, the RTA anneal temperature and times
were varied to show the various profiles that can be expected. A more typical plot is
represented by the 1000C RTA profile, showing a peak value around 1-2 microns with a tail
off for longer or shorter gate lengths.
Figure 2-33 A plot of Threshold Voltage vs. Gate Length for NMOS devices
Gate oxide thickness measurements are also required. Be careful here if oxide thickness is
measured with capacitance-voltage (C-V) methods, since quantum effects in very thin oxides
(less than 5nm) can lead to inaccuracies because the actual location of the peak concentration
of the accumulation charge is not at the interface as classic physics predicts but a short
distance into the silicon. Use the QUANTUM model in Atlas to match accumulation capacitance
with oxide thickness for very thin oxides.
Other useful electrical input information is data that won’t be used now but later for the
calibration process itself, testing the predictive nature of the simulation. Typical device
characteristics used for predictive testing includes threshold voltage versus gate length
measurements for a non-zero substrate bias.

2.5.2 Tuning Oxidation Parameters

During oxidation, interstitials are injected into the silicon substrate by the advancing
interface. The first parameter to tune is the fraction of consumed silicon atoms that are re-
injected back into the substrate as interstitials. In Athena, the related tuning parameter is
called THETA.0 and is defined in the INTERSTITIAL statement. THETA.0 has been found to
be slightly different for wet and dry oxides. The default value is reasonably accurate for dry
oxides but some tuning may be required for wet oxidation.
The major effect of interstitial injection during gate oxidation is to create enhanced diffusion
of the threshold adjust implant. The measured threshold voltage of the final device is very
sensitive to the dopant concentration near the silicon-gate oxide interface. Consequently,
threshold voltage measurements are a sensitive indicator of interstitial behavior. Oxidation,
however, is not the only source of interstitial injection. The source-drain and LDD implants
also induce a large concentration of interstitials. In order to isolate oxidation enhanced
diffusion, the threshold voltage of a long gate length device is used, preferably where L=20
m or more, so that the threshold voltage will be little influenced by damage near the source-
drain regions.
Interstitials injected by source-drain implant damage can travel up to 10 m along the surface
before recombination takes place. A gate length of 20 m is recommended as the minimum
gate length for calibration so this can allow the interstitials to diffuse 10 m along the surface
from both the source and drain ends without effecting diffusion near the center of the device.
In summary, tuning THETA.0 involves the comparison of modeled and measured threshold
voltage data for a long gate-length device.
THETA.0 can be rapidly tuned by taking a one dimensional (1D) vertical cutline through the
center of the gate and doing a 1D process simulation. You can either tune THETA.0 manually
or by using the Optimize function in DeckBuild. Theta.0 is tuned until the measured and
simulated data of the long channel threshold voltage correspond. The fine tuning of THETA.0
is performed by using a full 2D simulation.
Figure 2-34 shows a typical dependence of extracted threshold voltage on the THETA.0 tuning
parameter. Realistic values of THETA.0 correspond to the rising part of the curve. The glitch
in the curve is due to rounding errors in the EXTRACT statement used to calculate the threshold
voltage due to the automatic and independent mesh generated in the EXTRACT statement. The
mesh can be changed from its default value shown here to eliminate this effect. But close
examination reveals that the error is only a few millivolts off, which is accurate enough for
most process parameter extractions.

Figure 2-34 A Typical Dependence of Extracted Threshold Voltage

2.5.3 Tuning Implantation Parameters

You can now tune two implantation parameters by using the threshold voltage versus gate
length data. The peak value of threshold voltage for a given process flow (the reverse short
channel effect) will be a function of the initial implant damage caused by the LDD and
source-drain implants. Since these implants have a high total dose and damage, the tuning
parameter here is the clustering factor. In Athena, this parameter is called CLUST.FACT and is
defined in the CLUSTER statement. The higher the clustering factor, the greater the damage,
and the greater the diffusion, the greater the reverse short channel effect.
Figure 2-35 shows the effect on the threshold voltage of changing the CLUST.FACT parameter
for a typical process flow.
Figure 2-35 How Changing the clust.fact parameter affects the threshold voltage
The second implantation parameter that can now be tuned is the lateral spread of the implant
near the surface. In Athena, this parameter is called LAT.RATIO1 and is defined in the
IMPLANT statement. The lateral spread of the source-drain and LDD dopant is responsible for
the classical short channel effect, where the threshold voltage reduces for very short channel
lengths. Simply tune the LAT.RATIO1 parameter until the onset of classical short channel
effects of simulated and measured data correspond. If the LAT.RATIO1 is increased, the onset
of the classical short channel effect will occur for longer gate lengths.

2.5.4 Tuning Diffusion Parameters

The final part of the threshold voltage versus gate length curve can now be used to tune the
surface recombination rate of interstitials. In Athena, this parameter is called KSURF.0 and is
specified in the INTERSTITIAL statement. The surface recombination of interstitials will
dictate the roll-off rate of threshold voltage from its peak value (reverse short channel effect)
to the long gate length value. Once again, simply tune KSURF.0 until the long channel
threshold voltage roll off rate matches that of the measured data.
PMOS Tuning
PMOS devices are a special case since the boron doped Source/Drain implants overall tend to
absorb interstitials rather than emit them. The reverse short channel effect in buried channel
PMOS devices can be caused by high angle implants. If high angle implants are used, the
reverse short channel effect can be tuned using the LAT.RATIO1 parameter in the IMPLANT
statement.
2.5.5 Related Issues on using the Device Simulator Atlas for MOS Process Tuning
It should now be known that calibrating an Athena process file involves using the device
simulator Atlas to a significant extent. Hence, it’s imperative that the use of the device
simulator doesn’t create additional errors, rendering the process calibration results invalid.
Fortunately, the device physics involved in simulating the conditions required to extract a
threshold voltage are not demanding. The drain voltage required to extract a threshold voltage
is only 50-100mV so effects such as impact ionization can be neglected. The field
perpendicular to the gate is also relatively low around the threshold voltage so field effects in
this direction will do little effect. We recommend, however, using at least the models SRH and
CVT during the calculation. Other parameters for silicon are sufficiently well known for
silicon to the point that the results from the device simulator are reliable.
The first important point is to ensure that you let the device simulator calculate the work
function of the gate electrode from the simulated doping profile rather than assigning a value
to it. This means, making sure that the polysilicon gate is not itself defined as an electrode but
rather a layer of metal, usually aluminum, is deposited on top of the polysilicon gate.
Therefore, this metal layer is the film defined as the electrode. Do not assign a work function
to this deposited metal electrode to ensure that it behaves as an ohmic contact rather than a
Schottky contact. The effective work function of the poly gate will then be correctly
calculated from the doping profile in the polysilicon.
An important area for accuracy in MOSFETs is modeling the inversion region under the gate.
As it is, this charge that is responsible for current conduction in the device. The inversion
region charge under the gate-only extends approximately 30 Angstroms into the silicon. The
inversion region charge density under the gate falls off rapidly with depth into the silicon. It is
imperative that there are several mesh points in the Y direction in this inversion region to
model the drain current correctly. Accordingly, we recommend that the mesh spacing under
the gate be no more than 10 Angstroms (1 nm).
You would think that a 10 Angstrom mesh under the gate would result in a huge number of
mesh points. But, there only needs to be approximately three mesh points within the inversion
region in the Y direction. The grid spacing can increase rapidly in spacing away from the
oxide-silicon interface.

Figure 2-36 shows the effects of changing the mesh spacing at the interface on the simulated
drain current. You can see from this figure that too coarse of a mesh always results in too high
of a current simulated.
Figure 2-36 The effect of changing the mesh spacing at the interface on the simulated drain current
If contact resistance is a problem, then include it in the CONTACT statement. The resistance
added to the CONTACT statement should be the measured resistance per contact divided by the
number of contacts on each individual electrode. Obviously for D.C. measurements, the
resistance on the gate contact will have no effect on the results since no current flows in this
direction.
Checking the Predictive Powers of Tuned Process Parameters
If the process simulation has been correctly tuned, the process and device simulators should
have predictive powers. To check the validity of the tuning process, use a new set of electrical
data that was not used during the tuning process. For example, a good alternative set of data is
to check the threshold voltage versus gate length for a non-zero voltage applied to the
MOSFET body contact.
Conclusion
Using just one set of easily obtained measured electrical data, namely a plot of threshold
voltage versus gate length, you can obtain most of the tuning parameters required for accurate
process simulation. The other most important piece of data required is an accurate
measurement of the gate oxide thickness, which is routinely measured in any instance.
You have been given specific advice as to which process and device models to use for each
process in order to get the best results out of the simulation software. In particular, the correct
use of models for the implantation and diffusion processes is stressed, as this has a dramatic
effect on MOSFET characteristics, especially as anneal times and device dimensions
decrease.

Calibrating Athena for a Typical Bipolar Process Flow Tutorial
2.6 Calibrating Athena for a Typical Bipolar Process Flow

As with MOS calibration text, we assume you are familiar with the mechanics of making an
input file and using the correct methods and models (see Section 2.4 “Choosing Models In
SSuprem4”). For example, incorrect selection of diffusion models defined in the METHOD
statement would invalidate the remainder of the following section.
Calibrating a bipolar process flow entails matching the two parameters, base current and
collector current versus base emitter voltage to measure results throughout the full operating
range of the device. By implication, the current gain of the device (Ic/Ib) will also be
matched. All of the following paragraphs refer to the standard plot of collector and base
currents measured against the base-emitter voltage, Vbe, unless it’s specifically stated
otherwise. This standard IV graph is usually referred to as the Gummel Plot.
Another way of plotting the same information in a different format that can prove useful is a
plot of current gain, hfe, versus the log of the collector current. This graph, however, is a
derivation of the same information that makes it less clear as to which current is increasing or
decreasing for each change. Therefore, a less useful graph when it comes to understanding
exactly what is happening to the collector and base currents.
The full operating range of a bipolar junction transistor (BJT) consists of three general
regions defined by the current density injected into the base. These three operating regions are
usually described as low, medium, and high current injection regimes. The medium injection
region is the most important part of the curve to model correctly as this represents the typical
operating condition of the BJT. Each of the three operating regions is dominated by a
different physical phenomenon. Therefore, successful modeling of a BJT involves matching
both the base and collector currents in each of the three general operating regions, making a
total of six areas for calibration. The derived parameter, hfe, is also a good parameter to
monitor, since this is sensitive to errors in the ratio of collector to base current.
The following text suggests an approach and describes which of the six regions are effected
by each change. The general technique is to calibrate the parameters that have the greatest
effect on device performance in all regions first and then to move on to more subtle
phenomenon that effect certain parts of the base or collector currents or both. In general,
matching the collector current for all injection regions is less problematic than matching the
base current at the extremes of the injection regions. Consequently, there are more sections on
tailoring these parts of the curve. The text is divided into the following sections:
1. Tuning Base and Collector Currents – All Regions
2. Tuning the Base Current – All Regions
3. Tuning the Collector Current – All Regions
4. The Base Current Profile – Medium Injection
5. The Base Current Profile – Low Injection
6. Conclusions
If you follow this order, there should be a reasonable correlation between measured and
simulated data. Most of the tuning parameters, however, have some degree of
interdependency to the extent of which is also device design specific. Therefore, some degree
of iteration of the tuning parameters is to be expected.

When tuning bipolar transistors, there is a greater emphasis to accessing tuning parameters by
using the device simulator, Atlas, compared to optimizing MOSFETs, where most tuning
parameters are process-related. A powerful combination is the tuning of a BiCMOS process
where you can use the MOSFET part of the process flow to tune the process parameters,
while using the Bipolar part of the flow to tune Atlas. This technique should yield a high
degree of predictability in the results.
Tuning the process simulator parameters in Athena are mainly required to model effects, such
as the implantation induced defect enhanced diffusion responsible for the Emitter Push Effect,
which is essential to obtain the correct depth of the base-collector junction. The correct
process modeling of the out diffusion of dopant from the poly-emitter into the mono-
crystalline substrate is also critical to obtaining well-matched IV curves. Another critical
process modeling area is the base implant, because it is essential to match measured and
modeled base resistance for correct modeling of the collector current. These and other issues
are discussed in these sections.
2.6.1 Tuning Base and Collector Currents – All Regions

The most important parameter to model the general level of base and collector currents is the
device measurement temperature. The base and collector currents are strongly influenced by
temperature changes, as small as a few degrees centigrade. A significant effort should be
made to determine the exact temperature of the device during measurements before
calibration is attempted. This temperature should be input into Atlas in the MODELS statement
using the TEMPERATURE=<n> parameter. An increase in temperature will cause an increase in
base and collector currents.
2.6.2 Tuning the Base Current – All Regions

A critical region for poly-emitter bipolar devices is the interface between the poly-emitter and
the mono-crystalline silicon. This region is difficult to process simulate directly as the
interface between the polysilicon emitter and single crystalline silicon usually consists of a
thin, uneven and possibly non-continuous film of oxide. This is simulated by calibrating the
overall effect of this interface with Atlas. The tuning parameter is the surface recombination
velocity at this interface for electrons (VSURFN for PNP devices) or holes (VSURFP for NPN
devices). This will only be effective for thin emitters where at least a fraction of the holes (for
NPN devices) can reach the emitter before recombination.
The surface recombination velocity parameter not only affects the base current, it also affects
the base current in all of the operating regions. Therefore, it is a powerful parameter to
approximately match the base current and gain throughout the full operating range. In some
cases, the base current may be less affected in the very high and very low injection regions by
changes in the surface recombination velocity, and adding some scope to fine tuning the
profile of the base current versus base-emitter voltage curve.
It is important to define the poly-emitter as an electrode so it can define the interfacial surface
recombination velocity, VSURFN and VSURFP, using the CONTACT statement. This is in contrast
to the MOSFET calibration text where we strongly advise you not to define the polygate as an
electrode. Be sure not to get these two confused. The parameter that activates the
recombination velocity is SURF.REC, which is also in the CONTACT statement. For example,
an NPN BJT statement would be:
CONTACT NAME=emitter N.POLYSILICON SURF.REC VSURFP=1.5e5

A lower value of recombination velocity, VSURFP, will reduce the base current and increase
the gain, hfe. The reverse is also true.
2.6.3 Tuning the Collector Current – All Regions

Figure 2-37 shows the parameter that affects the collector current over the entire range is the
intrinsic base resistance. The base resistance is primarily determined by the dose of the base
implant(s). An increase in the base implant dose will decrease the intrinsic base resistance and
decrease the collector current in all injection regions. In some cases, however, the collector
current may be affected a little in the very high injection region, giving scope for fine tuning
the profile of collector current versus base-emitter voltage.
Figure 2-37 Effect of base doping profile on low injection base current in BJT
If the pinched or intrinsic base sheet resistance is a measured parameter, the simplest way to
match measured and simulated data is to make slight changes to the base implant dose so that
the simulated dose is not outside the expected error in actual implanted dose in conjunction
with the error in percentage activation.
In some designs, where the base contact is close to the collector contact or the base contact is
the substrate or is generally wide, the collector current can also influence all current injection
regions by specifying a surface recombination velocity at the base contact. For a typical
design with a buried n+ collector and surface contacts, the surface recombination velocity at
the base contact may have little affect on the collector current.

2.6.4 The Base Current Profile – Medium Injection

In Atlas, there are two major parameters that have a significant affect on the base current in
the medium injection regime. These parameters are the Poly-emitter Work Function and the
Bandgap Narrowing Effect. These parameters are described below.
Poly-emitter work function
If the poly-emitter is described as N.POLYSILICON in the CONTACT statement for an NPN
device, as already described, the Poly-emitter Work Function is then set to 4.17V and is
correct for saturation doped n++ polysilicon. But if the poly-emitter is not saturation-doped,
the work function will differ from this ideal and have a pronounced affect on the base current
and current gain in the medium injection regime as shown in Figure 2-38. The work function
of the poly-gate can vary from 4.17V for n++ poly-silicon to (4.17V + Eg) for p++
polysilicon, depending on the position of the Fermi-Energy. Changing the work function of
the poly-emitter by just 0.1V from 4.17V to 4.27V can often reduce the current gain in half in
the medium injection regime, so it’s very important to assign the correct value. The CONTACT
statement below assigns a work function of 4.27eV to the poly-emitter, while keeping the
other parameters the same as before.
CONTACT NAME=emitter SURF.REC VSURFP=1.5e5 WORKFUN=4.27
The poly-emitter work function can be calculated by measuring the position of the Fermi-
Energy at the poly-silicon/silicon interface relative to the conduction band and adding this
value to 4.17V. For example, if the Fermi-Energy is measured as being 0.1eV from the
conduction band edge, the work function of the poly-emitter set in the CONTACT statement
should be set to 4.17 + 0.1 = 4.27V.
Figure 2-38 Effect of emitter contact work function on bipolar gain

Bandgap Narrowing Effects

If the BIPOLAR parameter is stipulated in the MODELS statement in Atlas, bandgap narrowing
is included automatically. The inclusion of bandgap narrowing in the MODELS statement is
strongly advised since this phenomenon has a significant effect on the current gain of the
device. But, to validate the default Klaassen bandgap narrowing model, you should also use
the Klaassen mobility model. Use the additional keyword KLA in the MODELS statement to
activate this model. For example:
MODELS BIPOLAR KLA
The parameters in the Klaassen bandgap narrowing model are user-definable in the MATERIAL
statement and described in the Physics chapter of the Atlas User’s Manual. There are three
user-definable parameters for the Klaassen band gap narrowing model. The BGN.E parameter
has a linear dependency on doping concentration and has the default value of 6.92e-3 volts.
BGN.C has a square root dependency with doping concentration and has the default value of
0.5. BGN.N is the value of doping where band gap narrowing effectively starts to take effect
and has a default value of 1.3e17/cm3. The equivalent default setting consequently should be
written as:
MATERIAL BGN.E=6.92e-3 BGN.C=0.5 BGN.N=1.3e17
You can alter these parameters to modify the current gain of the device in the medium
injection regime. For example, reducing the linear parameter from 6.92e-3 to 6.5e-3 is
sufficient to cause a significant increase in current gain in the medium injection region.
Although the bandgap narrowing parameters affect both collector and base currents, the base
current is affected to a greater degree. The most sensitive plot to see the effect of small
changes to bandgap narrowing is a plot of current gain versus log of collector current. A
reduction in bandgap narrowing will result in an increase in current gain in the medium
current injection region.
2.6.5 The Base Current Profile – Low Injection

This is one case where there is an interdependency on one parameter, since the intrinsic base
resistance not only affects the collector current in all regions (see the previous section) Figure
2-37, however, also has an effect on the base current in the low injection region.
For a small range of implant doses around the optimum, the base doping concentration will
also affect the position of the knee or the rate or both of fall off of the base current in the low
injection operating region of the device. This is most noticeable as a loss of current gain in the
low injection region for the alternative standard plot of current gain versus collector current.
An increase in the base implant reduces the intrinsic resistance and typically increases the
base current in the low injection region, resulting in a decrease in current gain for very low
currents.
A similar effect to increasing the base doping is observed if the base doping is kept constant
but the overall doping is reduced in the mono-crystalline silicon region of the emitter. You can
tune the doping profile in the mono-crystalline region of the emitter using three parameters in
Athena. The main physical effect of these Athena parameters is to change the doping profile
of the emitter in the mono-crystalline silicon. These process parameters are as follows:
• The total interstitial concentration in the poly-emitter.
• The dopant segregation effects in the poly-emitter.
• The dopant velocity across the silicon/polysilicon boundary.

The first process parameter will affect how quickly the dopant in an implanted poly-emitter
reaches the silicon/polysilicon boundary during the RTA diffusion and therefore affects the
total diffusion of dopant into the single crystalline part of the emitter and the base width
doping profile.
The second process parameter affects dopant pile-up at the poly-silicon/silicon boundary and
therefore the source doping concentration at the mono-crystalline interface. Once again, this
will affect the overall doping profile of the emitter in the mono-crystalline region of the
device.
The third process parameter affects the velocity of transport of dopant across the polysilicon/
silicon boundary with similar effects to the parameters above.
You can use these parameters to tailor the emitter doping profile in the mono-crystalline
silicon region to match available measured data, usually in the form of SIMS or capacitance
information. An accurate profile of dopant in the poly-silicon part of the emitter is not too
important if measured data concerning interfacial dopant concentrations is available. This is
because the work function of the poly-emitter will be set in Atlas by defining the poly-emitter
as an electrode. All you need to calculate the correct work function at the poly-silicon emitter
is the interfacial doping concentration at the poly-silicon/silicon interface on the poly side of
the junction. See the “Poly-emitter work function” on 77 for setting the correct work function
for the poly-emitter .
Conclusions
By using a logical combination of tuning parameters available in both the process simulator
(Athena) and the device simulator (Atlas) and with the influence of each parameter, you can
get a good match for bipolar transistors for most device designs.
Since it is usually less problematic to match the collector current for all levels of applied
base-emitter voltage compared to the matching of base current, you will probably find that
more time is spent trying to match the base current for very small and very large values of
applied base-emitter voltage. You should, however, spend a good amount of time on making
sure that the correct process models are used in the process flow to reduce the overall
uncertainty as to which parameters require calibration.

Using Athena for Simulating SiGe Process Tutorial
2.7 Using Athena for Simulating SiGe Process

The recommended method for simulating SiGe process is to treat germanium as a dopant in
silicon rather than depositing the SiGe material. This is because it allows boron diffusivities
to be germanium concentration dependant by using the model.sige parameter in the METHOD
statement. You can also define SiGe related parameters in the MATERIAL statement.
The example below show a typical set of statements for depositing a 200A thick trapezoid
profile SiGe base region in an HBT may be as follows:
METHOD FULL.CPL MIN.TEMP=600 MODEL.SIGE
MATERIAL SILICON NIFACT.SIGE=100 EAFACT.SIGE=1.5 NO.FLIP
DEPOSIT SILICON THICK=0.02 DIVISIONS=8 C.BORON=5E18 \
C.GERMANIUM=1E22 F.GERMANIUM=1E21
DIFFUSE TIME=4 TEMP=650 NITRO PRESS=1
This section will give a brief explanation of the statements and possible variances. A more
detailed description of the individual parameters are given in the Athena notes files and in
Chapter 6 “Statements”.
2.7.1 METHOD Statement

FULL.CPL is the recommended diffusion model for boron to obtain the best accuracy. It is,
however, slower than other diffusion models due to the larger number and more complex
inter-relations taken into account.
MIN.TEMP is required if the deposition temperature is below 700°C, which is the minimum
calibrated temperature for Athena in standard processing. Select a MIN.TEMP value below the
deposition temperature otherwise, no diffusion will be calculated. The MODEL.SIGE
parameter invokes the silicon germanium models.
2.7.2 MATERIAL Statement

The following statement specifies reasonable user-definable parameters for the SiGe models
in the material silicon.
MATERIAL SILICON NIFACT.SIGE=100 EAFACT.SIGE=1.5 NO.FLIP
The NO.FLIP parameter prevents automatic mesh optimization, which preserves the user-
defined x grid spacing. Athena often will try to remove what it believes are excessive grid
points during DEPOSITION/ETCH statements. It is not advisable to remove mesh points in the
base of an HBT, which is the usual SiGe application.
2.7.3 DEPOSIT Statement

This statement is used to specify the material (silicon), thickness, gridding, and doping
parameters. Once you specify these parameters, specify the germanium dopant concentration
either as a constant or as a graded doping profile as shown in the example.
To calculate the doping required, the total number of atoms in silicon is taken to be 5e22/cm3
for simplicity. Therefore, a doping of 1e22/cm3 of germanium is a 20% concentration. Other
concentrations are calculated in proportion to this, so a final concentration at the end of the
deposition of 1e21 represents a 2% concentration of germanium. Therefore, the example

Using Athena for Simulating SiGe Process Tutorial
deposits a 200A thick SiGe film with an initial germanium concentration of 20% and a final
germanium concentration of 2%.
2.7.4 DIFFUSE Statement

Since this example above used the DEPOSIT statement, the thermal budget for deposition is
simulated by an inert diffusion for the deposition time. A typical deposition temperature is
around 650°C.
Generally, a typical SiGe HBT device would have a base profile consisting of boron and non-
boron doped regions together with a tapered germanium profile at both ends of the base. In
this case, simply specify the DEPOSIT and DIFFUSE statements in several stages. You can also
use several EPITAXY statements to do the same thing.
At the end of the process simulation, the germanium dopant profile has to be converted into
SiGe material with a variable X composition to pass the device into Atlas. To do this, type in
the following STRUCTURE command example.
STRUCT OUTFILE=HBT.STR SIGE.CONV
This will save a structure call HBT.str and converts the germanium dopant profile into the
correct X composition SiGe material.

Using Advanced Features of Athena Tutorial
2.8 Using Advanced Features of Athena

2.8.1 Structure Manipulation Tools
Using the Structure FLIP Capability
The Structure FLIP capability allows you to flip the structure in the x axis. The STRUCT
FLIP.Y statement causes the structure to be vertically flipped.
This operation can be useful if some process steps (e.g., etching, deposition, or implant) take
place from the backside of the wafer. By using this statement, you can flip a structure,
perform these steps, and then flip it back.
Using the Stretch Capability
In some cases, a device characterization as a function of length is of interest. For example, the
drain current characteristics depend strongly on the gate length. The Stretch capability makes
it possible to generate a number of MOSFET structures with different gate lengths from one
Athena simulation. The structure obtained so far in this tutorial (See Figure 2-22) has a gate
length of 0.6. To increase the gate length to 1.5, use the STRETCH command. To use this
capability, select StructureStretch in the Commands menu and the Athena Stretch menu
(Figure 2-39) will appear.
Figure 2-39 Athena Stretch Menu

Then, select Polysilicon from the Stretch Target Material menu. Next, set Stretch
Length to 1.5, and choose 10 as the number of Grid Divisions. Then, press the Write
button and the following command will appear in the input file:
# STRETCH TO 1.5 MICRONS
STRETCH LENGTH=1.5 POLY SNAP DIVISION=10
As a result, the polygate will be stretched from its initial length of 0.6  (left plot in Figure 2-
40) to 1.5µ (right plot in Figure 2-40).
Ten additional vertical grid lines will be inserted in the center of the gate area. The LENGTH
parameter of the STRETCH command can serve as a split parameter for the Virtual Wafer Fab
Split Experimentation capability. For more information about this capability, see the VWF
Automation and Production Tools User’s Manual.

Another use of the Stretch capability is in the simulation of large power device structures,
where active areas are uniform everywhere except in close proximity to the mask edges and
are separated from each other by long non-active or isolation regions. You can simulate a
shrunken structure and then stretch the active or non-active or both areas to the actual widths.
This will also save a tremendous amount of simulation time.
Figure 2-40 Using Stretch Function for a MOSFET Structure
Note: The stretch function can save a great deal of CPU time.

Using Athena In 1D Mode

You can increase the simulation speed by running Athena in 1D mode. Athena automatically
runs in 1D mode by default initially. The simulation will automatically converts to 2D mode
when you perform a two dimensional simulation process, such as ETCH or EXPOSE. Simple
operations such as conformal deposits, oxidation, and diffusion run faster in 1D mode.
The deposition and etch sequences displayed in Figure 2-41 show a sequence of 1D
depositions with an automatic conversion to 2D at the first etch.
Initial Fast 1D
silicon calculation
Epitaxy or Fast 1D
other calculation.
blanket
processing
Gate Automatically
formation transformed to
or other 2D calculation.
2D step
Process 2D calculation
completion
Figure 2-41 Automatic 1D to 2D Conversion

Figure 2-41 shows another aspect of 1D mode. In this case, the INITIALIZE command is
specified with the parameters ONE.D and X.LOC=<n> (see Section 6.30 “INITIALIZE”).
ONE.D specifies that a one dimensional calculation is to be done at the location X.LOC.
In the case of Figure 2-42, 1D profiles are generated at different X locations of a complicated
BiCMOS structure. This allows you to quickly check of the overall process using the 1D
mode.

Build a complete
2D process
N+ +
N P+ P+ description
P N (i.e. BiCMOS).
P+ N+
P - Substrate
Quickly calculate
1D profiles at key
locations using
the 1D mode.
Modify process
description to
produce desired
characteristics
at key locations.
Simulate the
complete 2D
N+ N+ P+ P+ process only
P N
once!
P+ N+
P - Substrate
Figure 2-42 Use of One Dimensional Mode
2.8.2 Deposition and Wet/Dry Etching using the Physical Models in Athena/Elite
This section describes the deposition and etch capabilities of the Elite module of Athena
using Athena/Elite default machines
To use Athena/Elite’s physically based deposition and etch models, use at least one of the
following steps:
1. Use one of the predefined machines described in Section 4.4 “Etch Models”.
2. Within your input file, modify the specification of one of these predefined machines to fit
your process
3. Define a custom machine
4. Invoke a prepared file that defines machines of interest
For example, one of the predefined machines is named PE4450. This machine deposits
aluminum at a rate of 1 micron/minute from a hemispheric source.
To simulate the effects of two minutes of operation of this machine, open the Athena Deposit
Menu and select the Machine checkbox. The section PARAMETERS TO RUN THE
DEFINED MACHINE will appear in the menu (See Figure 2-43). In this section, specify
PE4450 as the Machine Name, the Time of run (2.0), and the time units (menu box beside
the Time of run field) as minutes.

We recommend that you specify the Total number of grid layers in the deposited material
region. If you set this number to 10, it will insert the following Athena/Elite DEPOSIT
statement:
# USING DEFAULT DEPOSIT MACHINE PE4450
DEPOSIT MACHINE=PE4450 TIME=2.0 MINUTES DIVISIONS=10
You can specify impurity concentrations in the deposited region in the Impurity
concentration section of the Athena Deposit Menu by clicking on the Impurities box.
Figure 2-43 Athena Deposit Menu with Machine Section

Modifying Athena/Elite Default Machines

The file athenamod defines PE4450 as follows (Notice that a ( \ ) is used to concatenate or
continue a long input line):
RATE.DEPO MACHINE=PE4450 ALUMINUM \
U.M SIGMA.DEP=.35 HEMISPHE DEP.RATE=1.0 \
ANGLE1=72 ANGLE2=-70
The machine is modeled with a hemispherical deposition model. The deposition rate is 1
micron/minute. The logical parameter, U.M, specifies what units are used, in this case,
microns per minute. Finally, the angles of incidence of the hemispherical deposition with
respect to the surface normal are specified with the ANGLE1 and ANGLE2 parameters. You can
modify these characteristics of the machine PE4450 by copying the specification to the input
file and using an ASCII editor. For example:
RATE.DEPO MACHINE=PE4450 ALUMINUM \
U.M SIGMA.DEP=.35 HEMISPHE DEP.RATE=.5 \
ANGLE1=72 ANGLE2=-70
redefines machine PE4450 to have a deposition rate of 0.5 micron/minute.
Defining Elite Deposition Machines
You can define your own deposition machine using the Athena Rate Deposit menu (Figure 2-
44). To open this menu, select ProcessDepositRate Deposit in the Commands menu.
Machine definition requires the specification of five general parameters and one or several
model-specific parameters. The general parameters that must be specified are the following:
• Machine name (e.g., TEST01): This parameter uniquely identifies the machine.
• Material name (e.g., aluminum): A user-defined material.
• Machine (model) type (e.g., unidirectional): You can select one of six models by
pressing the appropriate button.
• Deposition (rate units specifier e.g., A/min): You can select one of seven unit specifiers
from the menu.
• Deposition rate (e.g., 1000): This parameter is specified in the user-selected units.
The SIGMA.DEP parameter is optional and defaults to 0.2.
The SMOOTH.WIN and SMOOTH.STEP parameters provides an alternative to a complete reflow
calculation. It allows a geometric averaging over a window of width (SMOOTH.WIN) microns
that is performed over a number of steps (SMOOTH.STEP). These parameters perform a post-
deposition smoothing that effectively emulates a reflow process. The wider smoothing
window produces a more intensive surface redistribution of the deposit material. The default
number of smoothing operations (1) is adequate for most applications.
One or several model-specific parameters are attributed to each model. For example, only the
ANGLE1 parameter is required for the unidirectional model. Table 2-4 indicates which
parameters are required for each model. The Machine Type section of the Athena Rate
Deposit Menu includes only those parameters that are relevant to the selected model. Each
parameter has a default value which will be inserted in the input file. If some of the
parameters are undefined, the simulation may be invalid or may produce unpredictable
results.

If the Athena Rate Deposit menu is set as shown in Figure 2-44, the following RATE.DEPO
statement will be inserted into the input file.
RATE.DEPO MACHINE=TEST ALUMINUM U.M SIGMA.DEP=0.2
SMOOTH.WIN=0.1 SMOOTH.STEP=1 DUALDIREC
DEP.RATE=1.0 ANGLE1=45.0 ANGLE2= -45.0
\
Figure 2-44 Athena Rate Deposit Menu

Table 2-4 Deposition Model Required Parameters

Models
Parameters
CVD UNI DUAL HEMI CONIC PLANET MONTE1 MONTE2 CUSTOM1 CUSTOM2
dep.rate yes yes yes yes yes yes yes yes optional yes
step.cov yes no no no no no no no no no
angle1 no yes yes yes yes yes/no* yes yes no no
angle2 no no yes yes no yes no no no no
c.axis no no no no yes yes no no no no
p.axis no no no no yes yes no no no no
dist.pl no no no no no yes/no * no no no no
sigma.dep no optional optional optional optional optional optional optional no yes
smooth.win no optional optional optional optional optional optional optional optional optional
smooth.step no optional optional optional optional optional optional optional optional optional
* To use the planetary model, either the ANGLE1 or the DIST.PL parameter must be specified. These parameters are mutually exclusive.

Defining Elite Etch Machines

An Athena/Elite etch machine can be defined using the Athena Rate Etch Menu (Figure 2-
45).
To open this menu, select ProcessEtchRate Etch in the Commands menu. The
machine definition requires the specification of four general parameters and one or several
model-specific parameters.
The general parameters that must be specified are as follows:
• Machine name (e.g. TEST02): This parameter uniquely identifies the machine.
• Material name (e.g., silicon): A user -defined material can also be specified.
• Machine type (e.g., Wet Etch): You can select one of two models by pressing the
appropriate button.
• Etch rate units specifier (e.g., A/min): You can select one of seven unit specifiers from the
menu.
One or several model-specific parameters are attributed to each model. For example, only the
ISOTROPIC rate parameter is required for the Wet Etch model.
Table 2-5 indicates which parameters are required for each of the three models. The
Parameters for Specific Machine Type section of the Rate Etch menu includes only those
parameters, which are relevant to the selected model.
If the Athena Rate Etch Menu is set as shown in Figure 2-45, the following RATE.ETCH
statement will be inserted into the input file:
# TEST02 ETCHING MACHINE
RATE.ETCH MACHINE=TEST02 SILICON U.M WET.ETCH ISOTROPIC=0.03
Figure 2-45 Athena Rate Etch menu

If several materials are present in the structure to be etched, etch rates for each material type
should be specified in separate RATE.ETCH statements.

Table 2-5 Allowable Etch Model Parameters

Parameters wet.etch rie
isotropic yes yes
directional no yes
divergence no yes
chemical no yes
Using A Specified Etch Machine

When etch rates for a specific machine are specified using RATE.ETCH statements, you can
simulate the effects of the operation of this machine. To simulate the etch process using a
specified etch machine, open the Athena Etch Menu and check the Etching Machine box.
The Parameters to Run the Defined Machine section will appear in the menu (See Figure
2-46).
Figure 2-46 Athena Parameters to Run the Define Machine Etch Menu Section
The Machine Name (TEST02), the time units (e.g., minutes) and the Time of run (e.g., 1.0)
must be specified. There are also two tuning parameters that control time stepping during the
etch process. To improve the smoothness of the etch surface, decrease the maximum time step
parameter, DT.MAX, from its default value of 10 percent of the specified Time of Run value.
If you set the Athena Etch Menu as shown in Figure 2-46, the following ETCH statement will
appear in the input file when you press the WRITE button.
# 1 MINUTE ETCHING USING TEST02 ETCH MACHINE
ETCH MACHINE=TEST02 TIME=1.0 MINUTES DT.MAX=0.25
A new parameter, DX.MULT, will be added to the ETCH statement to allow enhanced
discretization during individual Elite Etch steps. Increasing the value of DX.MULT from its
default value of 1.0 will result in larger surface segments and a reduced discretization.

Decreasing DX.MULT will result in better discretization in both space and time during the etch
calculation. Reducing the value of this parameter allows realistic modeling of wet etches that
previously were poorly resolved. Use the DX.MULT is preferable to the use of DT.MAX.
2.8.3 MaskViews Interface

This section describes an alternative to the manual specification of grid and etch steps
described in Section 2.3 “Creating a Device Structure Using Athena”.
Defining Initial Rectangular Grid Using MaskViews
An initial rectangular grid can also be defined by using Silvaco’s IC layout editor,
MaskViews. MaskViews is designed specifically for interfacing IC layout information with
process and device simulators. For more detailed information about MaskViews and its
interface with DeckBuild, see the MaskViews User’s manual. This section gives several
practical suggestions on how to prepare a good initial grid for Athena.
With MaskViews, you can omit Athena mesh definition statements because you can include
the gridding information in the layout file. When using MaskViews to provide line
information, DeckBuild will comment out existing line commands when it loads the
MaskViews information.
Load the example, 43.11 anex11.in, from the ATHENA-COMPLEX section of the
DeckBuild Examples Window (See Figure 2-2). Then, select MaskViewsStarting
MaskViews from the Tools menu of DeckBuild to open MaskViews Layout Files Popup
(Figure 2-47).
Figure 2-47 MaskViews Layout Files Popup

Choose the anex11.lay layout file from the scrolling list and press the Start MaskViews
button. The MaskViews window will then appear as shown in Figure 2-48.

Figure 2-48 MaskViews Window

This section will describe how to modify a grid specification for Athena. First, set the grid in
the Y direction by selecting GridY... from the Define menu. Figure 2-49 shows the Vertical
Grid Control popup will appear.
You can add, modify, and delete the lines for Athena initial rectangular grid exactly the same
way as using the Athena Mesh Define menu from DeckBuild as previously described. Note
that the Distance parameter is equivalent to the location parameter in Athena. Also, the
Add button is equivalent to the Insert button of the Athena Mesh Define Menu. Then, press
Return after entering the Distance or the Spacing values.
If the Distance and Spacing are set as shown in Figure 2-49, the grid will be the same in the
Y-direction as the grid produced using the Athena Mesh Define Menu.

Figure 2-49 Vertical Grid Control Popup

MaskViews also controls the initial Athena grid in the X-direction. MaskViews generates
Athena line statements for each mask edge on valid layers crossed by a cutline. The grid
spacing and the validation of layers can be set by the Athena Grid Control Menu (See Figure
2-50). To open this menu, select GridX... from the Define menu.
Figure 2-50 Athena Grid Control Menu

Figure 2-50 shows the line locations and spacings preset for the POLY layer. This set of
parameters means that for each POLY edge crossed by a cutline, three line statements are to be
inserted into the Athena input file. The first line will be located exactly at the edge and the
spacing will be 0.05m. The second line will be inside the POLY layer, 0.3m from the edge,
and spacing at this line is 0.15m. The third line will be outside the POLY layer, 0.2m from
the edge, and its spacing will be set to 0.1m. You can choose the current edit layer by
selecting the Name button (located underneath the Layers Menu) for the layer in the key list
of MaskViews (Figure 2-48). If you select AAD, then only set one line for an edge of the AAD
layer because offset distances are equal to 0.0. We recommend that only one line be set for
unimportant layers. It’s also important to validate only those layers that are going to be used
in Athena MASK statements.
Select Write Cutline File from the File menu. Then, the prompt Select first end of the
ATHENA cross section will appear at the bottom panel of the MaskViews window. Press the
select (left) mouse button at the desired point in the layout (e.g., within the VIA2 region in the
upper left corner of the layout). You will be prompted to select another end of the cross
section line. Then, drag the pointer and press the select mouse button on the other end of the
selected cutline. Figure 2-51 shows the Athena Cutline Popup, which will then appear. This
shows the exact location of the cutline. You can now preview the mask and grid information
generated by MaskViews. Press the Preview button and the Display Masks Window will
appear as shown in Figure 2-52.
Figure 2-51 Athena Cutline Popup

Figure 2-52 Display Masks Window

The additional information on the number of lines, points, and triangles is also displayed in
this window. If the grid does not appear as shown in Figure 2-52. Check Display Masks and
the third option (Masks and Grid) in the Defaults menu of the Preferences (Figure 2-53).

Figure 2-53 Preferences Menu

To select another cutline location, press the Done button in the Athena Cutline Popup and
repeat the cutline selection process for the desired cutlines one at a time. If you’re not
satisfied with the grid, you can modify the X... or Y... or both settings. You can then preview
the modified grid without selecting another cutline. For example, if the Spacing at edge in
the Athena Cutline Popup (see Figure 2-51) is decreased from 0.05 to 0.025, a finer grid will
be obtained at both POLY gate edges.
When the location of a cutline and the corresponding grid are satisfactory, the cutline
information can be stored or used either as a Cut file or as a cutline object.
You can save the Cut file by pressing the Write button in the Athena Cutline Popup (see
Figure 2-51). You can then load this file into DeckBuild for use in Athena by selecting
MaskViewsCutfiles... from the Tools menu in the DeckBuild window. The MaskViews
Cut Files popup (Figure 2-54 left) will then appear. Select the desired *.sec* file and press
Load.

Figure 2-54 MaskViews Cutline Files Window

There is another way of selecting and using cutline information in Athena. This involves
select and drag-and-drop operation. To do this, using the Right mouse button, change
Category in the MaskViews Cut Files menu from Disk Files to Drag & Drop (see Figure 2-
54 right). Press the Select mouse button anywhere within the Display Masks window (Figure
2-52) and the cutline icon will appear. Without releasing the Select mouse button, drag the
icon into the Drop & Drag window and drop it by releasing the Select mouse button.
Several cutlines with different locations and grids can be dragged and dropped in this fashion.
You can then load any of them into DeckBuild by clicking on the Load button.
When Athena is loaded with a cutline, DeckBuild will comment out all existing line
statements and will automatically run line statements generated by MaskViews. For example,
the following output will appear in the DeckBuild Text Subwindow if the default.sec
generated for the CMOS Inverter is loaded.
ATHENA> LINE X LOC=0.000 SPAC=0.100 TAG=LEFT


ATHENA> LINE X LOC=2.600 SPAC=0.100 TAG=RIGHT
ATHENA> LINE Y LOC=0.00 SPAC=0.03 TAG=TOP
ATHENA> LINE Y LOC=1.00 SPAC=0.10 TAG=BOTTOM
Using MaskViews for Generating Masks in Athena
The dry etching capability of Athena and the physical etching capability of Athena/Elite can
be used in conjunction with the mask generating capability provided by DeckBuild and
MaskViews. A cutline loaded into DeckBuild has information on the x-location of the
photomask edges. You should specify the sequence of mask creation and stripping steps in the
Athena input file. This can be done by selecting the CommandsProcessPhotoMask
option from DeckBuild when Athena is active, which will open the DeckBuild: Athena
Mask menu. (Figure 2-55).
When you select a name (e.g., POLY) from the Edit layer list, press the Apply Mask button
and the following lines will appear in the input file.
# DEFINING POLY MASK
MASK NAME="POLY"
Figure 2-55 Athena Mask Popup

Other parameters of the DeckBuild: ATHENA Mask menu (Bias, Misalign, Delta CD, and
Shrink) allow you to modify or experiment with mask size or position or both (see Section
4.9 in the DeckBuild User’s Manual)
During runtime, DeckBuild converts the MASK statement into a DEPOSIT statement, followed
by a series of ETCH statements. The mask thickness and material type are defined in the
MaskViews Layers Popup (Figure 2-56) in the Define menu of MaskViews.

Figure 2-56 MaskViews Layers Menu

Two types of mask material are available: Photoresist and Barrier. The real thickness of a
photoresist layer should be specified because it can be used as a mask for implantation.
Barrier is a fictitious material. It is impenetrable for any implants and can serve only as a
masking material. This material is implemented in Athena for the purpose of simplifying
simulation of mask deposition over highly non-flat structures. Therefore, it is recommended
to use barrier thickness of 0.1 microns or less in most cases. A region to be etched may be any
area not containing a mask on a clear field layer or any area containing a mask on a dark field
area. You can specify this in the Field section in the MaskViews Layers Popup. In the case of
the POLY mask and cutline in Figure 2-56, the barrier layer will be etched to the left of x=0.8,
and to the right to x=1.8. The following echo output will appear in the DeckBuild Text
Subwindow as the result of defining of the POLY mask:
ATHENA> # DEFINING POLY MASK
ATHENA> ## MASK NAME="POLY"
ATHENA> DEPO BARRIER THICK=0.1 DIVIS=1
ATHENA> STRUCT OUTFILE=.HISTORY.9
ATHENA> ETCH BARRIER LEFT P1.X=0.8
ATHENA> ETCH BARRIER RIGHT P1.X=1.8
ATHENA> STRUCT OUTFILE=.HISTORY.11

If the Reverse Mask box is checked in the Athena Photo popup, the following lines will be
inserted into the input file:
# DEFINING POLY MASK
MASK NAME="POLY" REVERSE
and the effect of the field attribute is reversed (i.e., the barrier area will be etched between
x=0.8 and x=1.8).
When the mask is defined, the Athena dry etch capability can be used to etch the specified
thickness of a material not covered by the mask. After the dry etch is complete, strip the mask
by clicking the Strip Mask button in the Athena Mask popup (Figure 2-55). A typical mask
definition fragment should appear as follows:
# POLY DEFINITION
MASK NAME="POLY"
ETCH POLY THICK=0.5
STRIP
If the cutline from Figure 2-54 is loaded, this will give the structure shown in upper plot of
Figure 2-57. If the reverse parameter is added, the structure will appear as shown in the lower
plot of Figure 2-57.

Figure 2-57 Using Mask Capability for POLY Definition

Using Athena/Optolith Tutorial
2.9 Using Athena/Optolith

2.9.1 Overview
Athena/OPTOLITH is designed as an optical lithography tool integrated into a complete
process framework. Specific functions of Athena/OPTOLITH include 2D aerial image
formation, 2D photoresist exposure and development, post exposure bake, and post
processing capabilities such as CD extraction for generating SMILE plots. This section of the
tutorial describes Athena/OPTOLITH input/output and the following basic operations for
creating a typical input file for optical lithography:
• Creating an input mask using MaskViews or the LAYOUT command
• Designing custom or standard illumination systems
• Projection Fourier plane filtering
• Imaging controls
• Properties of materials
• Structure exposure, post exposure bake, and development
• CD extraction, SMILE plots, and looping procedures
This section of the tutorial assumes that you are familiar with the general operation of Athena.
This includes familiarity with the command language used to generate structures, as well as a
general knowledge of the use of the Interactive tools. Specific features that refer particularly
to OPTOLITH will be explained here.

2.9.2 Creating A Mask

A mask can be created using the MaskViews tool or by using the LAYOUT command.
MaskViews facilitates the creation of complicated masks and can import different mask data
formats, such as the GDS2 stream format. In the case of simple masks containing one or two
features, it may be simpler to use the LAYOUT command.
MaskViews
Once you select MaskViews from the Tools/MaskViews menu, press the Start
MaskViews button and the MaskViews window will appear. Then, press the Preferences
button and the MaskViews Properties popup will appear (See Figure 2-58).
Figure 2-58 MaskViews Preferences Popup showing the Simulation Menu

Open the Simulator menu and select the ATHENA/OPTOLITH menu item. Customized
controls for MaskViews/OPTOLITH will appear in the MaskViews window. The colored
buttons on the right side of the window are discrete controls for phase in degrees and intensity
transmittance. The buttons first appear as phase. To change to transmittance, open the Phases
menu above the buttons and select the Transmittances menu item. This will change the
buttons from phase to transmittance controls. Continuous controls for phase and
transmittance are located directly below the colored buttons.
The mask can now be designed using the mouse driven line writer following the description
outlined in the MaskViews User’s Manual. Once the mask is created, it should be saved to a
file with a name ending in a .lay extension for future editing. It is important to be aware that
there are two types of files that can be saved from a MaskViews layout information.
The first type is the layout file. This file includes the information about layers and mask
features. To store this information, select the Files Save menu item in the MaskViews
screen. The second type of file that can be saved from MaskViews is a file that is similar to
the layout file but is written to interface with Athena/OPTOLITH. The file to be used by
OPTOLITH is created by pressing the Write File button in the MaskViews window. The
Optolith Simulation Popup will appear (Figure 2-59).
Figure 2-59 OPTOLITH Simulation Control Popup

Enter the desired file name, which should end with a .sec extension, and proceed to the next
step. Note that at the bottom of the MaskViews window the message: Select first corner of
OPTOLITH simulation area will appear. MaskViews is now prepared for the selection of
the image window. The image window describes the area where intensity will be calculated.
Click on the desired area for intensity calculation to create the first corner of the OPTOLITH
simulation area. The message: Select the other corner of OPTOLITH simulation area will
then appear at the bottom of the MaskViews window. Click on the desired second corner.
Once this second point is selected, the coordinates of the image window’s lower left and
upper right corners will be displayed in the OPTOLITH Simulation Control Popup. Press the
Write button to save the OPTOLITH mask file. The input file created by MaskViews is
loaded into OPTOLITH by the IMAGE command, which is described in Section 2.9.5
“Imaging Control”.
To modify the layers, open the Define menu in the MaskViews Window (See Figure 2-48)
and select Layers menu item. The Layers Popup (Figure 2-60) will then appear.

Figure 2-60 Layers Popup

If you select Dark, the field background will be dark and the features will have the intensity
transmittance (T), where T is user-defined. If you select Clear, the intensity transmittance
automatically becomes [1 - T].
Only rectangular features are used in the imaging module. MaskViews automatically converts
triangles or polygons to a set of parallel rectangles. Finer resolution on these rectangles can be
obtained by changing the resolution on the Screen... popup under the Define menu.
Mask Layout
In the LAYOUT command, each mask feature is defined with one command line. For example:
LAYOUT X.LO=-0.5 Z.LO=-5.0 X.HI=0.5 Z.HI=5.0 TRANS=1 PHASE=0
defines a 1µ wide line that is 10 long. The mask has an intensity transmittance of one and a
phase of 0°. The LAYOUT command can be repeated as often as desired. The number of mask
features is limited only to the amount of memory available. The LAYOUT LAY.CLEAR
command will remove all previous mask features from memory. Overlapping mask features
will cause an error. The OPAQUE and CLEAR parameters can be specified in the IMAGE
command. This will not reverse polarity as it does in MaskViews.

2.9.3 Illumination System

The Illumination System is defined using two statements: ILLUMINATION and
ILLUM.FILTER. ILLUMINATION defines the illuminating wavelength, the possible x and z tilt
of the optical system and the relative intensity, which is usually set to 1. ILLUM.FILTER
defines the shape of the illumination system. The general shapes available are CIRCLE,
SQUARE, GAUSSIAN, ANTIGUASS, and SHRINC. The extent of the source must be defined to be
within a square centered at the origin as shown in Figure 2-61.
The extent of the source is defined by the coherence parameter, SIGMA. SIGMA defines the
radius for circular sources (CIRCLE, GAUSSIAN, and ANTIGAUSS), the x and y intercepts for
square sources, and the radius of each individual SHRINC source element as shown in Figure
2-62.
In all cases, anything outside of the square defined by SIGMA=1 will be ignored. The SHRINC
source position is defined by the RADIUS and ANGLE parameters as shown in Figure 2-62. The
SHRINC source can be defined by the command:
ILLUM.FILTER SHRINC RADIUS=0.25 ANGLE=45 SIGMA=0.1
Figure 2-61 Maximum Extent of the Source Region

Figure 2-62 Three Different Source Types

Arbitrary sources can be defined by using the ANGLE and RADIUS parameters. Phase and
intensity transmittance of each source element are controlled by the parameters PHASE and
TRANSMIT. By positioning each source element in the source region, you can simulate any
type of source. To simulate a SHRINC source, enter the following command lines:
ILLUM.FILTER CIRCLE SIGMA=0.1 RADIUS=0.25 ANGLE=45
If overlapping sources are defined, a warning is issued and the most recent source is used. If
the overlap is partial, only the overlap area is overwritten by the most recent source.
Annular filters can also be superimposed on the source. There are two types of annular filters:
Square and Circle. Annular filters have a multiplicative effect on the source. Because of this,
be careful when defining a complex source and a complex filter. The following example of an
annular source of inner radius 0.4 and outer radius 0.6 is given below.
ILLUM.FILTER CIRCLE SIGMA=0.6
ILLUM.FILTER CIRCLE INNER.RAD=0.0 OUTER.RAD=0.4 TRANSMIT=0.0
In the first statement, the SIGMA parameter defines the outer radius. In the second statement,
an opaque spot is defined as an annular filter (see Figure 2-63).
The source must be described before the filter or the command will be ignored and a warning
will be given. The CLEAR.FILTER parameter is used to remove all pre-existing filters and
sources.

Figure 2-63 Annular Source

2.9.4 The Projection System

The Projection System is defined using two statements: PROJECTION and PUPIL.FILTER.
The PROJECTION command is used to define the numerical aperture and flare of the
projection system. The PUPIL.FILTER command describes the shape of the projection
system and the possible filters of the projection system. The shape of the projector pupil can
be square or circular. The circular pupil has the option of having a Gaussian or anti-Gaussian
transmittance profile. Filtering of the Fourier spectrum can be performed by using annular
filters. The filters have a multiplicative effect on the transmittance and phase in the projector
pupil. The following example creates an opaque square at the origin:
PUPIL.FILTER SQUARE
PUPIL.FILTER SQUARE INNER.RAD=0.0 OUTER.RAD=0.1 TRANS=0.0
This creates the following projection pupil (Figure 2-64):
Figure 2-64 Projection Pupil

The maximum extent of the projector pupil plane is +1 or -1 in both dimensions. A filter
exceeding these dimensions will be ignored and a warning will be issued.

2.9.5 Imaging Control

The image calculation is done by the IMAGE command and its associated parameters. The
mask to be imaged will already be defined either by a MaskViews *.sec file or by the LAYOUT
command. If a MaskViews *.sec file is used, the IMAGE command will be of this form:
IMAGE INFILE=*.sec ...
If the mask is defined using the LAYOUT command, the mask features will be stored in
memory and the only required input related to mask features is the OPAQUE/CLEAR
specification. OPAQUE specifies the background intensity transmittance to be zero. CLEAR
specifies the background intensity transmittance to be one. OPAQUE is the default setting.
OPAQUE and CLEAR cannot be used with an input file from MaskViews.
The Image Window (not the Computational Window) is specified with the parameters:
WIN.X.LOW, WIN.Z.LOW, WIN.X.HI, and WIN.Z.HI. These parameters define the
minimum and maximum range of the x and z values (see Figure 2-65). The aerial image is
then calculated only inside this window. This allows for faster computation when you only
want a cross section. If you want a simple cross section, set the window parameters for z
(WIN.Z.LO, WIN.Z.HI) to the same value for a cross section parallel to the x-axis.
This value(WIN.Z.LOW = WIN.Z.HI) gives the location of the cross-section.
Figure 2-65 The Image Window can be Placed Anywhere in the XZ Window
The resolution in the image window can be controlled by two different sets of parameters.
The first set is DX and DZ. DX and DZ are the resolution in micrometers for the x and z
directions. The second set, X.POINTS and Z.POINTS, is based on the number of points in
each direction. The resolution will be the length of the side of the image window divided by
the number of points. To study the defocus of the aerial image, use the DEFOCUS parameter.
DEFOCUS uses units of micrometers.

The N.PUPIL parameter specifies the computational window. If N.PUPIL is not specified, it is
automatically calculated to a size that encompasses all mask features. In these cases, you can
set the computational window manually using the following formula:
length  NA
N.PUPIL = ----------------------------- * 2 2-1
lambda
where length is the intercept coordinate for the x and z axes of a square centered at the origin
that delimits the Computational Window as shown in Figure 2-66.
Figure 2-66 The Computational Window is Always Centered at the Origin
Once you calculate the image, you can store it in a standard structure file by using the
STRUCTURE OUTFILE=*.STR INTENSITY command.
The INTENSITY modifier identifies the file to be different than a standard structure. This file
can later be initialized into memory and used without running the imaging module. To
initialize an intensity file, type:
INITIALIZE INFILE=*.STR INTENSITY
The intensity modifier again specifies the type of file. An intensity file initialized in this
fashion is useful only for exposures that use the vertical propagation model.
N.PUPIL also affects the accuracy of the aerial image calculation. A higher N.PUPIL value
increases the number of source points by a factor (2*N.PUPIL+1) squared, and increases the
accuracy and the computation time.
Note: The Image Window and the Computational Window are not linked. The computational window is
automatically adjusted to include all mask features, unless otherwise specified in the IMAGE command.
This means that the entire mask will be used in the image calculation. You can use the Image Window to
calculate a part of the entire image to increase the simulation speed. You can override the selected image in
the IMAGE command by specifying new window coordinates.

Aerial image intensity distributions can be added together by specifying MULT.IMAGE on

repetitions of the IMAGE command. You can add any number of images together. The first
IMAGE statement shouldn’t contain the boolean parameter, MULT.IMAGE, because the
preceding aerial images are erased from memory. You can weigh of the aerial images by using
the INTENSITY parameter on the ILLUMINATION command.
ONE.DIM is a new parameter that has been added to the IMAGE command. It allows you to
calculate one dimensional aerial images. This is used primarily for increasing speed in the
exposure calculation for repetitive line width calculations.
2.9.6 Defining Material Properties

There are two statements in Athena/OPTOLITH that relate to material properties: OPTICAL
and RATE.DEVELOP. The OPTICAL command sets the complex index of refraction for a single
material at a given wavelength. The RATE.DEVELOP command sets development rate
parameters for each resist defined in the resist library. Default values for these material
parameters are located in the athenamod file, which can be viewed in DeckBuild by selecting
the Models... item from the Command menu. You can change any of these parameters by
entering the command with the new values. For example, to change the index of refraction of
silicon at the wavelength of 0.365 µm, enter:
OPTICAL SILICON LAMBDA=0.365 REFRAC.REAL=6.522 REFRAC.IMAG=2.705
To enter resist parameters at wavelength of 0.407µm, use the RATE.DEVELOP command as
follows:
RATE.DEVELOP NAME.RESIST=AZ135OJ LAMBDA=0.407 \
A.DILL=0.88 B.DILL=0.077 C.DILL=0.018 E1.DILL=5.63 \
E2.DILL=7.43 E3.DILL=-12.6
Photoresist parameters for development or diffusivity (DIX.0, DIX.E) can be entered
separately from exposure parameters without specifying the wavelength. The photoresist
name must always be specified. When specifying Dill exposure parameters (A, B, and C),
specify the wavelength as these parameters vary with wavelength.
2.9.7 Structure Exposure

Exposure, post exposure bake, and development each have separate statements (EXPOSE,
BAKE, and DEVELOP) respectively. In order to use these three statements, some initial
requirements must be met. First, an intensity cross section or Fourier Spectrum data must be
available. Second, you must create a structure including photoresist using the techniques
described in Section 2.3 “Creating a Device Structure Using Athena”.
This intensity cross section can come from three different places. The first method is by
running the imaging module prior to exposure. This puts the intensity data array into memory.
The second method is by initializing with an intensity data array that has been stored in a
standard structure file (see Section 2.9.5 “Imaging Control”) using the following command:
INITIALIZE INFILE=*.STR INTENSITY
The INTENSITY qualifier lets Athena know that this is an intensity file, as opposed to a
standard structure file. After this command is entered, the intensity data array will be placed
in memory. Wavelength will be stored in this file and can be changed only by re-running the
imaging module.

The third method of entering an intensity cross section is through a user data file. The file
should contain the wavelength, the number of data points, and the intensity and position of
each point. The first line of this file should contain the wavelength in micrometers. The
second line should contain the number of points. The following line should contain the
position and then the intensity of the first point on the same line. This should be repeated for
each point. This input file is read in the EXPOSE command using the following format:
EXPOSE INFILE=*.EXP
Once the intensity array is initialized or when the Fourier spectrum data is in memory through
the IMAGE command, you can expose a structure if it exists in memory and if it has
photoresist as its top layer(s). You can either create the structure in the input file or initialize it
as described in Section 2.3 “Creating a Device Structure Using Athena”.
The EXPOSE command has many parameters that control the accuracy and speed of the
exposure simulation, as well as related imaging parameters. The following parameters control
simulation speed and accuracy and are unnecessary for a preliminary simulation.
FLATNESS, NUM.REFL, FRONT.REFL, BACK.REFL, ALL.MATS
The most important of these parameters is the FLATNESS parameter. If FLATNESS is set equal
to zero, the algorithm uses the entire grid for the calculation and may lengthen the simulation
time. The remaining parameters refer to the image to be exposed. Both TE and TM modes are
available in exposure, but they must be performed separately. Select TE by adding the
PERPENDICULAR parameter to the EXPOSE command, or select TM by entering the PARALLEL
parameter. TE is the default. The exposure dose is also defined in the EXPOSURE command in
units of mJ/cm2 using the DOSE parameter.
Exposures can be made with either coherent or incoherent sources. Coherent sources are
described by SIGMA=0.01 in the IMAGE command. This defines a small enough source so
that only one discretization point is included. If a large SIGMA is defined and discretization of
the source allows at least three source points in the x (or z) direction, then three points from
the source will be used in the bulk image calculation with equal weight given to each point.
The points chosen will be the central point and the outermost points, or the dimension of the
chosen cross section (x or z). If multiple sources are defined using the ILLUM.FILTER
command, then the central point of each SOURCE defined is used for calculating the bulk
image in the exposure. The latter allows an arbitrary amount of source points to be simulated
for the bulk image calculation. This is done by specifying many small adjacent sources and
one point will be taken from each source.
You can add bulk image exposures together by specifying MULT.EXPOSE on repetitions of the
EXPOSE command. Any number of exposures can be added together. The first EXPOSE
statement should not contain the boolean parameter MULT.EXPOSE because the preceding
exposures are erased from memory. You can weigh the exposures by using the DOSE
parameter on the EXPOSE command.
The final four parameters: X.CROSS, Z.CROSS, CROSS.VALUE, and X.ORIGIN all refer to the
aerial image cross section. The boolean parameters (X.CROSS or Z.CROSS) define the cross-
section to be parallel to the x-axis or the z-axis respectively. CROSS.VALUE specifies the z
location of X.CROSS, or the x location of Z.CROSS. These parameters are especially useful
when several cross sections from the one large, two dimensional aerial image will be
simulated. X.CROSS is the default. If CROSS.VALUE is unspecified, the center of the image
window, defined in the imaging module, will be used as CROSS.VALUE. X.ORIGIN allows the
aerial image cross section to be shifted laterally in the two dimensional exposure simulation.

Post Exposure Bake

The BAKE command can be invoked by using only one parameter DIFF.LENGTH (the diffusion
length). For a post exposure bake of 60 seconds at a temperature of 125°C, the recommended
diffusion length is between 0.05 and 0.1 micrometers. The BAKE command can also be used
by specifying TIME and TEMPERATURE in C°.
Development
The development module offers a choice of six different development models. As mentioned
before, model parameters are specified in the RATE.DEVELOP command. After the
development model is selected, the three primary parameters for the DEVELOP command are:
TIME (in seconds), STEPS, and SUBSTEPS.
TIME is the total development time. STEPS specifies the number of times the structure has to
be regridded. SUBSTEPS is the total number of times the development line has to be moved.
Each substep is performed for a time increment equal to TIME/(STEPS*SUBSTEPS). After
each regridding of the structure, you can dump out a standard structure file to show the
progress of the development. To do this, specify the parameter DUMP=1. To name the structure
file to be dumped, specify DUMP.PREFIX=<name> and the structure will be created in the
local directory with the name <name>**.str, where ** is the current development time.
Post Development Bake
A physically-based reflow of the developed photoresist is available. Specify it by using the
BAKE command and the boolean parameter REFLOW along with TIME and TEMPERATURE.

Adaptive Meshing Tutorial
2.10 Adaptive Meshing

2.10.1 Introduction to Mesh Adaption
Athena has a built-in mesh adaption module that automatically adapts the grid to dopant
profiles. Used together with implantation and diffusion, the module can achieve more
optimized accuracy of a given profile’s representation for a given number of grid points. This
relieves you to some extent from the time consuming mesh generation task in the simulation
structure preparation stage. It will also improve the accuracy and speed of the subsequent
diffusion/oxidation/epitaxy stages where impurity profiles change with time.
The algorithm used was suggested in [3], [4]. It uses an efficient local error estimator and a
triangulation scheme suitable for complex two-dimensional moving boundary problems.
Adaption During Ion Implantation
Ion implant is a common process step to introduce impurities into the substrate to form active
device regions. Prior to the implant step, it is difficult to determine the required mesh density
distribution because the exact dopant profile is unknown before processing. Thus, you can
only estimate the profile and required mesh. It’s a time consuming process to specify mesh
generation statements to create the mesh with a density conforming to an estimated profile.
Graphical tools, such as Devedit, can make this easier. But it can’t totally eliminate the
process. With the Adaptive Meshing module, you can overcome these difficulties to a larger
extent.
The program uses an iterative algorithm to determine the required mesh density distribution
to accurately conform to the implanted profile, and will automatically generate the additional
required mesh. The algorithm is illustrated with the flow chart depicted in Figure 2-67.

Figure 2-67 Flow Chart of Mesh Adaption Algorithm

Adaption During A Heat Cycle
During the diffusion/oxidation/epitaxy processes, impurity profiles are usually changing
continually with each elapsed time step. An initially generated optimal mesh will not conform
to the time varying dopant profile. If the impurity profiles change substantially during the
process, the mesh density distribution will be different from the dopant contour distribution,
causing both accuracy and speed problems.
During simulation, the total time of a diffusion/oxidation/epitaxy process is usually divided
into many small time steps, with profiles changing gradually between time steps. Using the
Adaptive Meshing module, you can perform a mesh adaption after each time step. This
allows the mesh to conform to the dopant after a time step.
The difference between the dopant contours and the change in the mesh density distribution
will thus be limited to the difference of dopant profiles between time steps. This difference is
substantially smaller than that over total diffusion time. Thus, mesh adaption can allow more
accuracy and minimize the mesh density for the dopant representation at any given time.

A Simple Example:
GO ATHENA
INIT SILICON C. ARSENIC=10E14
DIFF TIME=50 TEMP=950 DRYO2
DEPOSIT POLY LEFT PL.X=1.2
ETCH POLY LEFT PL.X=1.2
STRUCT OUTF=MOS_0.STR
#PERFORM ADAPTIVE MESHING FOR BOTH IMPLANT AND DIFFUSION
METHOD ADAPT
IMPLANT BORON DOSE=1.0E13 ENGERY=15 PERSON TILT=0
DEPOSIT OXIDE THICK=.35 DIV=6
ETCH OXIDE THICK=.35
IMPLANT BORON DOSE=1.0E14 ENGERY=15 PERSON TILT=0
DIFFUSE TEMP=1000 TIME=30
QUIT
LISTING 1: A SIMPLE EXAMPLE OF IMPLANT ADAPTIVE MESHING

This simple example creates a LDD MOS device structure. The initial simple mesh is specified
with the four LINE statements. This initial mesh is referred to as a base mesh and options for
its formation will be discussed the “The Mechanics of the Base Mesh Formation” on 120.
After some initial 1D processing, the adaptive meshing function is invoked. Subsequently,
automatically adds mesh that conforms well to the two implanted Boron profiles. During the
final DIFFUSE statement, Boron has been driven down into the substrate and tessellated with
the initial simple mesh. The mesh adaptation module adapts after each time step. This results
in meshing conforming to the Boron profile throughout the diffusion process.
The mesh adaption module is invoked during the simulation by specifying boolean flag
ADAPT on the METHOD command preceding IMPLANT, DIFFUSE, or EPITAXY statements. The
syntax behind this simple example using the mesh adaption module is shown below.
Three statements are available to access the mesh adaptation module, they are briefly
described as the following. The METHOD statement is used to control numerical algorithms.
When METHOD ADAPT is specified, the mesh adaptation algorithm will be used. If you specify
METHOD ADAPT=false, the mesh adaptation algorithm will be turned off. ADAPT is off by
default.

ADAPT specifies that the adaptive meshing should be performed on any of the following
IMPLANT, DIFFUSE or EPITAXY statements. Adaptation is performed by following each step
on each DIFFUSE/EXPITAXY statement. IMPLANT.MES specifies which adapting algorithm to
use on IMPLANT statements. Currently, IMPLANT.MES=0 corresponds to the University of
Florida’s algorithm. This is the default. Also currently, this is the only recommended
algorithm. There are four other parameters on the METHOD statement that specify mesh
smoothing.
They are as follows:
• ETCH.SMOOTH specifies that mesh smooth operation will be performed after etch.
• DEPO.SMOOTH specifies that mesh smooth operation will be performed after deposit.
• DIFF.SMOOTH specifies that mesh smooth operation will be performed after diffusion.
• STEP.SMOOTH specifies that mesh smooth operation will be performed after each
diffusion time step. These four parameters are currently set as default.
The ADAPT.PAR statement is used to set parameters to adjust the mesh adaptation process.
The parameters available on the ADAPT.PAR statement are the following. Specify material
regions to be adapted, such as SILICON, OXDIDE, and POLYSILICON. This may be one or
several materials at a time. The default impurities include such as I.BORON or I.ARSENIC.
Specify impurities to be adapted on. This may be one or several impurities at a time. The
DISABLE parameter specifies materials/impurities given disabled to be effective on mesh
adapting or smoothing. The MAX.ERR parameter specifies the maximum errors allowed before
adding points to the mesh (unitless). Errors calculated above this value cause points to be
added. The MIN.ERR parameter specifies the minimum error below which points may be
deleted from the mesh (unitless). Error calculated below this value causes points to be
removed. Both MAX.ERR and MIN.ERR are calculated using the Bank-Weiser Error Estimator
which is defined as
2   Ci 
c = h * -------------- 2-2
 Ci 
where h is the average of the edge lengths associated with node i. Ci is the impurity
concentration at node i. The parameter, CONC.MIN, specifies the minimum impurity
concentration below which adapting will stop (units 1.0/cm3).
The ADAPT.MESH statement is used to do mesh adaptation for a given device structure without
coupling implant diffusion/epitaxy to the process. Therefore, the mesh adaptation module can
be used to assist the manual mesh generation process. The following parameters are available
on the ADAPT.MESH statement:
• ADAPT specifies a stand alone adaptive meshing step should be performed to refine or
relax the current mesh based on the material/impurity specification given on the
ADAPT.PAR statement (default false).
• ADAPT.COUNT specifies that stand alone annealing be performed during the execution of
the ADAPT.MESH statement (default false).
• SMTH.COUNT specifies the number of smooth loops during the smoothing algorithm.

Adaptive Meshing Control

Adaptive meshing may be used in several different modes and has several statements to
control it. This section describes the adaptive meshing related statements and how to use
them. Table 2-6 list these statements.
Table 2-6 Summary of Adaptive Meshing Control
Parameter Description
METHOD Switches the various automated adaption modes on and off.

ADAPT.MESH Invokes a stand alone adaption of the mesh at a specific point.
ADAPT.PAR Control both the stand alone adaption and the automatic adaption meshing
criteria.
GRID.MODEL Describes an external template file, containing mesh related statements specific to a
general technology or device.
BASE.MESH Defines the 1D starting point of a mesh for an adaptive mesh based simulation.
BASE.PAR Specifies the adaption criteria for the base mesh only.
The Mechanics of the Base Mesh Formation

Athena uses adaptive meshing in both 1D and 2D modes. ADAPT.PAR parameters control both
these modes. The concept of the Base Mesh, however, needs to be described. A typical
simulation (e.g., a MOS) is simulated in 1D initially and then switched to 2D during mid-
process flow, perhaps at the Poly Gate definition process step. Here, the mesh is extruded
from 1D to 2D and the result is the base mesh. The Base Mesh then forms the basis, and is the
starting point, for 2D Adaptive Meshing. The mesh quality of this base mesh is important for
success of future adaption for example, during source-drain implants and anneals.
2.10.2 Interface Mesh Control

The Interface Mesh Control is used to control the mesh in the vicinity of a material interface.
This function allows you to add grid lines, for example, to run along under the gate of MOSFET
at some distance from the Si/Si02 interface. The Interface Mesh Control is often useful for
adding mesh as required by highly mesh dependent mobility models during a following
device simulation. It is also useful to be able to add mesh for better segregation modeling.
The ADAPT ADD.I.LINE=n command controls the addition of a new mesh line. Two materials
are specified as parameters to the command defining an interface or a set of interfaces. The
mesh line is added to MATERIAL1 as follows:
ADAPT.MESH ADD.I.LINE=0.001 MATERIAL1 / MATERIAL2
For example, in the case of adding an additional mesh line to the SILICON in the channel
region of a MOSFET:
ADAPT.MESH ADD.I.LINE=0.001 SILICON / OXIDE
The structural transition from 1D to 2D to create a base mesh is controlled by the BASE.PAR
parameters. Figure 2-68 indicates the flow of events towards the formation of a base mesh
and beyond in the case of MOSFET device.

Figure 2-68 MOSFET Device Mesh Formation Flow

The base mesh quality is important to allow a subsequent adaption in 2D. The adjacent ratio
of elements, both in 1D and 2D relate directly to the smoothness of the final mesh quality.
The generation of a high quality adapted mesh starts with the BASE.MESH command. Here,
the 1D mesh is defined from where the final 2D mesh will evolve. The BASE.MESH command
defines a 1D structure as a stack of up to five layers. Five layers are used to define the five
layers of a Bipolar device. Each layer is described as having a thickness: SURF.LY,
ACTIVE.LY EPI.LY, SUB.LY, and BACK.LY and an associated mesh spacing per layer,
SURF.DY, ACTIVE.DY, EPI.DY, SUB.DY, and BACK.DY. The whole structure can also be offset
in space with the point of origin determining the top left hand corner of the structure. The
OFFSET.X and OFFSET.Y parameters are used for this purpose.
An example of using the offset command might be defining the starting surface of an initial
structure an epi thickness below the zero position. That way, the subsequent geometrical
calculations are made easier.

Figure 2-69 indicates the relationship of the BASE.MESH command to the initial 1D structure
mesh.
Figure 2-69 Initial 1D Structure Relationship

The BASE.MESH parameters should be considered alongside the BASE.PAR parameters. When
forming a base mesh, there are three objectives to remember regarding the quality of mesh.
These objectives are as follows:
• 1D dopant information is neither lost in the 2D transition nor overly refined upon,
resulting in overly dense BASE.MESH (See Figures 2-70 and 2-71).
• Little or few flat triangles exist in regions and materials of importance. (See Figure 2-74).
• The adjacent triangle ratio, in both X, and Y directions, is not abrupt in spacial regions of
importance to the device (See Figure 2-75).

Controlling the quality of the BASE.MESH, formed at the 1D-2D transition, is achieved with
the BASE.PAR command parameter. Specific materials can be assigned different parameters.
The GRAD.SPACE parameter controls the Vertical Adjacent Triangle Ratio Quality, while the
RATIO.BOX parameter controls the lateral Adjacent Triangle Ratio. These two statements can
be thought of as operating upon the 1D and 2D simulation segments respectively during 1D
simulation. Only the adjacent spacing ratio can be controlled in the vertical profile with the
GRAD.SPACE parameter. Subsequently, at the point of 2D transition, the RATIO.BOX
parameter is used to trade off mesh quality for mesh density.
The INIT command includes the parameters: WIDTH.STR and DEPTH.STR. These parameters
define the size of the initial structure and will truncate the previous BASE.MESH defined
structure.
Figures 2-74 and 2-75 show an example of this base mesh and of the subsequent 2D diffusion.
Figure 2-70 Mesh that is too coarse leads to Dopant Information Loss
Figure 2-71 Too Dense Mesh Causes Too Much CPU Time during Subsequent Simulation

Figure 2-72 Large and reasonable Adjacent Triangle Ratios
Figure 2-73 Base Mesh Formation

Figure 2-74 Automatic Base Mesh Generation
Figure 2-75 Ion Implantation Adaption on New Base Mesh

Chapter 3
SSuprem4 Models
Diffusion Models SSuprem4 Models
3.1 Diffusion Models

The diffusion models in Athena describe how implanted profiles of dopants/defects (see the
Note below) redistribute themselves during thermal treatment, due to concentration gradients
and internal electric fields. When modelling the actual diffusion process, there are additional
effects to consider such as impurity clustering, activation, and how interfaces are treated.
Fundamentals of the models described in this section could be found in [5], [6], and [7].
Note: In the following sections, the terms impurity and dopant shall be used interchangeably, although an impurity
doesn’t necessarily have to be a dopant. Also the term, defect, shall mean the same as point defect, unless
otherwise indicated in the context.
Diffusion of dopants and point defects in SSuprem4 is described by a number of user-

specifiable models. The three most basic models are the following:
• The Fermi diffusion model.
• The two dimensional diffusion model.
• The fully coupled diffusion model.
The models are natural extensions of each other in the sense that the Fermi model is included
in the two dimensional model, which is included in the fully coupled model. The two
significant differences between them are the way point defects are represented and treated
throughout the simulation, and how the specific dopant diffusivities are formulated. The
selection of which model to use will depend upon the existence or the generation of point
defects during the diffusion process and the dopant concentrations within the silicon. Careful
reading of the following sections is critical to understanding which model to use.
All three models rely on the concept of Pair Diffusion, which says that a dopant atom cannot
diffuse on its own – it needs the assistance of a point defect (a silicon self interstitial or a
lattice vacancy) in the near vicinity as a diffusion vehicle. If there is a non-vanishing binding
energy between the two, they can move as one entity (a pair) through a number of jumps and
inversion cycles before eventually breaking up. When speaking of dopant diffusivity within
the scope of these models, one actually means the diffusivity of the pair as a whole. A point
defect, however, can either diffuse freely or as a participant in a dopant-defect pair. The
diffusivity of a free point defect can actually be different from the diffusivity of a point defect
pair.
All diffusion models in Athena also use the concept of Chemical and Active Concentration
Values. The chemical concentration is the actual implanted value of the dopant but when
dopants are present at high concentrations, clustering or electrical deactivation can occur so
that the electrically active concentration may be less than the corresponding chemical
concentration. This is described in Section 3.1.6 “Electrical Deactivation and Clustering
Models”.
Athena creates structures that can have multiple materials and interfaces such as the
polysilicon-oxide-silicon interface in MOSFETs. Each interface within Athena has boundary
or interface conditions that model impurity segregation. The model details are described later
in this chapter. You should, however, be aware that the gas/solid interface (the surface of the
silicon if exposed) and solid/solid interfaces have been strictly modelled within Athena.
Effects such as dopant loss from exposed silicon and dopant pile-up at interfaces are
simulated.

3.1.1 Mathematical Description

The mathematical definition of a diffusion model includes the following specifications for
every diffusing species present:
• a Continuity Equation (often called a Diffusion Equation).
• one or more flux terms.
• a set of boundary and interregional interface conditions.
In the case of impurity diffusion in semiconductors, we need a set of equations for each
dopant present and for each type of point defect if point defects are explicitly represented in
the model. Since dopants can only diffuse as participants in dopant-defect pairs, the dopant
continuity equation is actually a continuity equation for defect-dopant pairs.
The formulation of the continuity equation have a number of built-in assumptions:
• Electronic processes take place on a time scale, which is much smaller than the time scale
of all other processes (adiabatic approximation).
• The pairing reaction between dopants and defects is assumed to always be in equilibrium.
This may not be the case, especially at a low temperature, but would pose a much harder
and more CPU intensive numerical problem to solve.
• Mobile dopants are electrically active and vice versa.
Models that explicitly take pair populations into account have been implemented by various
research groups. But all these models suffer from the lack of established experimental data,
such as binding energies, or pairing coefficients, for which reason the predictability of these
models is questionable. The lack of data, especially for the energy levels of the different
charge states of the point defects in the band gap at typical diffusion temperatures, poses a
serious gap in our knowledge. Some of these energy levels have been measured in low
temperature experiments such as DLTS (deep level transient spectroscopy), but no one knows
how these levels adjust themselves relative to the band edges when the band gap narrows as a
function of increasing temperature.
Van Vechten [8] has theoretically argued that the acceptor states (0/- and -/=) and the donor
states (+/++ and +/0) of the mono vacancy follow the conduction band edge with increasing
temperature. Mathiot [5], however, chooses to scale the positions of the energy levels relative
to the band edges with the size of the band gap.
In addition to the models described above, which are all specific for dopants and point defects
in silicon, there is a smaller number of hard coded models that are used for other materials
such as oxide or poly.
In the sections that follow, we apply standard notation used in the literature for dopants, point
defects (interstitials and vacancies) and the different charge states as shown in Table 3-1. In
Table 3-1, the x designates the neutral charge state. - is a single negatively charged state. = is
a double negatively charged state.

Table 3-1 Notational standards in diffusion literature
Physical Entity Generic Symbol Replacement Values
Dopant A B, P, As, Sb,...

Point Defect X I, V
Charge State c x, -, =, +, ++
Many physical entities or parameters are temperature-dependent. In Athena, this dependence

upon temperature is modelled by the Arrhenius expression (unless otherwise indicated):
Q  T  = Q.0 exp  – ---------- 

Q.E
3-1
 kT 
where:
• Q.0 is the pre-exponential factor,
• Q.E is the activation energy,
• k is the Boltzmann constant,
• T is the absolute temperature.
Generic Diffusion Equation
All diffusion models, whether they are the Fermi, the two dimensional or the fully coupled
model, follow the same generic mathematical form of a continuity equation. A continuity
equation merely expresses particle conservation, that is, the rate of change with time of the
number of particles in a unit volume must equal the number of particles that leave that
volume through diffusion, plus the number of particles that are either created or annihilated in
the volume due to various source and sink terms.
This basic continuity equation for the diffusion of some particle species (C) in a piece of
semiconductor material is a simple Second Order Fick’s Equation [9]:
 C Ch
- = – JA + S
----------- 3-2
t
where CCh is the total particle (chemical) concentration, JA is the flux of mobile particles,  
is the gradient operator, and S accounts for all source and sink terms. The difference between
the total (chemical) concentration and the actual mobile concentration is described in a later
section entitled Section 3.1.6 “Electrical Deactivation and Clustering Models”. In
semiconductor diffusion problems, there are generally two contributors to the particle flux.
The first contributor is an Entropy Driven Term, which is proportional to the concentration
gradient of mobile particles. The coefficient of proportionality, DA, is called the diffusivity.
The second contributor is a Drift Term, which is proportional to the local electric field. Notice
that if there are several types of electrically charged species present, this term establishes a
coupling between them, since all charged particles both contribute to and are influenced by
the local electric field.
The Flux Term, JA, can be written as:

JA = –DA  C   CA + CA  E 3-3
where CA designates the mobile impurity concentration,  is the mobility, and E is the electric
field. It should also be observed that Equation 3-3 is non-linear, since both the diffusivity DA
and the electric field E in general depend on the concentration of all present species. In
thermodynamical equilibrium, the Einstein relation relates mobility and diffusivity through
the expression D = kT
------  .
q
Substituting for  in Equation 3-3 is writing the particle charge as a signed integer, ZA, times
the elementary charge, q, giving us this Flux Expression.
J A = – D A  C    C A – Z A C A qE
------- 3-4
 kT 
In insulator and conductor materials, the electric field is zero. In semiconductor materials, the
electric field is given by:
E = –  = – kT
------  n 3-5
qn
where is the electrostatic potential and n is the electron concentration. If charge neutrality is
assumed, then the electron concentration may be rewritten as:
ND – NA  ND – NA 2
- +  --------------------  + n
2
n = ------------------- 3-6
2  2  i
where ND and NA are the electrically active donor and acceptor impurity concentrations, and
ni is the intrinsic carrier concentration calculated as:
n i = NI.0  exp  – ------------ T

NI.E NI.POW
3-7
 kT 
where you can specify the NI.0, NI.E, and NI.POW parameters in the MATERIAL statement.
The electrically active and mobile impurity concentrations are equivalent.
Boundary conditions
Boundary conditions within Athena are of mixed type and are expressed mathematically as:
  CA +    n CA  = R 3-8
where ( are real numbers and nCA designates the flux of CA across the boundary. The
right hand term, R, accounts for all source terms on the boundary. Boundary conditions are
applied at two main regions.
The first region is at the top of the simulation region (the surface). The second region is at the
inter-regional interfaces for which the species in question only has a meaningful existence in
one of the region materials (e.g., an interstitial on a silicon/oxide interface).

Interface conditions
Between any two regions there must be some control on how any impurity species can exist
in the vicinity of the interface. For every such interface, you must specify a Concentration
Jump Condition and a Flux Jump Condition.
The Concentration Jump Condition accounts for discontinuities in particle concentrations
across interfaces and encompasses particle transport across material interfaces due to
different solid solubility ratios of the impurity species in the two materials.
The Flux Jump Condition enables the formulation of interface source and sink terms such as
surface recombination, particle injection, and particle pile-up at a moving interface.
For all species, no flux boundary conditions are employed on the sides and at the bottom of
the simulation structure. This is hardwired into the software, which means it is not user-
definable.
3.1.2 The Fermi Model

The Fermi Model assumes that point defect populations are in thermodynamical equilibrium
and thus need no direct representation. All effects of the point defects on dopant diffusion are
built into the pair diffusivities. The main advantage for using the Fermi Diffusion model is it
will greatly improve the simulation speed, since it does not directly represent point defects
and only needs to simulate the diffusion of dopants. Also, the Fermi Model usually results in
an easier numerical problem due to the avoidance of “numerical stiffness”. But since point
defects are not directly simulated, the Fermi model cannot deal with certain process
conditions in which the defect populations are not in equilibrium, such as in wet oxidation
(where Oxidation Enhanced Diffusion (OED) is important), emitter-base diffusions and
wherever implantation results in an initial high level of implant damage.
To use the Fermi Model, specify FERMI parameter in the METHOD statement.
In the Fermi Model, each dopant obeys a continuity equation of the form:
 C Ch
  D AX   C A – Z A C A qE
-------
------------- =  kT  3-9
t X = I, V
where CCh is the chemical impurity concentration, ZA is the particle charge (+1 for donors
and -1 for acceptors), DAV and DAI are the joint contributions to the dopant diffusivity from
dopant-vacancy and dopant-interstitial pairs in different charge states [5]. CA is the mobile
impurity concentration and E is the electric field. The terms DAV and DAI depend on both the
position of the Fermi level as well as temperature and are expressed as:
– n 1 n 2 + n –1 + + n –2
D AX (T,----) = D AX + D AX  ---- + D AX  ---- + D AX  ---- + D AX  ----
n x =
3-10
ni n  n  n  n 
i i i i
where the temperature dependency is embedded in the intrinsic pair diffusivities, which are
specified by Arrhenius expressions of the type:
 c 
c c
D AX = D.0 AX exp  – D.E AX
-
----------------- 3-11
 kT 

Table 3-2 shows the names of the pre-exponential factors, D.0, and activation energies, D.E,
for each of the charge states, c, of the various intrinsic pair diffusivity terms.
Pair charge states beyond two are unlikely to occur, which is why they have been omitted.
Also, for most dopants it is seldom that more than three of the terms above are non-vanishing.
Table 3-2 Table of intrinsic pair diffusivities for different pair types
Pair Charge State Pre--exponential factor Activation Energy
AV  DVX.0 DVX.E
AV - DVM.0 DVM.E
AV = DVMM.0 DVMM.E
AV + DVP.0 DVP.E
AV ++ DVPP.0 DVPP.E
AI  DIX.0 DIX.E
AI - DIM.0 DIM.E
AI = DIMM.0 DIMM.E
AI + DIP.0 DIP.E
AI ++ DIPP.0 DIPP.E
Note: Since the point defect populations are by definition assumed to be in equilibrium in the Fermi model, there
are no separate continuity or boundary condition equations for these species. Additionally, neither the
vacancy concentration, CI, nor the interstitial concentration, CV, appear explicitly in Equations 3-9, 3-10, or
3-11.
3.1.3 Impurity Segregation Model

In multilayer structures, dopant segregation across material interfaces must be considered.
Such interfaces can represent either a solid/solid interface or a gas/solid interface (the
surface).
Interface segregation is modeled empirically by a first order kinetic model for the
interregional flux:

C1
F s = h 12   --------
- – C 2 3-12
 M 12 
where:
• C1 and C2 are the particle concentrations in the immediate vicinity of the interface in the
regions 1 and 2.
• h12 is the interface transport velocity.
• M12 is the segregation coefficient.
The transport velocity and segregation coefficients are temperature-dependent parameters
defined through the following Arrhenius expressions:
h 12 = TRN.0  exp  – -----------------

TRN.E
3-13
 kT 
M 12 = SEG.0  exp  – SEG.E

----------------- 3-14
 kT 
You can specify the parameters: TRN.0, TRN.E, SEG.0 and SEG.E in the IMPURITY statement.
All parameters are specified for only one direction, which is from region 1 to region 2.
The following is an example of the syntax used to change the segregation coefficients
between oxide and silicon. Two material names separated by a / to indicate the combination
and the ordering of materials for which these parameters are specified.
IMPURITY I.PHOSPHORUS SILICON /OXIDE SEG.0=30 TRN.0=1.66E–7
Interface Trap Model (Dose Loss Model)
You can simulate the effect of dose loss at silicon/oxide interface by specifying the
DOSE.LOSS parameter in the METHOD statement. This model is based on the theory that the
dopant diffusing through silicon/oxide interface can be trapped into the trap sides located at
the interface [10], [11]. A modified equation for impurity flux is used in the Dose Loss
Model.
F m = S m  C m  1 – C T  C Tmax  – M m C T  3-15
where m=1,2 correspond to Si and SiO2, Mm = Cmss/CTmax, M12 = M1/M2, and S1 = 10S2.
The CT parameter is the real density of occupied trap sites at the interface and is found by
solving the following equation:
C T
= F1 + F2 3-16
t
where CTmax = 6.8  1014cm-2 for phosphorus and 21014cm-2 for other dopants. The dose
loss transport coefficient S1 is calculated through the following Arrhenius expression:
S 1 = TRNDL.0  exp  – -------------------------

TRNDL.E
3-17
 kT 
where the TRNDL.0 and TRNDL.E parameters can be specified in the IMPURITY statement.

3.1.4 The Two Dimensional Model

In this model, the point defect populations are directly represented and evolved in time. If
there is a super/supra saturation of point defects, it will affect the dopant diffusivity through a
simple scale factor, which goes to unity as the actual defect concentration approaches the
equilibrium defect concentration. Therefore with equilibrium defect profiles, the Two
Dimensional Model merely reduces to the Fermi Model, although in a more computational
inefficient manner, since solving for point defects is not required. The pair coupling between
defects and dopants in this model is assumed to be one-way. The diffusion of dopants is
highly influenced by the diffusion of point defects, while the diffusion of the point defects,
however, is regarded as totally independent of dopant diffusion. Stated in physical terms, this
corresponds to a pairing between defects and dopants with zero binding energy.
To turn on the Two Dimensional Model, specify parameter TWO.DIM in the METHOD
statement. The Two Dimensional Model is based on the Fermi Model, so read the Fermi
Model description before continuing. The major difference between the Fermi Model and the
Two Dimensional Model is the direct representation and evolution of non-equilibrium point
defect populations. Therefore, there are three different sets of governing diffusion equations:
one for dopants, one for point defect interstitials and one of point defect vacancies. In
addition, you also need to take into account the {311} cluster formation and dissolution, bulk
and interface recombination, and the generation of point defects through oxidation. Each of
these are described in detail in the following sections.
Dopants
The continuity equation for dopants is:
 C Ch
-----------
t
- = –
 J AX 3-18
X = I, V
 CX   C X  qE
J AX = – f X D AX   C A ---------
-  – Z A  C A ---------
- ------- 3-19
 C*   C *  kT
X X
where CI and CV are the actual concentrations of interstitials and vacancies, and C*I and C*V
are the equilibrium concentrations. The fX factor is an empirical defect factor, which for
interstitials is assumed to be temperature-dependent through the following Arrhenius
expression:
f I = FI.0 exp  – FI.E

----------- 3-20
 kT 
where the FI.0 and FI.E parameters can be specified in the IMPURITY statement. The value
of fI is clamped to a number between 0 and 1. The equivalent term for vacancies is calculated
according to:
fV =  1 – fI  3-21
The formulation of equation (Equations 3-18 and 3-19) is similar to the Fermi Diffusion
Model except for two elements. The additional term CX  C X * has been added to model the

enhancement or retardation of diffusion due to non-equilibrium point defect concentrations.

The term fX takes into account the knowledge that some impurities diffuse more by
interstitials than by vacancies or vice versa. Although this dependency is of a
phenomenological character, it seems reasonable, and is the one used by most diffusion
simulators to account for the diffusion enhancement of dopants during oxidation enhanced
diffusion (OED) or transient enhanced diffusion (TED).
Remember the point defect equilibrium concentrations are temperature as well as Fermi level
dependent and can be calculated from the following expressions:
n +1 n +2 n –1 n –2
neu + neg  ---- + dneg  ---- + pos  ---- + dpos  ----
n  n  n  n 
C X = C X i -------------------------------------------------------------------------------------------------------------------------------------------------------------------
i i i i - 3-22
neu + neg + dneg + pos + dpos
where CX*i represents the equilibrium defect concentrations of interstitials and vacancies
under intrinsic conditions, and the weight factors: neu, neg, dneg, pos, and dpos account for
the distribution of defects of different charge states under intrinsic conditions. All of these are
assumed to be temperature dependent through Arrhenius expressions of the following type:
neu = NEU.0 exp  – ------------------

NEU.E
3-23
kT
where the pre-exponential factors and activation energies, in this case NEU.0 and NEU.E, can
be specified in the VACANCY and INTERSTITIAL statements. Table 3-3 shows the complete set
of corresponding parameters.
For dopants, the boundary and interface conditions are identical to the ones stated in the
Fermi Model.
Table 3-3 Parameters for charge statistics and intrinsic point defect concentrations
Entity Pre-exponential factor Activation Energy
neu NEU.0 NEU.E
neg NEG.0 NEG.E
dneg DNEG.0 DNEG.E
pos POS.0 POS.E
dpos DPOS.0 DPOS.E
CSTAR.0 CSTAR.E
C X
i
Interstitials
The interstitial profile evolves with the following continuity equation.
 CI
-------- = – J I – R B + R T – R  311 3-24
t

where JI is the flux of interstitials, RB is the bulk recombination rate of interstitials, RT

accounts for the capture or emission of interstitials by traps, and R{311} is the recombination
rate of {311} clusters. Each of these terms are described below.
The interstitial flux, JI, is calculated according to [5] with:
*  CI 
– J I = D I C I   ------- 3-25
 C*I 
which correctly accounts for the effect of an electric field on the charged portion of the
CI
interstitials, taking the gradient of the normalized interstitial concentration, --------

. DI is the
CI
diffusivity of free interstitials. Don’t confuse it with the pair diffusivity DAI, which was
mentioned in Section 3.1.2 “The Fermi Model”.
DI is calculated once again with the following Arrhenius expression:
D I = D.0 exp  – D.E

---------- 3-26
 kT 
where the pre-exponential factor and activation energy (D.0 and D.E) can be set in the
INTERSTITIAL statement.
The bulk recombination rate (RB) is a simple reaction between vacancies and interstitials that
assumes that any interstitial will recombine with any vacancy, regardless of their charged
states. This assumption may overestimate the recombination rate. The equation is expressed
as:
*
R B = K r  C I C V – C I C V* 3-27
where Kr is the bulk combination coefficient and specified as:
K r = KR.0 exp  – KR.E

------------- 3-28
 kT 
where the parameters KR.0 and KR.E are user-definable in the INTERSTITIAL statement.
The interstitial trap rate (RT) model was first introduced by Griffin [12] to explain some of the
wide variety of diffusion coefficients extracted from different experimental conditions. The
Trap Equation, which describes the evolution of the empty trap population in time, is:
 C ET 
C I  C T – C ET 
e
- = – K T C ET C I – -----------------
R = ----------- 3-29
t 
1–e
where:
• CT is the total trap concentration.
• KT is the trap capture rate.
• CET is the empty trap concentration.
• CI is the interstitial concentration.


• CI is the equilibrium interstitial concentration.
• e* is the equilibrium empty trap to total trap ratio, e½ C ET

 C .
T
Both KT and e* are Arrhenius expressions that can be set in the TRAP statement with the total
trap concentration, RT, with the parameters shown in Table 3-4.
Table 3-4 Parameters for interstitial traps
KT KTRAP.0 KTRAP.E
e* FRAC.0 FRAC.E
CT TOTAL —
The trap equation is either derived from the simple reaction:

I + T  IT 3-30
or posed as a rate equation:
 C ET
- = – K T C I C ET + K r  C T – C ET 
----------- 3-31
t
where Kr is the trap emission rate. In equilibrium, the left hand side of Equation 3-31 must
vanish, giving:
C I  C ET  C 
K T C I C ET
 T e -
= K r C T – C ET  K r = K T ------------------------------------ = K T C I -------------
   3-32
   1 – e
1 –  C ET  C 
 T
Substituting this value for Kr into Equation 3-31 leads to the expression in Equation 3-32.
The recombination rate of {311} clusters (R{311}) in Equation 3-24 accounts for the release
rate of {311} interstitial clusters, which are small, rod-like defects that have been observed in
Transmission Electron Microscopy (TEM) studies after medium to high dose implantation of
impurities into silicon. Since a large fraction, if not all, of the excess interstitials after
implantation are believed to exist in this form, the time scale for dissolution of {311} clusters
plays an important role for the duration of any Transient Enhanced Diffusion (TED).
Think of these volume defects as small pockets of interstitials distributed throughout certain
parts of the doped regions, which are released during annealing; thus, acting as bulk sources
of point defects.

Note: Actually, {311} defects are believed to be created from excess free interstitials during the earliest part of the
annealing cycle through a process called Ostwald ripening. But here, they’re considered as existing
immediately after the implantation.
The cluster release rate follows a simple exponential decay in time specified by:
C 311 
- = f  x   --- exp  – --
1 t
R  311  – ----------------- 3-33
t    
where f(x) is the as-implanted profile of {311} clusters, and  is an Arrhenius type
temperature dependent time constant calculated from:
 = TAU.311.0 exp  – TAU.311.E

-----------------------------

3-34
kT
where the TAU.311.0 and TAU.311.E parameters can be specified in the CLUSTER statement.
The profile, f(x), of the {311} clusters is created from a previous IMPLANT statement. For
more information, see Section 3.5.5 “Ion Implantation Damage”.
To activate this model, a previous IMPLANT statement has to introduce {311} clusters with the
CLUSTER.DAM flag in the METHOD statement. For example:
METHOD CLUSTER.DAM
CLUSTER BORON MIN.CLUSTER=1.0E17 MAX.CLUSTER=1.0E19 CLUST.FAC=1.4\
TAU.311.0=8.33e-16 TAU.311.E=-3.6 SILICON
IMPLANT BORON ENERGY=100 DOSE=1E15
Here, the METHOD statement switches the model on, and the CLUSTER (optional) statement
decides the location and scaling of the {311} cluster profile. In this example, clusters are
present in the regions of the substrate, where the chemical boron concentration is between
1.0e17 cm–3 and 1.0e19 cm–3, which are scaled by a factor of 1.4 relative to the boron
concentration.
Notice that the activation energy for TAU.311.E must be specified as negative, since the time
constant decreases with rising temperature.
Interstitial Generation and Recombination at Interfaces
Interfaces present a moving boundary problem during a thermal oxidation. In this instance,
there will be a recombination rate at the interface, which will vary as a function of the
interface velocity. Also as a consequence of the silicon being consumed, there is a significant
injection of interstitials into the substrate. Within Athena, this is modelled by an interstitial
flux boundary condition, as described by Hu [13].
 n C I + K s  C I – C I  = G I 3-35
where:
• nCI is the projection of the interstitial flux vector on an inward pointing unit vector
normal to the boundary.
• KS is the effective surface recombination rate for interstitials.

• GI is the generation rate at the interface of interstitials during annealing in an oxidizing

ambient.
In other words, Equation 3-35 shows that the number of interstitials generated on the surface,
minus the number of interstitials that recombine there, must equal the number of interstitials
that diffuse from the surface/interface into the substrate.
The effective surface recombination rate, KS, depends on the motion of the interface during
oxidation according to:
K
 v i  POW
K S = K SURF K RAT  ------------------ +1 3-36
 v i – max
where vi is the velocity of the interface, vi-max is the maximum interface velocity, and the
parameters KSURF, KRAT, and KPOW are calculated according to the following equations:
K SURF = KSURF.0 exp  – -------------------------

KSURF.E
3-37
 kT 
K RAT = KRAT.0 exp  – ---------------------

KRAT.E
3-38
 kT 
K POW = KPOW.0 exp  – -----------------------

KPOW.E
3-39
 kT 
where the pre-exponential factor and exponent terms can be defined in the INTERSTITIAL
and VACANCY statements.
Surface recombination plays an important role in the relaxation of perturbed point defect
profiles back to their equilibrium values, which cannot happen by bulk recombination alone.
The surface generation rate, GI, controls the injection of point defects into the silicon during
oxidation.
Two models have been implemented into Athena: the default model, GROWTH.INJ and
TIME.INJ. The moving interface can inject point defects into silicon and polysilicon.
The GROWTH.INJ parameter in the VACANCY or INTERSTITIAL statement will activate or
deactivate the growth dependent injection model. By default, this model is always turned on
and is described mathematically by the following equation.
G
 v i  POW
G I =   VMOLE  v i   ------------------ 3-40
 v i – max
where:
•  is the fraction of silicon atoms consumed during growth that are injected into the bulk as
self interstitials.
• VMOLE is the lattice density of the consumed material.
• GPOW is a power parameter.
The values  and GPOW are calculated from the following equations:

 = THETA.0 exp  – THETA.E

-------------------------

kT
3-41
G POW = GPOW.0 exp  – ------------------------

GPOW.E
3-42
 kT 
and the THETA.0, THETA.E, GPOW.0, GPOW.E, and VMOLE parameters can be specified in the
INTERSTITIAL and VACANCY statements.
As a general rule, the ratio    K RAT  K SURF  should be maintained reasonably constant during
calibration. The entities, vi and vi–max have the same meaning as described for surface
recombination. The maximum interface velocity, vi–max, cannot be manipulated directly and
will change only when oxidation characteristics change.
The TIME.INJ parameter in the VACANCY or INTERSTITIAL statement activates the time
dependent injection model. It is defined as:
T POW
G IV = A  t + t 0  3-43
where t is the total diffusion time in seconds, and A, t0 and Tpow are free parameters used for
calibration purposes and are calculated according to the following equations:
A = A.0 exp  – A.E

--------- 3-44
 kT 
t 0 = T0.0 exp  – T0.E

------------ 3-45
 kT 
T POW = TPOW.0 exp  – TPOW.E

----------------------- 3-46
 kT 
Table 3-5 shows all user-specifiable model parameters for point defect boundary and injection
conditions.
Table 3-5 Parameters for specifying point defect boundary and injection conditions
KSURF KSURF.0 KSURF.E
KRAT KRAT.0 KRAT.E
KPOW KPOW.0 KPOW.E
A A.0 A.E
t0 t0.0 t0.E

Table 3-5 Parameters for specifying point defect boundary and injection conditions
TPOW TPOW.0 TPOW.E
 THETA.0 THETA.E
VMOLE VMOLE —
GPOW GPOW.0 GPOW.E
KSURF KSURF.0 KSURF.E
Vacancies
The diffusion and flux equations for vacancies are largely similar to the interstitial equations
described above.
C
---------V- = –   J V – R B 3-47
t
Here, JV is the flux of vacancies and RB is the bulk recombination rate.

The Vacancy Flux Expression is:
 CV 
– J V = D V C V   ---------
- 3-48
C  
V
which correctly accounts for the effect of an electric field on the charged portion of the
V C
vacancies by taking the gradient of the normalized concentration --------

. The term DV is the
CV
diffusivity of free vacancies, not to be confused with the pair diffusivity DAV, which was
mentioned in Section 3.1.2 “The Fermi Model”. The vacancy diffusivity is set according to
the following equation:
D V = D.0 exp  – D.E

---------- 3-49
 kT 
where the D.0 and D.E parameters are set on the VACANCY statement.
The bulk recombination rate, RB, is a simple reaction between vacancies and interstitials that
assumes that any interstitial is equally likely to recombine with any vacancy, regardless of
their charged states. This assumption may overestimate the recombination rate but is a
commonly applied assumption. The equation is expressed as:
*
R B = K r  C I C V – C I C*V  3-50
where Kr is the bulk combination coefficient and is specified as:

K r = KR.0 exp  – -------------

KR.E
3-51
 kT 
where the parameters KR.0 and KR.E are user-definable in the INTERSTITIAL statement.
This is the same equation for bulk recombination as described earlier for interstitial bulk
recombination.
The boundary/interface conditions for vacancies are set similarly to those for interstitials
except that the VACANCY statements should be used instead of all INTERSTITIAL statements.
Dislocation Loop Based Enhanced Bulk Recombination
A topic of some debate in recent literature has been the creation of Dislocation Loops.
Currently, the exact physical nature of these defects is still under investigation. It is believed
that they arise from amorphizing implants and only exist at the edge of the amorphous layer.
It has been suggested that these loops grow through the absorption of interstitials during
oxidation and perhaps shrink by the emission of interstitials when annealed in nitrogen.
Due to the lack of proper scientific description, only a simple recombination model has been
implemented into Athena. This model introduces an additional sink of interstitials that is
described by the following expression:
R loop = damalpha  C I – C I   3-52

 
where damalpha is a parameter you can set on the INTERSTITIAL statement. This additional
recombination is only applied to a region of the amorphizing implant controlled by the
parameters MIN.LOOP and MAX.LOOP in the DISLOC.LOOP statement.
As an example, the following statements will produce a region of dislocation loops where the
as-implanted phosphorus concentration is between 1e16 cm-3 and 1e18 cm-3. The damalpha
parameter is then set to 1e8 in this region. For example:
DISLOC.LOOP MIN.LOOP=1e16 MAX.LOOP=1e18 PHOSPHORUS
INTERSTITIAL SILICON DAMALPHA=1e8
IMPLANT PHOSPHORUS DOSE=1e15 ENERGY=120
This model should only be chosen when either the Two Dimensional or Fully Coupled Model
is also used. The two things to apply to this model are as follows:
• an implant has to create an amorphous layer
• immediately after the implant there is an anneal in a wet ambient
Even when these two criteria are met it is suggested to only apply this model when it is
needed to match experimental results.
The Steady State Diffusion Model
The Steady State Diffusion Model is a variant of the two dimensional diffusion model, which
assumes the point defect profiles are in a steady state. It is turned on with the METHOD
STEADY command.

Important Notes about Defect Diffusion

Point defects have larger diffusivities than dopants and may therefore diffuse down to the
bottom of the structure during a simulation. If the simulation structure is too shallow, you
may get an unrealistic high defect concentration in the regions where dopant profiles are
present and consequently too much dopant diffusion. Therefore, you may need to extend the
depth of the simulation space to provide an adequate sink for the point defects. To determine
the depth of the structure, you can estimate the characteristic defect diffusion lengths using:
l = DX  t 3-53
where DX is the defect diffusivity and t is the total diffusion time. Simulations show that a
depth of 20 to 50 microns is required in most cases. This restriction on the minimum structure
depth poses a threat to computational efficiency, whenever diffusion models that include
point defects are employed. But since the fine structure of the defect profiles near the bottom
of the structure is not a feature of particular interest for processing purposes, you can reduce
the computational cost by making the grid very coarse in this region.
Time Step Control
When using diffusion models that include the explicit representation and evolution of point
defects, be aware of time stepping issues. Although step size control between iterations is
fully automated, you can still specify the size of the initial time step. Different initial time
step sizes can be specified for dopants and point defects, respectively, by using the parameters
INIT.TIME and PDINIT.TIME. For example, the command
METHOD INIT.TIME=0.001 PDINIT.TIME=0.001
would set the initial time step to 1 millisecond for both dopants and point defects. Default
values are INIT.TIME=0.1 seconds and PDINIT.TIME=1.0E-5 seconds.
Note: There is no guarantee that the program will actually use these values. For this initial time step, the only
purpose of these parameters is to make it feasible for you to give the program a hint about an appropriate
initial time step size.
3.1.5 The Fully Coupled Model

The Fully Coupled Diffusion Model is identical to the Two Dimensional Model. Be familar
with that model before reading any further. The one important difference is that the diffusion
of the defects is now influenced by the diffusion of the dopants by the addition of the joint
pair fluxes to the flux terms in the governing equation of the defects. Therefore, there is a true
two-way interaction between the diffusion of dopants and the diffusion of point defects,
which gives this model its name. The fully coupled model is slightly more CPU-intensive
than the two dimensional model, but encompasses the capability of reproducing certain
important aspects of semiconductor processing such as the Emitter Push Effect in the case of
phosphorus diffusion.
From a physical viewpoint, however, this original fully coupled model suffers from the
shortcoming of not explicitly representing pairs, and the consequential lack of a subdivision
of defects and dopants into paired and non-paired fractions. Therefore, this model cannot
reproduce the saturation of the dopant diffusivity that is believed to occur at very high
damage concentration due to a total pairing of dopants. In other words, the model relies on the

dilute approximation (i.e., the assumption that the concentration of pairs is much smaller than
both the dopant and the defect concentrations). To use the Fully Coupled Model, specify
parameter FULL.CPL in the METHOD statement.
The Fully Coupled Model establishes a two-way coupling between the diffusion of dopants
and point defects respectively by adding the joint dopant-defect pair fluxes to the flux terms
of the defect equations, which then become:
   
 C + C c = –    J V + J c – R
----
t

V  AV



 AV



3-54
A ,c A ,c
   
 C +  = –   J + 
----
t  I

C
AI
c

 I

J
AI

c  – R B + R T – R  311 

3-55
A ,c A ,c
where summations run over all dopants and pair charge states. The rest of the Fully Coupled
Model Equations are identical to those in the Two Dimensional Model described in the
previous section. The effect of the correction terms only displays itself at very high dopant or
high implant damage conditions or both, where the Fermi level enhancement and point defect
supersaturation will increase the dopant diffusivities significantly.
High Concentration Extension to the Fully-Coupled Model
This extension to the fully coupled model takes into account additional higher order defect-
dopant/defect pairing the extra point defect recombination mechanisms. This model was
developed at Stanford University [14] to include higher order dopant-defect interactions in
the cases, where the number of dopant-defect pairs are significant. This is the case for high
dopant concentration in silicon. It is activated by the following command:
METHOD HIGH.CONC FULL.CPL
It is an extension of the basic fully coupled model and may only be used in conjunction with
the METHOD FULL.CPL command. This model includes two extra bulk recombination
reactions and two extra Si/SiO2 interface recombination reactions.
In the bulk, extra terms for point defect recombination apply at high concentrations, where
statistically, a high level of dopant-defect pairing is prevalent
D PI + D V
= IIFACTOR   ----------------------
PI – V
Kr - K 3-56
 DI + DV  r
D I + D PV
= IVFACTOR   ----------------------
I – PV
Kr - K 3-57
 DI + DV  r
The extra model parameters calibrate the ratio of effective capture cross sections of dopant-
defect/defect to defect/defect recombination mechanisms. The IIFACTOR and IVFACTOR
parameters can be set in the INTERSTITAL statement as follows:
INTERSTITIAL SILICON IVFACTOR=<n> IIFACTOR=<n>

RTA Diffusion Modelling

SSuprem4 has the capability to model rapid thermal annealing (RTA) processes within the
framework of existing diffusion models (i.e., the two dimensional model and the fully
coupled model). Since RTA is a short time thermal cycle involving steep temperature ramping
to high temperatures, Transient enhanced diffusion (TED) will dominate whenever a
significant amount of lattice damage is prevalent. Because the amount of dopant diffusion is
intimately coupled to the evolution of the point defect populations, you can calibrate these
models to RTA conditions by tuning the point defect related parameters. The ratio of
interstitial damage in the form of {311} clusters to that in the form of free interstitials and the
characteristic time for dissolution of interstitial clusters are important parameters to include
when setting up SSPUREM4 for an RTA scenario.
Table 3-6 shows an approximate time for completion of about 95% of the TED at various
temperatures.
Table 3-6 The approximate duration of TED at various annealing temperatures
Annealing Temperature (C) Time for completion of 95% of TED
600 390 hours

700 3.3 hours
750 30 minutes
800 3.7 minutes
850 43 seconds
900 8.3 seconds
950 1.9 seconds
1000 0.48 seconds
1050 0.13 seconds

3.1.6 Electrical Deactivation and Clustering Models

When dopants are present at high concentrations, the electrically active (mobile)
concentration, Cact, may be less than the corresponding chemical concentration, Cchem.
In order for an impurity to become electrically active in a semiconductor material, it must be
incorporated into a substitutional lattice site, which then will contribute with a carrier to either
the valence band (an acceptor impurity) or the conduction band (a donor impurity). Above
certain dopant concentrations, however, it is impossible to incorporate more dopants into
substitutional lattice sites. The excess dopants are said to be non-active.
The threshold where the deactivation occurs is often called the solid solubility limit, since
impurities can exist in different phases in the crystal. But for this section, we’ll call it
deactivation threshold. Therefore, it isn’t well-defined which phase transition the solid
solubility limit might refer to. For example, excess dopants could be participating in small
clusters or larger precipitates. Deactivation threshold would be a more proper designation for
this limit and will be used throughout the rest of this section. The notation, C th
act will be used
for the deactivation threshold. Therefore, for all the models described in this section, the
following points are assumed for each dopant type:
• Dopants in excess of the deactivation threshold are considered electrically inactive (i.e.,
they do not contribute to the carrier populations).
• Additionally, dopants in excess of the deactivation threshold are considered to be
immobile (i.e., they cannot diffuse).
Electrical Activation Model
The Electrical Activation Model is used to calculate which level of dopant concentration the
deactivation occurs. For this purpose, two different Electrical Activation Models are used for
all diffusion models:
• An AsV Clustering Model for arsenic or similar models for other impurities.
• A Semi-empirical Activation Model, based on Table B-13 in Appendix B “Default
Coefficients”, for all other dopants.
These models can be activated in the IMPURITY statement by parameters CLUSTER.ACT and
SOL.SOLUB respectively. By default, the CLUSTER.ACT model is used only for As in silicon
and polysilicon.
The AsV Clustering Model used in SSuprem4 is based on the simple reaction:
+ = –
As + V  AsV 3-58
Assuming that Equation 3-58 is always in equilibrium, the following equation describes the
relationship between the chemical and the active arsenic concentration.
 2 C 
C chem = C act + C clust = C act  1 + C TN  ----  ------------
2 n V
 n  C 
- 3-59
 i V
Since there is a cyclic dependency between the active arsenic concentration, Cact, and the
carrier concentration, n, an initial guess for the value of n has to be made at the start of the
simulation. In addition, C V  C V is (by definition) set to unity when running the Fermi
Diffusion Model (see Section 3.1.2 “The Fermi Model”).

The clustering coefficient, CTN, is temperature dependent according to the following

equation.
C TN = CTN.0 exp  – -----------------  n i

CTN.E
3-60
 kT 
Here, the CTN.0 and CTN.E parameters can be defined in the IMPURITY statement.
You can specify CLUSTER.ACT, CTN.0 and CTN.E parameters for other acceptors in the
IMPURITY statement. But be aware the model isn’t elaborated for other impurities and these
parameters are unknown. The model can also be empirically used for acceptors (e.g., indium).
The following equation is based on acceptor-interstitial clusters with empirical parameters
CTP.0 and CTP.E.
 2 p 2 C 
C chem = C act  1 + C TP  ---- -------I-  3-61
 n  C
 i I 
The Semi-empirical Activation Model, based on Table B-13 in Appendix B “Default

Coefficients”, which is used for all other dopants except arsenic, uses a two-step scheme to
calculate the active dopant concentration.
1. The program interpolates into a table of experimental (e.g., temperature and deactivation
th
threshold) data pairs and finds a concentration independent deactivation threshold, C act ,
that corresponds to the current simulation temperature. You can set these pairs in the
IMPURITY statement by assigning values to the parameters, SS.TEMP and SS.CONC. The
temperature should be specified in Celsius.
2. A logarithmic concentration dependency is incorporated by setting the final deactivation
threshold to the value:

 th
 th C act  C act – b th
th  C  1 +  1 – b  ln ------------------------------------- ; C act > C act
C act =  act 1–b 3-62

 th
 C act ; C act  C act

where the parameter b must be in the range of [0.8, 1.0]. Parameter b can be specified as
ACT.FACTOR in the IMPURITY statement. The effect of Equation 3-62 is to produce a
rounding in the top of the active profile that slightly follows the form of the chemical profile.

Transient Activation Model

The Transient Activation Model assumes that dopants, after an implant, are inactive. A
certain time is required before the dopants become active. After an ion is implanted into
silicon, this model assumes that all dopants are inactive and may not be activated immediately
but become gradually active instead. The Transient Activation Model simulates this behavior
and applies it to activating the implant dopants.
The following equation for the active concentration CA is solved.
eq
 CC – CA
 C Chem – C A  = --------------------
- 3-63
t A
CC is the chemical concentration of the dopant, and CAeq is the equilibrium active
concentration calculated either from solid solubility or clustering model (for arsenic) as
defined in the previous section. The parameter  is the time constant for activation, which is
a function of temperature, and is calculated using the following Arrhenius expression.
 A = TRACT.0  exp  – TRACT.E

------------------------- 

3-64
kT
The initial condition at time t=0 for Equation 3-63 is specified by

eq
CA = min  C A , TRACT.MIN  n i  3-65
t=0
where ni is the intrinsic carrier concentration. Therefore, implantation activation will occur
immediately up to a level of TRACT.MIN  ni, after the active concentration is calculated
according to Equations 3-59 and 3-62.
To activate the transient activation model, set the CLUST.TRANS parameter in the METHOD
statement and specify the TRACT.0, TRACT.E, and TRACT.MIN parameters in the IMPURITY
statement.
The defaults for B, P, As and Sb are TRACT.0=8e-16sec, TRACT.E=-4.2, and
TRACT.MIN=1.0 (TRACT.MIN for Phosphorus is 2.0).

3.1.7 Grain-based Polysilicon Diffusion Model

The mechanism for impurity diffusion in polysilicon is different than that of crystalline
silicon. Polysilicon has a micro-structure of small (compared to the interesting device
regions) crystalline regions called grains. These grains are separated by grain boundaries
which occupy a certain spatial volume and are connected to form a complex network. The
texture and morphology of the grain structure depends on the deposition conditions and on
subsequent thermal treatment (during which recrystallization can occur). Impurities inside the
grain will diffuse differently than those in the grain boundaries. Dopant will also transport
through grain and grain boundary interfaces.
A model for impurity diffusion in polysilicon outlined in [18], [19], and [20] is incorporated
in Athena. To use polysilicon diffusion, specify the POLY.DIFF parameter in the METHOD
statement. Most of parameters, which control the model, have a prefix PD and are specified in
the IMPURITY statement. In this model, the concentration of each impurity Ci is split into two
g
components, namely the concentration within grain interior C i and concentration in the grain
gb
boundary C i . The impurity diffusion within grain interior is simulated by the standard
model used for crystalline silicon (see Equations 3-2 to 3-7):
g
Ci
=   Di C i – ZD i C i 1--- n + G i
g g g g
----------- 3-66
t  n 
g
where diffusivity of impurity i within grains D i is calculated exactly as in Equations 3-10
and 3-11.
The diffusion in the grain boundary is assumed to be constant and very rapid:
gb
C i 2 gb gb
- =   Di Ci  – Gi
----------- 3-67
t
gb
where D i is diffusivity of impurity i along the grain boundaries:
= PD.DIX.0  exp  – – ------------------------------ 

gb PD.DIX.E
Di 3-68
 kT 
The PD.DIX.0 and PD.DIX.E parameters are specified in the IMPURITY statement. The last
term Gi in Equations 3-66 and 3-67 controls impurity segregation between grain interior and
grain boundaries:
g
 Ci gb – 1
G i =  ---------- – Ci   3-69
 p seg 
where pseg is segregation coefficient and is the rate of segregation specified as PD.TAU in the
IMPURITY statement. The segregation process brings the ratio between grain interior and
g gb
grain boundary concentration to a steady state condition where C i  C i = P seg . Initial
conditions for Equations 3-66 and 3-67 are determined by setting the PD.CRATIO parameter

in the IMPURITY statement. This parameter specifies the initial ratio between impurity
gb gb g
concentration in the grain boundary C i and total concentration C i + C i .
The grain boundary segregation Pseg is calculated according the model suggested in [20]
AQ s Q 0 –1
- exp  ------
P seg = ---------------------- 3-70
L g  t N Si  kT 
where Qs is the density of segregation sites at the grain boundary specified by the
PD.SEGSITES parameter in the IMPURITY statement, NSi is the atomic density of crystalline
silicon (2.5 1022 atoms/cm3), A is the entropy factor specified by the PD.EFACT parameter in
the IMPURITY statement, Q0 is the segregation activation energy specified by the PD.SEG.E
parameter in the IMPURITY statement, and Lg(t) is the time-dependent grain size calculated
according to the grain growth model suggested in [19]:
L 0 + ------------------------------------------  exp  – – -------------------------------------------   t

2 PD.GROWTH.0 PD.GROWTH.E
Lg  t  = 3-71
kT  kT 
where L0 is the initial polysilicon grain size, which should be specified by GR.SIZE
parameter in the DEPOSIT POLYSILICON statement (GR.SIZE.F parameter allows to have
linearly graded grain size within deposited polysilicon layer), the grain growth parameters
PD.GROWTH.0 and PD.GROWTH.E are specified in the IMPURITY statement, and t is the
elapsed diffusion time.
The segregation boundary condition at the polysilicon/silicon boundary is also modified
when POLY.DIFF model is used. The default segregation coefficient M12 in Equation 3-12 is
increased by a factor, which depends on the impurity concentration in the grain boundaries:
 gb 
Ci
M12 = M 12 1 + PD.SEG.GBSI  ------------------------
  3-72
 gb g
 Ci + C 
where M*12 is modified segregation coefficient and the PD.SEG.GBSI parameter is specified
in the IMPURITY statement.

Advanced Diffusion Models SSuprem4 Models
3.2 Advanced Diffusion Models

The ultimate goal of TCAD simulation is to compute the electrical characteristics of a given
device by using only process-related data as input parameters. Since the electrical
characteristics of the device are heavily dependent on the distribution of the electrically active
impurity (dopants), resulting from the entire thermal processing sequence, it is important that
the diffusion models used in the process simulation are as accurate as possible. This is
particularly important for deep sub-micron processes. Therefore for these emerging
technologies, 2D or even 3D phenomena are expected to be of growing importance, whereas
there is presently no accurate technique to measure multi-dimensional dopant profiles.
Consequently, the active dopant 2D distributions can only be obtained by simulation based on
models that are as reliable as possible. It has become clear that the "abnormal" behaviors of
dopant diffusion in silicon are caused by non-equilibrium point defects. These are induced by
the diffusion process itself (emitter push effect caused by high concentration of phosphorus
diffusion), or injected into the substrate by external treatments, such as oxidation or
silicidation. Otherwise, they result from the ion implantation used to introduce the dopants
into the silicon substrate. With the necessary decrease of the thermal budget due to the
shrinkage of the device dimensions, these transient phenomena become key issues for
accurate dopant diffusion simulation.
Figure 3-1 The model consists of three parts: the classical dopant diffusion model, the interstitials clusters
model and the model of mixed dopant clustering.
The new model of dopant diffusion implemented in Athena is called PLS and was developed
in collaboration with CNRS-Phase (Strasbourg, France), CEA-LETI (Grenoble, France) and
SILVACO. This model is up to date with actual physical models and contains only physical
parameters [21]. It consists of three parts: the classical dopant diffusion (CDD) model, the
interstitials clusters (IC) model and the model of mixed dopant-defect clustering (DDC). This
section describes the three parts of the model and how the new model differs from the fully
coupled model.

The main physical points taken into account in the models are the following:
• Dopant diffusion of all species is assisted by both vacancies (V) and self-interstitials (I).
These point-defects exist in various charge states and their relative concentrations depend
on the local Fermi level position (i.e. on the local dopant concentration).
• Both I and V have strong binding energies with the dopant atoms and consequently the
diffusing species are dopant-defect pairs (the isolated substitutional dopants are
immobile). These impurity-defect pairs in their various charge states are not assumed to
be in local equilibrium with the free substitutional dopant atoms and the free defects. In
the PLS model, at high dopant concentrations, the concentrations of these pairs are not
considered to be negligible with respect to the substitutional (active) dopant
concentration. Therefore, the pair concentrations are explicitly taken into account to
compute the total dopant concentration and the Fermi level position (i.e., carrier
concentration). Consequently, a partial self-compensation takes place at high doping
concentrations, which contributes to the differences between total and active
concentrations and affects the variations of the extrinsic diffusivities as a function of the
total doping.
• The flux of each diffusing species (dopant-defect pairs and free defects) includes the drift
terms caused by the built-in electric field, due to the dopant gradients.
• Both I and V are not considered to be in local equilibrium but can be annihilated by
bimolecular recombination. This feature of the CDD model performs annihilations
between not only the free defects but also between the impurity defects pairs, which play
the role of recombination centers. Therefore, the I-V recombination rate is strongly
enhanced at high dopant concentration.
• It is now well established that transient enhanced diffusion (TED) is strongly correlated
with the evolution of the self-interstitial supersaturation governed by the nucleation and
evolution during the high temperature anneal of a variety of extended defect structures,
such as the interstitials clusters. Thus, the predictive process modeling of the deep
submicron MOSFET technologies requires the development of accurate diffusion models
which take into account the full set of interactions between dopants and the point or
extended defects (clusters). The PLS model, coupled with the BCA implantation, allows
you to calculate the evolution of the clusters. Therefore, it is unneccessary to artificially
add {311} or other clusters because the model automatically generates them and
calculates their evolutions according to the Ostwald ripening theory.
• At concentrations near the solid solubility limit, a dynamic clustering model is
considered. Immobile complexes (AsnV or BnIm) are formed, which result in decrease of
the effective diffusivity and increase of the inactive dopant concentration. These
complexes are not assumed to be in local equilibrium with the other species.
• When the dopant concentration exceeds a few 1020cm-3, the dopant-vacancy pairs can no
longer be considered as isolated entities because the vacancies can interact with more
than one dopant atom.

3.2.1 Classical Model of Dopant Diffusion (CDD)

The basic idea of the model is isolated substitutional dopant atoms (As) are immobile. The
dopant diffusion occurs only through the migration of dopant--self-interstitial (AI) and of
dopant--vacancy (AV) pairs. Moreover, in this enhanced model, local equilibrium is not
assumed between the pairs and their components (unlike the original CNET model [5]). All
possible charge states of the free defects and of the pairs have been considered and their
relatives concentrations depending on the local Fermi level position.
To turn on the CDD model, specify PLS parameter in the METHOD statement. All physical
parameters of the model can be modified in the *.mod files. To specify the location of these
files, use the B.MOD, P.MOD, AS.MOD, IC.MOD, and VI.MOD parameters in the DIFFUSE
statement. By default all these files are located in the $SILVACO/lib/athena/
<version_number>/common/pls directory.
Charge States
Point Defects
The result of diffusion studies in metals and ionic crystals have led to the establishment of
several basic atomic diffusion mechanisms. These mechanisms dominate the interpretation of
silicon diffusion experiments with the exception that in silicon there is a very wide energy
range available to the Fermi level. Therefore, a given lattice defect can appear in a variety of
ionized states. The fundamental principles of thermodynamic predict that such defects will
exist in equilibrium at all temperatures above 0° K, because the presence of such defects
0 0
minimizes the free energy of the crystal. The entities Veq and Ieq are the equilibrium
defect concentrations for vacancies and silicon self-interstitials in their neutral charge state.
The weight factors  and  account for the different charge states for distribution of point
defects under extrinsic conditions. Each  and  is assumed to be temperature dependent
through Arrhenius expressions.
For point defects (V or I), five various charge states are considered:
s  s   n i s 0 s s n s

I =  ---- I , V =   ----i V 0
 n  n
3-73
where s is one of the charge states -2, -1, 0,+1, or +2. All parameters  and are specified in
the charge state statements of the defect.mod file.
With this consideration, the equilibrium concentrations Veq and Ieq are estimated as
 s   n i s  s   n i s
 
eq 0 eq
I = I eq  ---- , V = V 0  ---- . 3-74
 n eq  n
Under intrinsic conditions (i.e., n=p=ni), the equilibrium concentration can be simply written
as:
s s

eq eq

0 0
I i = I eq , Vi = V eq . 3-75

Equilibrium concentration for vacancies and silicon self-interstitials are defined as a simple
Arrhenius functions:
V
E
I
 Ef 
I i = KI exp  – -----f- , V i = K V exp  – --------
eq 0 eq 0
3-76
 kT  kT 
I V 0 0
where E f and E f represent respectively the formation energy for I and V, and K I , K V are
coefficients. These parameters are specified in the defect.mod file.
Dopant-defect Pairs
Using the same assumption as for point defects, the dopant-defect pairs are defined by the
charge states of -1, 0, +1.
For boron pairs we have
K  s----------
– 1- s+1
 2 
 ------------ 
2
s 0 s
-  ---------------  n
s
=  -----------
p - BI
BI BI , p BI = -------------------- 3-77
 p s    s-----------
+ 3- i,
BI K  s-----------
+ 1-  2 
BI
 2  
K  s----------
– 1-  s-----------
+ 1-
 2   2 
s p  s 0 s 
=  ------------- -  -----------------  n i,
BV
BV BV , p BV = --------------------- 3-78
 p s    s-----------
+ 3-
BV K  s-----------
+ 1-
 2 
 2 
BV 
The BV pairs exist, however, with positive (+1) and neutral (0) charge states only. Moreover,
it is well established BI pairs exhibit a negative-U behavior and neutral state is unstable.
Therefore, the three charge states BI+, BI0 and BI- are considered.
For donor impurities, the three charge states for dopant-interstitial pairs are also considered.
Though only AV- and AV0 are known to exist (E-centers).
K  s-----------
+ 1-  s----------
– 1-
 2   2 
s  n  s 0 s
AI =  ------------ AI ,
AI
-  --------------  n
n AI = -------------------- 3-79
s i,
 n AI  K  s----------
– 1-  s----------
 2 
– 3-
AI
 2  
K  s-----------
+ 1-  s----------
– 1-
 2   2 
s  n  s 0 s
AV =  ------------- AV ,
AV
-  ---------------  n
n AV = --------------------- 3-80
s i,
 n AV  K  s----------– 1-  s----------
 2 
– 3-
 2  
AV

The KAI(s) and KAV(s) parameters for both donors and acceptors represent the pairing
coefficients for the dopant-interstitial and dopant-vacancy pairs and are defined by following
formula:
E AI  s   E AV  s 
= KAI s  exp  – -----------
0 0
K s - , K s =K s exp  – -------------
- 3-81
AI  kT  AV AV
 kT 
Fully Coupled Equations

The point defects, dopant-defect pairs and active dopant evolve according to the following
continuity equations:
tot
 I - = –   J + GR + GR + GR
----------- AV – I,
t I IV AI
tot
-------------
V - = –   J + GR + GR + GR
AI – V,
t V IV AV
tot
---------------
AI - = –   J – GR + GR 3-82
AI – V,
t AI AI
tot
 AV - = –   J – GR + GR
----------------- AV – I,
t AV AV
A
--------s- = GR AI + GR AV – GR AV – I – GR AI – V
t
Here, Xtot is the total concentration of the species X, and [As] is active dopant concentration in
substitutional position. J represents the flux of species X, and GRX is a generation-
recombination term corresponding to the reactions that contain X.
Flux Equations
Point Defects
As the migration rate of a given species may depend on its charge state, two distributions
must take into account in the flux equation. The first one is the Fickian term and the second is
the Nerstian term. Thus, the flux can be written for a given species X with the charge state s:
= – D s   X – sX 
s s p
J ------- . 3-83
X  p
s
X
For point defects, the PLS model makes the assumption that the diffusivity is independent of
the charge state at high temperature. This implies that for vacancy and silicon self-interstitial
at any charge state s the following is always true.
DI = D s , DV = D s 3-84
I V
Moreover, it is assumed that diffusivities at high temperature follow the simple Arrhenius
law:

m m
D I = DI exp  – -------
EI 
D V = D exp  – --------
0 0 EV 
- , - 3-85
 kT  V  kT 
m m
where E I and E V represent respectively the migration energy for self-interstitials and
0 0 m m
vacancies. Parameters D I , D V , E I , and E V are defined in the defect.mod file.
2
Finally, taking into account that np = n i and
–1 –1
s s n s s s n s
   ----    ----
tot tot
I = I , V = V , 3-86
ni ni
s s
the flux equations for Itot and Vtot are defined as follows:
 s   n i s
 s ----
 n
J -
tot tot s
J tot = s = – D I DI – I --------------------------------- ------n- 3-87
I I
s i n s n
s
  ----
n
s
 s   n i s
 s ----
 n

J
tot tot s
J tot = s = – D V DV – V ---------------------------------- ------n- 3-88
V V
s i n s n
s   ----
 n
s
Dopant--Defects Pairs
V I
The diffusion of dopant--defect pairs implies on the fact that diffusivities D A and D A are
calculated from the basic parameters for the pairs defined by Equation 3-77 through 3-81 as
follows:
D D
I eq V eq
D A = f I D Ai = s AI , D A =  1 – f I D Ai = s AV , 3-89
AI AV
s s
where fI is the interstitialcy component under intrinsic conditions, and DAi is the intrinsic
diffusivity of the impurity A. The fI and DAi parameters are defined for each dopant in the
corresponding dopant.mod file.
It should be emphasized that several relationships exist between this various parameters.
Therefore, you can decrease the number of free fitting parameters. From physical point of
view, it can be safely assumed the ratio between the diffusivities of the various charge states
must be equal. Consequently, we have
D – D 0 D – D 0
s s
AI
------------ AI
- =   AI  ,
- = ----------------- AV
------------- AV
- =   AV  ,
- = ------------------ 3-90
D 0 D + D 0 D +
AI AI AV AV

where s=1 for donors and s= -1 for acceptors. These diffusivities are free parameters when
the various coupling parameters in Equation 3-81 are known. Therefore, you can calculate
them through the experimentally known intrinsic diffusivities DAi and the interstitialcy
component fI. The  AI and  AV parameters are defined as an Arrhenius functions.
Finally, the flux equations, for instance, for boron--point defect pairs are defined as follows:
p- s 1  s 1 
 ---

s
 BI  K 
BI
s
 n i

J tot = s
J s = – D 0 ------------------------------------------------------------- - 3-91
BI BI BI s1  s1 
 

p
s K s ---- 
BI  n 
i
s
 p s1  s 1  
 sK s  ---- 
 
 BI n i 
s -
  BI tot – BI tot ------------------------------------------------ ------p- ,
 
p s1  s1  p 
 K s  ---- 

 BI  n i 
 s 
 p  1  s1 
s
 
s
BV
 K s  ----
BV n i

J tot = J s = –D 0 s
----------------------------------------------------------------  3-92
BV BV BV s1  s1 
 

p
s K s ---- 
BV  n i
s
 p s1  s1  
 sK s  ---- 
 
 BV  n i 
  BV – BV --------------------------------------------------
tot tot s 
------p- ,
 
p s1  s1  p 
 K s  ---- 

 BV  n i 
 s 
where s1=s+1.
The equations for pairs formed by donor impurities are completely symmetrical to the
equations above.
Generation-Recombination Terms
The generation-recombination terms GRX in Equation 3-82 describe the evolution of
particular reactions. For example, GRAV represents the formation of the dopant-vacancy:
f
kAV
A s + V  AV 3-93
r
kAV

Formation of Pairs
In the model, the formation of dopant-defect pairs is taken into account by simulating the
following reactions:
f
kAI
– 0 –
As I  AI
+ 3-94
r
kAI
f
kAV
– + 0
As + V  AV 3-95
r
kAV
where kf and kr are the reaction kinetic constants for each reaction. They are define as
k AI = 4  R D I ,
f eff r f
k AI = k 0 K 0 3-96
AI AI
k AV = 4  R D V ,
f eff r f
k AV = k AV  K 0 3-97
AV
where KAI0, KAV0 are defined by Equation 3-81, and Reff is the silicon lattice constant. Thus,
the generation-recombination terms in Equation 3-82 are as follows:
f – 0 r 0
GR AI = k AI A s I – k AI AI , 3-98
f – + r +
GR AV = k AV A s V – k AV AV 3-99
Frenckel Pair Recombination

During annealing, many of the interstitials and vacancies recombine either at the surface or in
the bulk. The driving force for this reaction is to change both interstitial concentration Itot and
vacancy concentration Vtot toward their equilibrium concentrations I eq and V eq . Moreover, it
tot tot
is clearly shown that defect recombination strongly depends on the impurity concentration.
The following reaction is also considered.
f
kVI
 
V + V   0 3-100
r
kVI
The recombination rate can be written as follows:

tot tot eq eq
GR IV = k BM  I eq V eq – I V . 3-101

Here:
 
s  s  ni s
k BM = k a    +    ---- – 1

 s
 n 

3-102
and
k a = 4  a 0  D I + D V , 3-103
where a0=2.35A is the distance between two separated silicon atoms in crystal.
Bimolecular Recombination
As an alternative to the direct recombination of point defects in reaction of Equation 3-100, it
is also possible for I and V to recombine through reactions such as
f f
kAV – I kAI – V
0 + + 0 + +
AV + I  As  AI + V  As . 3-104
r r
kAV – I kAI – V
In these cases, the annihilation of the Frenckel pairs implies a dissociation of a dopant-defect
pairs.
Recombination at the Surface
Understanding of the mechanisms that determine the interaction of interstitials and vacancies
with the interfaces is getting more important, because the implantation energies of dopants
and the temperatures for the thermal treatments become lower and the devices are fabricated
closer to the surface. It has been also demonstrated that the fundamentals of the point defect
properties are critically important in accurate prediction of device behavior. For example, the
reverse short channel effect.
The recombination flux of silicon self-interstitials at a nonoxidizing interface  is given by
the formula:
D tot tot D
= -----I  I – I eq  , = ------V-  V – V eq  ,
tot tot
JI JV 3-105
 LI  LV
where LV and LI are called the recombination length at the surface. These parameters can be
used for adjusting the recombination rate at the surface and are specified in the defect.mod
file.
In the case of dopant implantation, an exodiffusion may occur during the thermal treatment.
Therefore, you can write the dopant flux at the surface as:
tot tot
JAI =  AI AI , JAV =  AV AV 3-106
 
where AI and AV are the exodiffusion coefficients for each dopant-defect pair and are
defined in dopand.mod files. These parameters are defined by the following Arrhenius
functions:

E  AI E  AV
 AI =  AI exp  – -----------
- ,  AV =  AV exp  – ------------
- .
0 0
3-107
kT  kT 
3.2.2 Solid Solubility Model

The model for precipitation is assumed to be a constant solid solubility cut-off. This means
that all solute atoms above the solid solubility level will form a precipitate almost
instantaneously. This model is activated by adding the SS parameter to the METHOD PLS
statement.
The rate equation for the solid solubility model can be formulated as follows [26]:
D   As –A  T   for A s > A ss  T ,
A pr 
- = D  A pr  A s – A ss  T   +
--------  3-108
t 0 for A s  A ss  T ,
where Apr is the concentration of the precipitate and  is the effective length of capture. The
solid solubility Ass(T) is defined using Arrhenius expressions in the corresponding
dopant.mod file.
3.2.3 Interstitials Clusters Model (IC)

Point defects in crystalline materials inherently have high free energy. Free interstitials in
silicon are thermodynamically unstable because of their unpaired electron orbitals and
induced lattice strain. At high concentrations, the interstitials clusters are formed to reduce
free energy. Many of interstitial cluster species have been observed for many years (e.g.,
{311} defects and dislocation loops). The interstitial cluster configurations are believed to
occur mainly in ion-implanted silicon. The formation and dissolution of interstitial clusters
are simulated to correctly predict TED.
To activate the interstitial clusters models, use the following statement:
METHOD PLS IC
During a typical rapid thermal annealing, various type of clusters (small clusters, {311}
defects, perfect and faulted loops) evolve according to a competitive growth mechanism
named Ostwald ripening. The driving force for this evolution is the reduction of the formation
energy per interstitial of these clusters as they grow in size and change their crystallographic
structures.
In IC model, a cluster containing n interstitials (In) evolve to a cluster of size n+1 by
interaction with a free interstitial according to the following reactions:
f f f
k2 k3 kn

0 0 0 0 0 0 0
I + I  I2  I2 0
+ I  I 3 0 In – 1 + I  In  3-109
r r r
k2 k3 kn
where according to [22]:
g
I
D Ef – E f  n 
k n = 4  R eff D I exp  – --------n- , k n = 6 -----I  n exp  – -------------------------
f r
3-110
 kT   

2 kT

Here, Reff represents the effective capture radius. The elementary jump length  is equal to the
inter-atomic distance a0=2.35°A. n is the number of dissociating sites, E f is the self

I
interstitial formation energy defined in Equation 3-76, and Ef(n) is the formation energy per
interstitial for clusters of size n. The number of reactions taken into account is specified in the
ic.mod file.
The value of effective energy barrier gn in Equation 3-110 can be represented as the sum of
two components:
  I tot  
 g n = max  – k B T ln  ----------- + E f  n  ,0 . 3-111
  I eq  
tot
The first of the two terms represents the change in free energy associated to the change in
chemical potential when an interstitial jumps from the supersaturated phase to the cluster. The
second term is the formation energy per interstitials. This parameter is a function of the size
and of the crystallographic structure of the cluster. The values of the formation energy per
interstitial are specified in the ic.mod file. The first nine parameters corresponding to the
formation energy per interstitial for clusters of size 2 through size 10 are defined in [23]. For
larger clusters, the energy of formation per interstitial is calculated by the following empirical
formula:
E f  n  = A log  1 + --------------------
1 - + E  311     . 3-112
 C f
n + B
 311 
The empirical parameters of the model A, B, C and E f are specified in the ic.mod file.
Thus, the system of coupled equations for the IC model can be written as:
N
tot
 I - = –  · J + GR + GR + 2GR
-----------
t I IV AI IC  2  +  GR IC  n ,
n=3
tot
-------------
V - = –  J + GR + GR ,
· V
t IV AV
tot
---------------
AI - = –  J – GR + GR
· AI AI– V , 3-113
t AI
tot
 AV - = –  · J – GR + GR
----------------- AV – I ,
t AV AV
A
--------s- = GR AI + GR AV – GR AV – I – GR AI– V ,
t
 In
-------- = GR IC  n  – GR IC  n + 1 , n = 2 ,3 , ,N,
t
where the generation-recombination term:

f tot r
GR IC  n  = k n I I n – 1 – k n I n . 3-114

It is important to notice that the IC model is completely independent of the dopants involved
in the process.
3.2.4 Vacancy Cluster Model (VC)

In the VC model, a cluster containing m vacancies (Vm) evolves to a cluster of size m+1 by
interaction with a free vacancy according to the following reactions:
f f f
k2 k3 km

0 0 0 0 0
V + V  V2 , V2 + V 0  V 30 , Vm – 1 + V 0  V m0 , 3-115
r r r
k2 k3 km
where:
V
D E f – E f  m 
= 4  R eff D V , = 6 ------V-  m exp  – --------------------------
- , m = 2,  , M .
f r
km km 3-116
2  kT 

Here, Reff represents the effective capture radius and the elementary jump length  is equal to
V
the inter-atomic distance. m is the number of dissociating sites. E f is the vacancy formation
energy defined in Equation 3-76, and Ef(m) is the formation energy per vacancy for clusters
of size m.
To take into account the interactions VC with interstitials, the following reactions are added:
f f
f
k2 kI4 k Im + 1
0 0 0

0 0
V2 +I  V2 , V3 + I 0  V 30 , 0 0
Vm + I  Vm – 1 3-117
r r
k2 k V3 r
k Vm
Therefore, the VC model consists of Equations 3-113 and 3-118.

V
---------m- = GR VC  m  – GR VC  m + 1  3-118
t
According to the following reactions, Equations 3-115 and 3-117, there are two kinds of
summands in the generation-recombination part.
f 0 f 0 r 0 r 0
GR VC  m  = k m V V m – 1 + k Im + 1 I V m GR VC  m + 1  = k m V m + k m – 1 V m – 1 3-119
The equations for vacanies and intersititials will also contain the additional terms from
Equation 3-119.

3.2.5 Electrical Deactivation and Clustering Models (DDC)

At high doses of dopant, the electrically active concentration may be less than corresponding
chemical concentration. The impurity atom becomes activated inside semiconductor only if it
is incorporated into a substitutional lattice site. In this case, the activated atom will contribute
with a carrier to either the valence band (an acceptor impurity), or the conduction band (a
donor impurity). It has been observed that, even below solid solubility, a significant dopant
concentration can stay inactive. This effect can be explained by the formation of immobile
dopant-defect clusters, which is described by the DDC model. This is the third part of the PLS
model. The model strongly depends on the nature of the dopant and therefore is presented
separately for each type of dopant below. To activate the DDC model, add the DDC parameter
to the METHOD PLS statement.
Boron
In the case of boron, these clusters are named boron interstitial clusters (BIC). The BIC
species BnIm consists of n atoms of boron and m atoms of silicon self-interstitials. In absence
of any direct experimental data concerning the exact composition of these clusters, BIC
structure and charge states are chosen according to recent ab-initio theoretical calculation
[21]. Various possible path are considered for these clusters: a given cluster can grow or
dissolve by the addition or release of a silicon self interstitial or a boron-interstitial pair
(Figure 3-2).
Figure 3-2 BIC reaction paths [5]

By default, the DDC model is based on the formation and dissociation of four BIC’s species:
BI2, B2I, B3I2, and B4I3. Thus, the following reaction are added.
f
k BI k
f
2 + + B2 I
+ – 0
BI  B2 I 
0
BI + I  BI 2  Bs + r
3-120
r kB I
k BI 2
2
f f
k kB I
+ B3 I2 + + – 4 3
B 2 I + BI  B3 I2 
0 0
r
B 3 I 2 + BI r B4 I3 . 3-121
kB I kB I
3 2 4 3
By default, the model assumes that dopants just after implantation are inactive. As this model
does not assume any local equilibrium between each species, the activation of the dopant will
gradually evolve with time.
The kinetic constants kf and kr for each reactions are defined as:
DB I E b  B n I m 
-  exp  – ----------------------
f r
kB = 4  R eff D BI , kB = 6 -------- 3-122
n Im n Im

2 n  kT 
Reff and Eb(BnIm) are respectively the effective capture radius and the binding energy for each
Boron-Interstitial Cluster and can be specified and modified in the boron.mod file.
Phosphorus
For phosphorus, the similar situation can be considered [25]. Therefore, phosphorus atoms
can form clusters with self-interstitials. These clusters have been experimentally observed
and are called phosphorus interstitials clusters (PIC).
The following reactions are taken into consideration.
f f
k k
– – P2 I 0 0 PI 2 0
Ps + I  P2 I  PI +I  PI 2 
0
r r
3-123
kP k PI
2I 2
These and additional reactions and their parameters can be specified in the phosphorus.mod
file.

Arsenic
The case of arsenic is a little bit different. Arsenic migrates through the vacancy and
interstitial mechanisms with roughly the same proportion. It is well known that arsenic can
form clusters with vacancies of type AsnV. Therefore, the DDC model needs to take into
account the following reactions:
f f
k As V k As V
+ – 2 0 3 0
As 2 V + AsV 
0 0 0
As s + AsV  AsV , r
As 3 V + V , 3-124
r
k As V k As V
2 3
f
k As
4V
0 0
As 3 V + AsV 
0
As 4 V + V 0 3-125
r
k As V
4
The system of equations for DDC model is

tot
 I - = –  J + GR + GR + GR
----------- · I AV – I + GR AIC ,
t IV AI
tot
-------------
V - = –  J + GR + GR + GR
· V AI – V + GR AVC ,
t IV AV
tot
---------------
AI - = –  J – GR + GR
· AI AI – V + GR AIC ,
t AI
tot 3-126
 AV - = –  · J – GR + GR
----------------- AV – I + GR AVC ,
t AV AV
A
--------s- = –  · J AV + GR AV – GR AI – V – GR AV – I + GR AIC + GR AVC ,
t
------------
AIC- = GR ,
t AIC
 AVC- = GR ,
--------------
t AVC
where GRAVC, GRAIC are generation-recombination terms caused by reactions (Equations 3-

120 and 3-125).

3.2.6 Typical Examples

The following will show each part of the PLS model validated using specific experimental
results.
The CDD model is tested using a simple predeposition step. The interstitial cluster part of the
model is validated using the Cowern’s experiment [23], and the mixed cluster part of the
model is then analyzed using the Pelaz experiment [27]. To illustrate the improvements given
by the PLS model, we perform simulations within a very broad range of experimental
conditions from a standard implantation and diffusion step to a "state of the art" RTA
Predeposition
Boron
As a first indication that the PLS model is able to handle the complex couplings between
boron and the free point defects, we show in Figure 3-3 the result of the conventional
predeposition steps simulation (diffusion with a constant surface concentration).
Figure 3-3 Simulation of Boron predeposition using the CDD model at various temperature, time annealing
and surface concentration. (Crosses, pluses, and squares are experimental data from [28])

Although not corresponding to the modern deep sub-micron technologies, this simulation
represents the high dopant concentration features that reveal the complex couplings between
dopant and point defects. Therefore, it is considered as a meaningful basic test for any
advanced diffusion models.
Although this model has been developed for advanced silicon technologies, it still can be used
as the standard diffusion model for any diffusion step. For example in the case of buried layer
formation, the TED phenomena become irrelevant. Therefore, you can use only CDD part of
the PLS model while ignoring IC and DDC models.
Phosphorus
To illustrate the improvements that result from the CDD model, we show simulations of
phosphorus predeposition profiles at high and intermediate concentrations. The simulation
results are compared to the SIMS data of Yoshida and Matsumoto. This experiment represents
high dopant concentration features that reveal the complex couplings between dopants and
point defects. Therefore, it is considered as a meaningful test for advanced diffusion models.
Figure 3-4 Simulation of phosphorus predeposition at 900°C during 1 hour with various surface
concentration. (Squares, pluses, and crosses are experimental data from [29])
As our simulations prove, the PLS model accurately reproduces the experimental profiles
features. Particularly, the simulated profiles exhibit the enhanced tail with more or less

pronounced inflexion in the surface region. This inflexion is the result of the strong coupling
between the defect gradients and the dopant.
Cowern’s Experiment
To validate the IC model, we first try to compare our model prediction with the experimental
data obtained by Cowern et al. [23]. Briefly described, this experiment consists of observing
the diffusivity of two boron marker layers after a silicon implantation at 40 keV to a dose of
2 1013 atoms/cm2 (simulated in the left hand side from Figure 3-5). From these observations,
Cowern et al. have estimated the diffusion enhancement and the evolution of supersaturation
with time. On the right hand side of Figure 3-5, the PLS simulation of the supersaturation
evolution is shown.
Figure 3-5 Simulation of the Cowern experiment and extraction of the evolution of the supersaturation during
the annealing. (Experimental data are from [23])
Figure 3-5 shows the free interstitial supersaturation behavior with time. This quantity is
simply calculated using the ratio between local effective interstitial concentration and its
equilibrium value. From the experimental point of view, this quantity can be related to the
boron diffusion enhancement with respect to its thermal equilibrium diffusion.
The supersaturation evolution curve exhibits three parts:
• The first step characterized by a high supersaturation value corresponds to a large
acceleration of the dopant diffusion and can be explained by the presence of small
clusters.
• The second step exhibits a plateau slightly decreasing with time. It is explained by the
competitive growth between <311> defects known as the Ostwald ripening phenomena.
• The third step characterized by the supersaturation collapse is explained by the entire
dissolution of the IC population, due to the recombination at the surface.
It is clear that the CDD model alone (dashed lines in Figure 3-5) cannot reproduce these three
steps of the transient enhanced diffusion, since the curves are monotonically decreasing.

Pelaz Experiment
In this experiment, a boron marker is deposited at a depth of 0.15 m. To observe the boron
diffusion, Pelaz et al. have performed a silicon implant to generate a high interstitial
concentration at the surface. Unlike the Cowern experiment, the boron concentration is high
enough to allow the formation of BICs. Thus, this experiment exhibits a particular effects of
in the boron diffusion: an immobilization and disactivation of the dopant at high
concentrations even under the solid solubility limit.
Figure 3-6 demonstrates simulations of this particular experiment, using various parts of the
PLS model.
Figure 3-6 Simulation of the Pelaz experiment using various parts of the PLS model. (Experimental Data are
extracted from [27])
The results obtained with the full PLS model correspond to the experimental data. As
expected, the boron at the highest concentration stay immobile and are still inactive due to the
formation of mixed BICs. Moreover, it is clear that only the full model can explain this
behavior. Therefore, CDD and IC models cannot simulate this particular phenomenon
without DDC model involved.

Implantation Diffusion Experiment

The analysis above proves validity of each part and the entire PLS diffusion model. In
addition, we present several simulations relevant to specific processes important in the
modern VLSI technologies.
Experiment with boron implanted at 20 keV
Boron implantation at 20 keV is widely used in silicon technologies. Since the experiment
setup involves low temperature, such as 800°C, the effects of boron immobilization and
disactivation at high concentration take place. Therefore, you must use three parts of the PLS
model.
Figure 3-7 Simulation of 35 minutes boron diffusion at 800°C after an implantation at 20 keV
with a dose of 5  1014 cm-2. (Experimental data are extracted from [30])
As shown in Figure 3-7, PLS model results in an excellent fit with the experimental annealed
curves. The comparison of chemical and electrically active boron concentration shows that
even below the solid solubility, most of the dopant stay inactive due to the formation of mixed
dopant-defect clusters. By comparing results obtained with only CDD or IC model with the
full PLS (IC + DDC), it can be concluded that this type of clusters also consume free
interstitials which leads to reduction of the boron diffusion. This fact justifies one more time
that all three parts of the model: CDD, interstitial clusters and mixed dopant-defect clusters
are needed in simulation of post-implant diffusion.

Experiment with boron implanted at 2 keV

A most aggressive technology simulation has been carried out with a 2 keV boron implant
with a dose of 1014 cm-2 followed by a RTA at 1000°C during 10 seconds. This type of
simulation is difficult because it needs to take into account a large number of phenomena
including strong defect recombination at the surface, and the fact that TED duration is
function of the implant setup and subsequent diffusion duration. Despite of these difficulties,
Figure 3-8 shows a very good fit with experimental annealed data which proves the excellent
quality of the model.
Nevertheless, as presented in the literature, the TED effect decreases at very low implant
energy and high temperature annealing. This explain why the simulations done using CDD
and IC models and only CDD model sometimes give reasonable agreement with experiments.
Figure 3-8 Simulation of boron diffusion at 1000°C during 10 s after an implantation at 2 keV
with a dose of 1 1014cm-2. (Experimental data are from [31])
Experiment with arsenic implanted at 2 keV
The PLS model is not only design for the boron diffusion but also for other common dopant,
such as phosphorus or arsenic.

An Arsenic implantation has been performed with a energy of 2 keV at a dose of 1015 at/cm-2
and followed by a spike RTA at 950°C with a ramp up estimated at 100°C/s. In this typical
process example for ultra shallow junction, dopant is implanted at such a high dose that its
concentration reaches the solid solubility limit. In this case, most of the dopant at
concentration higher that this limits will precipitate in the early stage of the annealing.
Moreover, all other specific characteristics in arsenic diffusion is taken into account. In other
words:
• Arsenic migrate both through interstitial and vacancy mechanisms with roughly the same
proportion.
• Arsenic atoms form with vacancy any clusters, As2V or As4V.
The initial interstitial profile generated by the arsenic implant is modeled though a simple
"plus n" model.
Figure 3-9 Simulation of Arsenic diffusion after an implantation at 2 keV with a dose of 11014 cm-2 and a
spike anneal at 950°C with a temperature ramp rate estimated at 100°C/s. (Experimental data are from [32])
The result obtain with the full PLS model is in good agreement with experimental data. As
expected, most of the arsenic at concentration above the solid solubility limit precipitate
quickly and consequently immobilize the dopant.

Oxidation Models SSuprem4 Models
3.3 Oxidation Models

The fabrication of integrated circuit microelectronic structures and devices vitally depends on
the thermal oxidation process for the formation of gate dielectrics, device isolation regions,
spacer regions, and ion implantation mask regions. Particularly, the precise control of silicon
dioxide thickness as device geometries continue to scale to sub-micron dimensions.
In SSuprem4 silicon thermal oxidation is modeled when a DIFFUSION statement contains a
DRYO2, WETO2, F.O2, or F.H2O parameter. Oxidation takes place when there is an interface
between silicon (or polysilicon) and silicon dioxide or a silicon (polysilicon) surface is
exposed to an oxidizing ambient. SSuprem4 simulates polysilicon oxidation in a very similar
manner as silicon (almost all oxidation parameters for polysilicon are the same as for silicon).
SSuprem4 also allows oxidation completely through a silicon (polysilicon) layer. This is very
important in processes (e.g., poly buffered LOCOS) in which polysilicon regions are
completely consumed during oxidation.
Because exposed silicon surfaces usually have a thin native oxide layer, SSuprem4
automatically deposits a thin native oxide layer on all exposed silicon (polysilicon) surfaces
at the beginning of oxidation steps. The INITIAL parameter in the OXIDE statement
determines the layer’s thickness, which has a default value of 20 Å.
The two-dimensional oxidation models in SSuprem4 are based on the well-known linear-
parabolic theory of Deal and Grove [33]. Numerical aspects of the model implementation can
be found in [34]. Silicon oxidation is modeled by considering the following three processes:
1. Oxidant (e.g., H2O or O2) is transported from the ambient gas into the SiO2 layer at the
gas/SiO2 interface.
2. Oxidant is transported across the SiO2 layer until reaching the Si/SiO2 interface.
3. Oxidant, arriving at the Si/SiO2 interface, reacts with silicon to form a new layer of SiO2.
The transport of oxidant across the gas/SiO2 interface is given by
*
F 1 = h  C – C 0 n 0 3-127
where:
• h is the gas-phase mass-transport coefficient.
• C* is the equilibrium oxidant concentration in SiO2.
• C0 is the oxidant concentration in SiO2 at the gas/SiO2 interface.
• n0 is a unit vector normal to the gas/SiO2 interface pointing toward the silicon layer.
The equilibrium oxidant concentration in SiO2 is linearly related to the partial pressure of the
oxidant, P, in the gas by Henry’s law:
*
C = KP 3-128
where K is a constant.
Diffusion of oxidant molecules in the SiO2 is driven by a concentration gradient and is given
by Fick’s law as
F 2 = – D eff  C 3-129

where Deff is the effective oxidant (H2O or O2) diffusivity in the growing SiO2 layer. C is the
oxidant concentration in the oxide.
The reaction at the Silicon (or Polysilicon)/SiO2 interface between silicon and the oxidant is
expressed as
F 3 = kC i n i 3-130
where:
• k is the apparent surface reaction rate constant.
• Ci is the oxidant concentration at the Silicon (or Polysilicon)/SiO2 interface.
• ni is a unit vector normal to the Si/SiO2 pointing toward the silicon layer.
Under steady state conditions, the three fluxes are equal.
F = F1 = F 2 = F 3 3-131
By dividing the flux by N1, the number of oxidant molecules incorporated in a unit volume of
SiO2, and considering one dimensional growth, the growth rate of the oxide layer is given by
dx 0 F-
-------- = ----- 3-132
dt N1
where x0 is the oxide thickness. From Equation 3-127 and Equations 3-129–3-131, Equation
3-132 can be expressed as
dx 0 B -
-------- = ------------------ 3-133
dt A + 2x 0
where:
A = 2D eff  1--- + 1--- 3-134

 k h
*
B = 2Deff C
------- 3-135
N1
Equation 3-133 is modified for thin oxides (less than 500 Å) as follows:
dx 0 B -+R
-------- = ------------------ 3-136
dt A + 2x 0
where R is calculated according to [35]:

THINOX  E x0
R = THINOX  0 exp  ------------------------------- exp  ------------------------------
-  P THINOX  P 3-137
 kT   THINOX  L
where P is partial pressure. THINOX.0, THINOX.E, THINOX.L, and THINOX.P parameters are
specified in the OXIDE statement.

3.3.1 Numerical Oxidation Models

In the previous section, an introduction to one-dimensional oxidation modeling was
presented. This section describes the two-dimensional numerical oxidation models
implemented in SSuprem4. The numerical oxidation models build on the Deal-Grove
oxidation theory and provide the capability to simulate arbitrary two-dimensional structures.
The numerical oxidation models require solving the oxidant diffusion equation at incremental
time steps at discrete grid points in the growing SiO2 layer. The oxidant diffusion equation is
given by
 C- =   F
------ 3-138
t
where:
• C is the oxidant concentration in SiO2.
• t is the oxidation time.
• F is the oxidant flux.
Equation 3-138 is solved by substituting Equation 3-129 for F and defining appropriate
boundary conditions at material interfaces with SiO2. At the gas/SiO2 interface, Equation 3-
127 describes the interface transport flux of oxidant molecules, accounting for the boundary
condition at that interface. The boundary condition at the Silicon (or Polysilicon)/SiO2
interface is described by Equation 3-130. The flux at boundaries between SiO2 and other
materials in the simulation structure is set to zero. By solving Equation 3-138, the oxidant
concentration is determined at each grid point in the SiO2 layer. The SiO2 growth rate or Si/
SiO2 interface velocity, Vs, is determined at each point along the interface by combining
Equations 3-130 and 3-132 resulting in the following.
kC i n i
V s = -------------
- 3-139
N1
At each time step, Equation 3-139 is solved. The incremental oxide thickness grown is
calculated by multiplying Equation 3-139 by the time step. During the oxidation reaction
silicon atoms bond with the oxidant to form the SiO2 compound. Thus, silicon material is
removed during the oxidation process. The ratio of the silicon thickness consumed to form a
given thickness of SiO2 is specified using the ALPHA parameter in the OXIDE statement.
Equation 3-138 is sufficient to describe the motion of the Si/SiO2 interface if the oxide flow is
in the same direction as the growth (for planar oxidation structures). In most structures of
interest, the oxide flow is two dimensional. Therefore, additional equations have to be solved.
Both the Compress and Viscous models calculate the two-dimensional flow of oxide elements
by solving a simplified hydrodynamic creeping flow equation.

3.3.2 Compress Model

In addition to solving Equations 3-138 and 3-139, a simplification of the hydrodynamic flow
equation is solved to obtain the flow of oxide elements [37]. The Compress Model is
activated by specifying COMPRESS in the METHOD statement prior to a DIFFUSE statement.
The Compress Model is the default oxidation model in SSuprem4. Neglecting the
acceleration and gravitational terms in the hydrodynamic flow equation, the creeping-flow
equation is given by
2
 V =  P 3-140
where P is the hydrostatic pressure, V is velocity of oxide elements and  is the oxide
viscosity.
The oxide viscosity is calculated from the following equation:
E -
 = -------------------- 3-141
21 + 
where Young’s modulus E and Poisson’s ratio  are specified through elastic constants as
described in Section 3.10 “Stress Models”.
The oxide flow is treated as an incompressible fluid. By doing this, it implies the density of
the oxide is constant with respect to time. Applying this fact to the mass continuity equation,
the incompressibility condition is given as
V = 0 3-142
The incompressibility condition in Equation 3-142 is implemented by allowing a slight
compressibility of the flowing oxide. Thus, Equation 3-142 is modified to give the following
equation:
1 – 2
  V = ---------------- P 3-143

The solution of Equation 3-143 at each time step gives the velocity field of the flowing oxide
elements. The Compress Model is recommended for simulations of planar and non-planar
structures, where stress effects play a minor role in determining the oxide shape. When stress
effects are important, you can use the Viscous oxidation model.

Figure 3-10 shows a two-dimensional cross section of the structure resulting from a LOCOS
oxidation using the Compress Model.
Figure 3-10 Resulting Structure from a LOCOS Oxidation step using the Compress Model
3.3.3 Viscous Model

The Viscous Model solves the same flow equations as described in the previous section. This
model is activated by specifying the VISCOUS parameter in the METHOD statement prior to the
DIFFUSE statement. The Viscous Model calculates stresses in the growing oxide and creates
almost the same shape for the silicon/oxide interface as does the Compress model.
The stresses in the oxide are calculated as follows:
2  VISC.0  exp  ----------------------  v

– VISC.E
 kT   x  v y
 xx +  yy = --------------------------------------------------------------------  -------- + -------- 3-144
1–2 x y
– VISC.E  v v
 xx –  yy = 2  VISC.0  exp  ----------------------  --------x – --------y 3-145
kT x y
– VISC.E  v v
 xy = VISC.0  exp  ----------------------  --------x + --------y 3-146
kT x y

where vx and vy are the x and y components of flow velocity v respectively. VISC.0 and
VISC.E are the pre-exponential and activation energy respectively for viscosity are specified
on the MATERIAL statement.  is Poisson’s ratio specified through elastic constants as
described in Section 3.10 “Stress Models”.
The stress-dependent nonlinear model based on Eyring’s work [36] allows a description of
the real shape of LOCOS profiles with kinks on the interface. The model is turned on by
specifying the STRESS.DEP parameter the OXIDE statement. Using Equation 3-140 and
Equations 3-143–3-146, the non-linear solver first finds a linear solution for flow velocities
and stresses and then uses the stresses obtained to calculate the reduction factors for oxidant
diffusivity, Deff, oxide viscosity, , and the interface reaction rate constant k as follows:
i i – 1 V d   xx +  yy 
D eff = D eff  exp  -------------------------------------- 3-147
 kT 
V
 --------c-
i  i – 1   2kT
 =  -------------------------- 3-148
 Vc
sin h  ---------
 2kT
i i – 1 r Vr + t Vt
k = k  exp  – ------------------------------
- 3-149
 kT 
where i is the iteration. Vd, Vc, Vr, and Vt are the activation volumes (in Å3), specified in the
OXIDE statement.
 is the total shear stress:
2 2
 = 1---   xx –  yy  + 4  xy 3-150
2
r is the normal component of the total stress:

2 2
 r =  xx n x +  yy n y + 2  xy n x n 3-151
y
t is the tangential component of the total stress:

2 2
 t =  xx n y +  yy n x + 2  xy n x n 3-152
y
where nx and ny are the x and y components of the unit vector normal respectively.
The reduced parameters feed back to the next iteration. This process continues until the
accuracy criterion is met. Fast convergence of this process is not guaranteed. Oxidation
calculations by the stress-dependent model usually take much more CPU time than the
Compress Model.

Figure 3-11 shows the resulting structure from a LOCOS oxidation step using the stress
dependent Viscous Model.
Figure 3-11 Resulting Structure from a LOCOS Oxidation step using the Stress-Dependent Viscous Model
3.3.4 Linear Rate Constant

For short oxidation times and low oxidation temperatures the oxide growth is linearly related
to the oxidation time. The interface processes (oxidant transport across the gas/SiO2 interface
and oxidant reaction at the Si/SiO2 interface) are the determining factor in describing the
growth kinetics. In this regime, the oxide thickness can be approximated as
x 0   --- t
B
3-153
 A
where (B/A) is called the linear rate constant and is obtained by dividing Equation 3-135 and
Equation 3-134, resulting in the following equation:
*
B--- = C
-------  1--- + 1--- 3-154
 A N 1  h k
The equilibrium oxidant concentration in the oxide, C*, is defined by Equation 3-128. K in
Equation 3-128 is specified by the HENRY.COEF parameter in the OXIDE statement. The gas-
phase mass-transport coefficient, h, is given by the following Arrhenius relation:

 
h = TRN.0  exp  – TRN.E
----------------- 3-155
 kb T 
where the TRN.0 and TRN.E parameters are specified in the OXIDE statement. The interface
reaction rate constant, k, is determined from Equation 3-154 and experimentally determined
values of (B/A).
The linear rate constant is composed of several dependencies including orientation, pressure,
chlorine additions, and doping effects.
B--- =  B
---  B
---  B ---  B
---  B--- 3-156
 A  A i  A ori  A P  A Cl  A doping
(B/A)i is given by:

  
 LIN.L.0 exp  – LIN.L.E
--------------------- T < L.BREAK
  kb T 
B 
--- =  3-157
 A i  LIN.H.E

 LIN.H.0 exp  – ----------------------  T  L.BREAK 
  kb T 

which is the linear rate constant determined for oxidations on lightly-doped substrates
annealed at atmospheric pressure with no chlorine content in the ambient. The parameters
appearing in Equation 3-157 are specified in the OXIDE statement. The remaining factors in
Equation 3-156 are described in the following sections.

Orientation Dependence
The silicon substrate orientation is known to affect the oxidation kinetics [37, 38]. The
influence of orientation on the linear rate constant is modeled as (B/A)ori in Equation 3-156.
The orientation dependencies for <100> and <110> orientations are modeled by appropriate
reduction factors, and (B/A)ori for <111> substrates is unity.
Figure 3-12 shows the silicon dioxide thickness dependence as a function of the substrate
orientation for several oxidation temperatures.
Figure 3-12 Silicon Dioxide Thickness versus Time for Different Substrate Orientations and Temperatures

Pressure Dependence
High pressure silicon oxidation allows one to grow relatively thick SiO2 films while keeping
the temperature low so that dopant redistribution is reduced [39]. The pressure dependence in
the linear rate constant is given by
B--- = P
L.PDEP
3-158
 A P
where L.PDEP is specified on the OXIDE statement for each oxidant and P is the partial
pressure of the oxidizing gas.
Figure 3-13 shows the silicon dioxide thickness versus time with PRESSURE as a parameter.
Figure 3-13 Silicon Dioxide Thickness versus Time with Pressure as a Parameter

Chlorine Dependence
The addition of chlorine to the oxidation system results in better passivation and higher oxide
dielectric strength [33, 38]. For a dry oxygen ambient, chlorine introduction gives rise to a
higher oxidation rate. It has been suggested [38] that chlorine reacts with O2 to produce H2O
and Cl2 as products. The oxidation rate is higher in H2O ambients than in O2 ambients
because equilibrium concentration of H2O in the oxide is higher. A look-up table approach is
implemented to model the increase in the linear rate constant in Equation 3-156 though the
(B/A)Cl term. The table gives an enhancement factor to the linear rate constant as a function of
chlorine percentage and temperature. The default values for chlorine dependence are included
in Appendix B “Default Coefficients”.
The effects of adding chlorine to the oxidizing ambient is shown in Figure 3-14, where the
silicon dioxide thickness increases as more chlorine is added to the ambient.
Figure 3-14 Silicon Dioxide Thickness Versus Oxidation Time with HCl Percentage and Temperature as
Parameters

Doping Dependence
It is well known that SiO2 formation on highly-doped n-type and p-type substrates can be
enhanced compared to SiO2 formation on lightly-doped substrates [40]. The dependence of
silicon dioxide growth kinetics on doping concentration is manifested as part of the linear rate
constant, where the physical significance of the high doping levels has been explained
primarily as an electrical effect [40], [41]. This factor in the linear rate constant is given by
B--- – BAF.KE  V 
= 1 + BAF.K0  exp  -------------------------  -------
- – 1 3-159
 A doping  k T  V  
b i
where V* is the equilibrium vacancy concentration in silicon at the Si/SiO2 interface. Vi* is
the equilibrium vacancy concentration in intrinsic silicon. BAF.K0 and BAF.KE are specified
on the OXIDE statement.
The equilibrium vacancy concentration, composed of vacancy defects in different charged
states, depends on the Fermi level location [8], [42] and is given by
 n + n 2 n 2 
 i  +  ----i  ++ +  ---- n  –  ----
 +    = 
 1 +  ---- n   n n  n
 i i 
V = V i  ------------------------------------------------------------------------------------------------------------------  3-160
 1 +  + +  ++ +  +  – = 
 
 
where n is the electron concentration and ni is the intrinsic carrier concentration, and
 BAF.PE
+ = BAF.EBK exp  – --------------------- 3-161
 kb T 
BAF.PPE-  
 ++ = BAF.EBK exp  – ------------------------  3-162
 kb T 
–  
 = BAF.EBK exp  – BAF.NE
--------------------- 3-163
 k b T 
=  BAF.NNE
 = BAF.EBK exp  – -------------------------- 3-164
 kb T 
where , , , and  are fractions of the vacancy concentration which are positively,
double positively, negatively, and double negatively charged respectively.
Figure 3-15 shows a plot of V*/Vi* at 950°C for common silicon dopants. Notice that for n-
type dopants (V*/Vi*) increases as the doping concentration increases, but V*/Vi* remains
essentially constant for the p-type dopant. The increase in V*/Vi* for n-type dopants increases
the linear rate constant. This ultimately leads to thicker oxides when oxidizing highly-doped
n-type substrates due to a higher availability of unoccupied silicon lattice sites (vacancies) for
oxidant molecules to be incorporated.

The oxide thickness trend is shown in Figure 3-16, where the SiO2 thickness is plotted versus
doping concentration for common silicon dopants.
 
Figure 3-15 V  V i Ratio versus Doping Concentration
Figure 3-16 Simulated Silicon Dioxide Thickness vs. Doping Concentration for Common Silicon Dopants

3.3.5 Parabolic Rate Constant

For long oxidation times and high temperatures the oxide growth is parabolically related to
the oxidation time. The diffusion of oxidant in the oxide is the determining factor in
describing the growth kinetics. For these times and temperatures, the oxide thickness can be
approximated as:
2
x 0  Bt 3-165
where B is called the parabolic rate constant and is given by Equation 3-135. When using this
equation, the oxidant diffusivity, Deff, is determined from specified values for C*, N1, and
experimentally-determined values of B. The parabolic rate constant has been determined to
have dependencies on the ambient pressure and the chlorine content during oxidation and is
given by
B = B i  B P  B HCl 3-166
where:
  
 PAR.L.0 exp  – PAR.L.E
----------------------- T < P.BREAK
  kb T 

Bi =  3-167
  PAR.H.E
 PAR.H.0 exp  – ------------------------ T  P.BREAK
  kb T 

Bi is determined as a function of temperature and time for lightly-doped substrates annealed
at atmospheric pressure with no chlorine content in the ambient. The parameters in Equation
3-167 are specified for the appropriate oxidant species using the OXIDE statement. The
pressure dependence and chlorine dependence are described in the following sections.
Pressure Dependence
The effects of pressure on the kinetics of the silicon oxidation process have been studied by
Razouk et al. [39] for pyrogenic steam and Lie et al. [42] for dry oxygen. The parabolic rate
varies with pressure because of its dependence on the oxidant equilibrium concentration in
the oxide, C*, which is directly proportional to the partial pressure of the oxidizing gas. The
following relation is used to model this dependency.
P.PDEP
BP = P 3-168
Here, P is the partial pressure of the oxidizing gas in atmospheres and P.PDEP is specified on
the OXIDE statement.
See Figure 3-13 for a plot of SiO2 thickness as a function of time and pressure.

Chlorine Dependence
It has been observed that additions of chlorine during thermal oxidation also affect the
parabolic rate constant. One possible explanation is that as chlorine enters the oxide film, it
tends to cause the SiO2 lattice to become strained, which increases the oxidant diffusivity
[38].
Chlorine concentration dependence on the parabolic oxidation rate is modeled in a similar
manner to that of the linear rate constant. Given an HCl percentage, a look-up table is used to
determine an enhancement factor for the parabolic rate constant. Figure 3-14 shows the SiO2
thickness dependency on HCl percentage.
3.3.6 Mixed Ambient Oxidation

In practice, an oxidizing ambient may be a gas mixture consisting of more than one oxidant
and other impurities. The total oxidation rate will be the combined effect of all these species.
To simulate oxidation under a multi-gas ambient, SSuprem4 simultaneously calculates the
diffusion and oxidation of several ambient gases.
The capability is invoked by specifying the gas flow parameters: F.O2, F.H2, F.H2O, F.N2,
and F.HCL on the DIFFUSE statement. From the gas flow, the partial pressure of each gas is
calculated as:
Fj
P j = P total ------------- 3-169
 Fj
where Pj and Fj are partial pressure and gas flow rate for the jth gas respectively, and Ptotal is
the total pressure of the gas mixture (specified by the PRESSURE parameter on the DIFFUSE
statement). If only one oxidant gas is specified in the gas flow (i.e., only O2 or H2O with other
gases), oxidation is then modeled as previously described. Equation 3-169 determines the
pressure of the oxidant gas. If both F.H2 and F.O2 are specified, the reaction of H2 and O2 to
form H2O is assumed to occur. The partial pressure of H2O is then calculated before solving
the oxidation equations.
For ambients containing more than one oxidant (e.g., O2 and H2O), the partial pressure of
each oxidant is used to calculate C* for each species. From C*, k and Deff for each oxidant
species are calculated in a similar manner as that described in the pairs sections, respectively.
Equation 3-138 is solved for each oxidant to obtain each oxidant’s concentration distribution
in the growing SiO2. The contributions of each oxidizing species to the Si/SiO2 interface
velocity is calculated with the following equation:
k j C ij n ij
Vs =
 -------------------
N 1j
3-170
j
where Equation 3-139 has been used and j corresponds to the jth oxidant gas.
The flow equations are also calculated for a mixed ambient where both O2 and H2O exist and
COMPRESS or VISCOUS has been specified on the METHOD statement. The stress dependence of
Deff and k is a function of the composition of dry or wet oxide which depends on oxidation
history. Mixed ambient oxidation simulations take longer to solve than simple ambient
equations.

3.3.7 Analytical Oxidation Model

You can use the analytical oxidation models to simulate a limited set of simple structures.
Possible structures include a silicon substrate with an oxide layer deposited (or grown) on it.
Since you can only specify the mask at the left part of a simulated structure, oxidation will
only occur to the right of the mask edge. Analytical methods do not account for any real
material layer located to the right of the specified mask edge. As the oxide layer thickens, the
material is elevated, but growth rate and oxide shape are not affected. In all analytical models,
the initial silicon surface must be planar.
The ERFG model simulates the bird’s beak oxide shape under nitride masks of different
thicknesses [43]. The ERFG model consists of two models: ERF1 and ERF2. The ERF1
model describes the oxide growth under a thin nitride layer, where the stress from the nitride
mask layer is negligible. ERF2 model describes the oxide growth when nitride layer
thicknesses are large enough to cause stress in the oxide, which can result in the oxide layer
being pinched. When ERFG is specified, either the ERF1 or ERF2 model will be
automatically selected based on the structure under consideration. Both models are based on
the error-function shape of the oxide/silicon and oxide/ambient or oxide/nitride interfaces:
Z = A erfc  By + C  + D 3-171
The A, B, C, and D parameters are complex functions of several geometric parameters:
• initial thickness of oxide lox and nitride ln,
• current thickness Eox of oxide given by the Deal-Grove Model (Equation 3-132),
• the length of lateral oxidation under the nitride layer Lbb,
• and the lifting of the mask during oxidation H.
These functions are specified in the OXIDE statement. All defaults are taken from [43].
3.3.8 Recommendations for Successful Oxidation Simulations

Achieving successful oxidation simulations can be a frustrating task for a novice user of
process simulation software. This section highlights some of the most common barriers
encountered using process simulation to model oxidation steps, and describes how to
overcome those barriers with the proper methods for simulating these oxidation steps.
One of the most common errors made in simulating oxidation steps is improperly gridding of
the oxide structure. Improper gridding can result in jagged oxide shapes and errors in
resolving impurity distributions. As the oxide layer is growing, grid points are added at
predefined spacings. As silicon is being consumed, dopants are transported across the Si/SiO2
interface. It is important to obtain a well gridded oxide to properly account for dopant
redistribution during the oxidation step.
Growing Thin Oxides
A typical application where thin oxide growth is important is during a gate oxidation step of a
MOSFET which has a highly-doped polysilicon gate. By default, SSuprem4 uses a grid
spacing of 0.1 microns in the growing oxide layer. Thus, one grid layer will be added in the
growing oxide every 0.1 microns (or 1000 angstroms). This grid spacing is appropriate for
field oxidations, and hence the reason it is the default grid spacing in the growing oxide layer.
Using the default grid spacing in the oxide for typical gate oxidations in today’s MOS
technology results in no grid being added in the interior of the SiO2 layer. With no grid
present in the oxide to resolve the dopant diffusion in the oxide during subsequent processing,

the polysilicon dopant can penetrate into the underlying silicon substrate. This simulation
artifact can cause threshold voltages to be very different than expected.
To rectify this simulation artifact, you can control the number of grid layers added during the
oxidation with the GRID.OXIDE and GRIDINIT.OXIDE parameters in the METHOD statement.
You should place this statement before the gate oxidation diffusion step. Setting these
parameters to a value, which results in three or four grid layers in the gate oxide (e.g., 15
angstroms for a 60 angstrom gate oxide thickness), can alleviate this problem. We suggest
you to set back these parameters to the default values after the gate oxidation step.
Figure 3-17 shows a cross section of an NMOSFET with a highly-doped phosphorus
polysilicon gate. The default grid spacing in the oxide is used in Figure 3-17(b), while the
grid spacing is adjusted properly for Figure 3-17(a). By comparing these two figures, it is
obvious that phosphorus has penetrated through the gate oxide for Figure 3-17(b) but does not
penetrate through the gate oxide in Figure 3-17(a).
Figure 3-17 (a) MOSFET Structure with Proper Gridding in Gate Oxide, (b) MOSFET Structure with Default
Grid Spacing in Gate Oxide

Implantation Through Thermally-Grown Oxides and Dopant Loss During Subsequent Annealing
Frequently, dopants are implanted through thermally grown oxide layers. It is important to
have a proper grid spacing in the oxide through which the dopant is implanted for two
reasons. First, this will aid in determining the proper dopant profile in the oxide layer and the
underlying silicon. Secondly, proper gridding is required to resolved the dopant diffusion in
the oxide during subsequent processing steps. During annealing, the dopant will diffuse in
SiO2 and silicon and eventually evaporate into the ambient at the gas/SiO2 interface. If proper
gridding is not supplied in the growing oxide layer, the amount of dopant evaporating can be
underestimated, yielding a larger dose retained in the silicon substrate. The mechanism is
similar to what was described in the earlier sections. There may not be any grid points in the
interior of the growing SiO2 layer. The problem is again remedied by specifying more grid
layers to be added as the SiO2 layer grows.
Figure 3-18 shows a comparison of the resulting arsenic profiles in silicon using the default
grid spacing and a corrected grid spacing in the growing SiO2 layer. For this experiment, a
silicon dioxide layer was thermally grown. Arsenic was ion implanted through the SiO2/
Silicon structure. A subsequent annealing step followed which results in the profiles shown in
Figure 3-18.
Figure 3-18 Comparison of Arsenic Profiles in Silicon with Default Grid Spacing and Improved Grid Spacing
in the Growing SiO2 layer.

Oxidation Enhanced Diffusion (OED) / Oxidation Retarded Diffusion (ORD)

During silicon thermal oxidation, some of the dopant in silicon gets incorporated into the
growing SiO2 layer and some remains in silicon where it diffuses. As oxidation proceeds,
silicon lattice atoms become interstitial (interstitials are injected into silicon at the Si/SiO2
interface) as oxygen molecules are incorporated into the lattice to form SiO2. Due to the
injection of interstitial defects during oxidation, you can enhance dopant diffusivities. To
properly simulate this effect, you must include the creation and movement of point defects,
vacancies and interstitials in the simulation. By specifying TWO.DIM in the METHOD
statement before the oxidation step, non-equilibrium point defect concentrations (including
injection and recombination at the Si/SiO2 interface) are included in the simulation. For more
information on point defect diffusion kinetics, see Section 3.1 “Diffusion Models”.
Note: Figure 3-19 compares the boron concentration profiles after an oxidation step when point defects remain at
their equilibrium values (FERMI) and when point defects are allowed to obtain non-equilibrium values
(TWO.DIM). It is evident from Figure 3-19(a) that boron diffusion is enhanced for the TWO.DIM case. The
corresponding interstitial concentrations are shown in Figure 3-19(b). The interstitial concentration is above
the equilibrium interstitial concentration for the TWO.DIM case (thus, allowing oxidation enhanced diffusion
to be observed) but remains at equilibrium for the FERMI case.
Figure 3-19 (a) Boron Concentration Versus Depth (b) Corresponding Interstitial Concentration Versus Depth
You can also have a diffusion retardation effect during thermal oxidation. For dopants
diffusing primarily through a vacancy mechanism, you can reduce their diffusivities during
oxidation because of the recombination of vacancies with injected interstitials at the SiO2/
Silicon interface. Figure 3-20 shows an example of this phenomenon.

Figure 3-20(a) shows the resulting antimony concentration profiles after an oxidation step
where the FERMI and TWO.DIM models were used. In contrast to boron (Figure 3-19(a)), the
resulting antimony concentration profile is shallower for the TWO.DIM case when compared to
the FERMI case.
Figure 3-20(b) shows the reduced vacancy concentration, displaying the results from the
TWO.DIM and FERMI models.
Figure 3-20 (a) Antimony Concentration Versus Depth (b) Corresponding Vacancy Concentration Versus
Depth

Silicidation Model SSuprem4 Models
3.4 Silicidation Model

Silicide modeling capability is implemented in SSuprem4. Silicides are formed when a metal
reacts with silicon or polysilicon to create an intermediate phase. The conductivity of silicides
is typically orders of magnitude greater than that of highly doped n+ and p+ regions. Modern
CMOS technologies use silicides to reduce contact and interconnect resistances. Also, the use
of SALICIDE technology (self-aligned silicides) is a practical way to reduce poly gate
resistance and source and drain sheet resistance.
Silicidation process is invoked by depositing refractory metal layers on the exposed silicon/
poly surface and then specifying a thermal cycle in the DIFFUSE statement. There are four
standard refractory metals in Athena: Titanium, Tungsten, Platinum, and Cobalt.
Corresponding silicides are called TISIX, WSIX, PTSIX, and COSIX. User-defined metal
and corresponding silicide can be also specified using parameters MTTYPE and /MTTYPE in the
SILICIDE statement.
The modeling and understanding of silicide growth is nowhere near as developed as for
oxidation. But, you can consider simulation of silicidation process similar to that of oxidation.
It starts with insertion of a thin (0.002 microns) initial layer of silicide on the boundary
between silicon (or polysilicon) and corresponding metal. During each time step, growth
velocities are calculated for each point at both metal-silicide and silicon (or polysilicon)-
silicide interfaces. The growth velocity at the ith interface point is calculated as follows:
dx i n in
------- = k i C i --------
- 3-172
dt N 1i
where ki is the interface reaction rate coefficient, N1i is the number of silicon or metal
molecules per unit silicide material, and Ci is the silicon or metal concentration. nin is the
interface normal vector which points towards the silicon-poly or metal side. Similarly to
oxidation, this equation can be solved by applying an initial boundary condition xi = x0 at t =
0. The solution is
2
xi xi
- = t+
------- + ---------- 3-173
B BA
where parameters B =  2D i C i   N 1i and B  A =  k i C i   N 1i are equivalent to Deal-

Grove coefficients of classical oxidation model. The silicide growth data indicates that for
most silicides the rate limiting step is diffusion of silicon. This simplifies the Equation 3-174
to
2
xi = Bt 3-174
The silicide growth rates parameters are extracted from experimental data for TiSi2 [44], [45]
and CoSi2 [46]. For two other standard silicides PTSIX and WSIX as well as for user-defined
Silicides the TiSi2 growth rates are used. The silicide growth rates can be modified by
varying parameters D.0 and D.E for silicon (or interstitial) diffusivity in silicide Di, which are
specified in the INTERSTITIAL statement. Defaults are D.0=1.96 and D.E=1.81 for all
silicides.

Silicidation Model SSuprem4 Models
Silicide formation usually leads to a large volume decrease. The ratio between consumed
volumes of silicon and metal and resultant volume of silicide are specified by ALPHA
parameters in the SILICIDE statement. The default values for the ALPHA parameter are taken
from [47].
The 2D movement of growth interfaces and volume change cause the viscous flow of the
silicide layer. This silicide flow is modeled analogously to the compress model of oxidation,
where the equations solved are:
2
 V =  P 3-175
  V = –  --------------- P
1 – 2v
3-176

E
 = -------------------- 3-177
21 + v
where:
• V is the velocity.
• P is the pressure.
• µ is the viscosity.
• v is Poisson’s ratio.
• E is Young’s modulus.
The parameters v and E are specified through elastic constants as described in Section 3.10
“Stress Models”.

Ion Implantation Models SSuprem4 Models
3.5 Ion Implantation Models

Athena uses analytical and statistical techniques to model ion implantation. By default, the
analytic models are used. Analytical models are based on the reconstruction of implant
profiles from the calculated or measured distribution moments. The statistical technique uses
the physically based Monte Carlo calculation of ion trajectories to calculate the final
distribution of stopped particles.
3.5.1 Analytic Implant Models

Athena uses spatial moments to calculate ion implantation distributions. This calculation
method is based on range concepts from “Range Concepts and Heavy Ion Ranges” [48] in
which an ion-implantation profile is constructed from a previously prepared (calculated or
measured) set of moments. A 2D-distribution could be essentially considered a convolution
of a longitudinal (along the implant direction) 1D-distribution and a transverse (perpendicular
to implant direction) 1D-distribution.
In the rest of this section, we will first describe three 1D implant models and the method used
to calculate 1D profiles in multi-layered structures. Then, two models of transverse (lateral)
distribution and a method of construction of 2D implant profiles will be outlined. Finally,
three methods of implant parameter specification will be described.
Gaussian Implant Model
There are several ways to construct 1D profiles. The simplest way is using the Gaussian
distribution, which is specified by the GAUSS parameter in the IMPLANT statement:
2
– x – Rp 
 - exp --------------------------
C  x  = --------------------- 3-178
2  R  2R p
2
where  is the ion dose per square centimeter specified by the DOSE parameter. Rp is the
projected range.  Rp is the projected range straggling or standard deviation.
Pearson Implant Model
Generally, the Gaussian distribution is inadequate because real profiles are asymmetrical in
most cases. The simplest and most widely approved method for calculation of asymmetrical
ion-implantation profiles is the Pearson distribution, particularly the Pearson IV function.
Athena uses this function to obtain longitudinal implantation profiles.
The Pearson function refers to a family of distribution curves that result as a consequence of
solving the following differential equation:
df  x  = -----------------------------------------
------------  x – a f  x  - 3-179
dx 2
b0 + b1 x + b2 x
in which f(x) is the frequency function. The constants a, b0, b1 and b2 are related to the
moments of f(x) by:
 Rp    + 3 
a = – -------------------------------- 3-180
A

2 2
R p 4 – 3 
b 0 = – ----------------------------------------- 3-181
A
b1 = a 3-182
2
2 –  – 6 3-183
b 2 = – ---------------------------
A
2
where A = 10  - 12  - 18,  and  are the skewness and kurtosis respectively.
These Pearson distribution parameters are directly related to the four moments
(  1,  2,  3,  4 ) of the distribution f(x):
3 4
Rp = 1  Rp = 2  = ---------
-  = ---------
- 3-184
3 4
Rp Rp
 i is given by:

1 =  xf  x  dx 3-185
–
  x – R p  f  x  dx
i
i = i = 2, 3, 4 3-186
–
The forms of the solution of the Pearson Differential Equation depend upon the nature of the
roots in the equation b0+b1x+b2x2=0. There are various shapes of the Pearson curves. You
can find the complete classification of various Pearson curves found in “Atomic and Ion
Collision in Solids and at Surfaces” [49]. Obviously, only bell-shaped curves are applicable to
ion implantation profiles. It is readily shown by Ashworth, Oven, and Mundin [50] that f(x)
has a maximum when b0+b1x+b2x2 < 0. You can reformulate this as the following relation
between  and  :
 2 6 4 2 1  2
9   6  + 5  +   9    16 + 8  + 25   + 1   
 -
 > ------------------------------------------------------------------------------------------------------------------------------- 3-187
2
50 – 
2
with the additional constraint that  <50.

Only Pearson type IV has a single maximum at x = a+Rp and monotonic decay to zero on
both sides of the distribution. Therefore, Pearson type IV is usually used for ion implantation
profiles; it is the solution of Equation 3-178 when the following conditions are satisfied:
2 2 32
39  + 48 + 6   + 4  2
 = ---------------------------------------------------------------- and 0 <  < 32 3-188
2
32 – 
This gives the following formula for Pearson IV distribution:
1- b1
-------- ------ + 2a
2b 2 b2  2b  x – R  + b 
atan  -------------------------------------------
2 2 p 2
f  x  = K b0 + b1  x – Rp  + b2  x – Rp  exp – ----------------------------------- 3-189
 
4b 1 b 2 – b 1
2  4b b – b 2 
1 2 1
where K is defined by the constraint:

 f  x  dx = 1 . 3-190
–
2
In the narrow area of  –  plane where Pearson IV type criterion (Equation 3-188) is not
satisfied while bell-shaped profile criterion (Equation 3-187) holds Athena, by default, uses
other than type IV Pearson functions. These functions are bell-shaped but they are not
specified over the whole (  – ,   ) interval. Usually, this doesn’t affect the quality of
calculated profiles because the limits of these functions are situated far from their maximums.
If you want to use only Pearson-IV distribution, set the ANY.PEARSON parameter to FALSE. In
all cases when  and  do not satisfy one of the mentioned criteria, Athena will
automatically increase  up to the value that satisfies the criterion used. In the standard
Pearson model, the longitudinal dopant concentration is proportional to the ion dose  :
Cx = fx 3-191
This single Pearson approach (method) has been proved to give an adequate solution for
many ion/substrate/energy/dose combinations. But, there are many cases when the channeling
effects make the Single Pearson Method inadequate.

Dual Pearson Model

To extend applicability of the analytical approach toward profiles heavily affected by
channeling, Al Tasch [51] suggests the dual (or Double) Pearson Method. With this method,
the implant concentration is calculated as a linear combination of two Pearson functions:
C  x  = 1 f1  x  + 2 f2  x  3-192
where the dose is represented by each Pearson function f1,2(x). f1(x) and f2(x) are both
normalized, each with its own set of moments. The first Pearson function represents the
random scattering part (around the peak of the profile) and the second function represents the
channeling tail region. Equation 3-191 can be restated as:
C  x  =    f 1  x  +  1 –  f 2  x   3-193
where  =  1 +  2 is the total implantation dose and  =  1   .
To use dual Pearson distribution, supply nine parameters— four moments for each Pearson
function with the dose ratio  . The dual Pearson model will be used only when all nine
parameters are present (see “Specification of Implant Parameters in the Moments Statement”
on 206) and the AMORPHOUS parameter is not specified in the IMPLANT statement (the
default is CRYSTAL). Otherwise, the single Pearson formula will be used.
SIMS-Verified Dual Pearson (SVDP) Model
By default, Athena uses SIMS-Verified Dual Pearson (SVDP) implant models. These are
based on the tables from the University of Texas at Austin. These tables contain dual Pearson
moments for B, BF2, P, and As extracted from high quality implantation experiments are also
conducted by the University of Texas at Austin. Table 3-7 show these implantation tables
contain dose, energy, tilt, rotation angle, and screen oxide thickness dependence.
Table 3-7 Range of Validity of the SVDP Model in Athena
Ions Energy (keV) Dose (cm-2) Tilt Angle(°) Rotation Angle(°) Screen Oxide (Å)
B 1 100a 1013  0  10 0  360 native oxide  500b

81015
BF2 1 80c 1013  0  10 0  360 native oxide
81015
P 12 200d 1013  0 10 0  360 native oxide
81015
As 1 200e 1013  0  10 0  360 native oxide
81015
a Experimentally verified for 5-80keV. For energy range, 1-5keV, an interpolation between 5keV and 0.5keV calculated with UT-MARLOWE,
is used; an extrapolation is used for energy range 80± 100keV.
b Only for 15-80keV.
c Experimentally verified for 5-65keV. For energy ranges, 1-5keV and 65-80keV, the same procedures is used for boron.
d Experimentally verified for 15-80keV. Numerical extrapolation is outside this energy range.
e Experimentally verified for 5-180keV. Interpolation between 5keV and UT-MARLOWE calculated profile at 0.5keV.

If you choose a simulation outside the parameter ranges, described in Table 3-7, Athena will
not use the Dual Pearson Implant SVDP Models but will use the standard tables instead.
When using the Dual Pearson model, remember the following:
• For implant energies below 15keV, for boron, BF2 and arsenic, the simulation predicts the
dopant profiles for implants into a bare silicon surface (i.e., silicon wafer subjected to an
HF etch less than two hours before implantation). Low energy implant profiles at such
low implant energies are found to be extremely sensitive to the presence of a thin (0.5-
1.5nm) native oxide layer or disordered silicon layer on the wafer surface [52].
Remember this fact when using the model for the simulation of low energy ion
implantation and when performing implantations.
• For implant energies between 10keV and 15keV, the simulations are performed for boron,
BF2, and arsenic by using an interpolation between the Dual Pearson Model parameters at
15keV, and the Dual Pearson Model parameters at 10keV. The parameters at 15keV
correspond to implantation through a native oxide layer (0.5-1.5nm), while the parameters
at 10keV correspond to implantation into a bare silicon surface (i.e., silicon wafer
subjected to an HF etch less than two hours before implantation).
• For implant energies below 5keV, the models for boron, BF2, and arsenic have not been
verified experimentally. The simulations in this range of implant energy are performed
using an interpolation between experimentally verified Dual Pearson parameters at 5keV
and parameters based on UT-MARLOWE estimates at 0.5keV.
• The SIMS measurements upon which these profiles are based have a concentration
15 16 –2
sensitivity limit in the order of 5  10 to 2  10 cm , increasing with dose from
the implant. The profiles have been extended below these limits, following the trends that
occur within the sensitivity limits of the SIMS.
• The screen oxide thickness range has been verified from 1.5 to 40nm (only for boron and
15-80keV energy range). But the oxide range has been extended to 50nm.
Screen Oxide Thickness Parameter (S.OXIDE)
To specify screen oxide, use the S.OXIDE (µm) parameterin the IMPLANT statement. This
thickness is specified independently of any actual surface oxide in the structure. It is,
however, possible to automate the extraction of the surface oxide thickness for use with the
IMPLANT statement. An example is supplied demonstrating this.
S.OXIDE is another parameter for dual Pearson moments selection from the implant tables.
It’s up to you to select its value accordingly. The effect of this parameter is that it represents
ion implantation through a thin (050nm) surface oxide layer. The present algorithm in
Athena when encountering a multi-layered structure (see Section 3.5.2 “Multi-Layer
Implants”). For example, oxide/silicon switches to multi-material scaling technique for
evaluating the depth profile. This technique will combine two profiles — single Pearson for
the oxide and dual Pearson for silicon with S.OXIDE preferably set to the thickness of the
oxide.
There are two reasons why this separation between the surface oxide is present in the
structure before the IMPLANT statement and the S.OXIDE parameter. The first reason is
because the flexibility of using this parameter for different thin surface layers other than oxide
(with appropriate scaling of their thickness for stopping). The second reason the currently
restricted availability of moments with screen oxide in the tables (050nm, 1580keV and for
boron only). If you need a more precise dependence of the implantation profiles on the
surface screen oxide, use a single layer of silicon with S.OXIDE set to an appropriate value.

3.5.2 Multi-Layer Implants

To apply any of the described analytical distribution functions for structures that are
comprised from several different material layers, use a scaling method that’s mentioned in
this section. This is because stopping powers and range parameters are different in different
materials. This section will describe the implant scaling methods available in Athena.
DOSE.MATCH
The Dose-Matching Method was historically the first and is the most widely used [53]
method. The Dose-Matching Method is selected by the DOSE.MATCH parameter (default) in
the IMPLANT statement. With this method, the segment of the profile within ith layer is
calculated by:
C i  x  =  f  x – x t + x eff  3-194
where f(x) is the distribution function specified for this implant (Gaussian, Pearson, or Dual
Pearson) with moments corresponding to the ith layer; xt is the distance from the surface to the
top of the ith layer:
ki–1
xt =  tk 3-195
k=1
xt is the thickness of the kth layer; xeff is the effective thickness evaluated from:
x eff ki–1
 C i  x  dx =
 k 3-196
0 k=1
where  k is the portion of the total implant dose, which is consumed in the kth layer.
Obviously, for the first layer xeff = 0 and xt = 0.
RP.SCALE and MAX.SCALE
The other two methods for analytical calculation of implantation profiles in the layered
structures are projected range depth scaling (set by RP.EFF or RP.SCAL in the IMPLANT
statement) and maximal depth scaling (set by the MAX.SCALE parameter). These two methods
differ from the dose-matching method in the way the effective depth xeff is calculated and in
the normalization of the partial profiles in the layers. Like in the dose-matching method, the
distribution in the first layer is calculated directly from the moments corresponding to the first
layer without any corrections. For subsequent layers, the implant distribution is calculated by
the formulae:
C i  x  = Nf  x – x t + x eff  3-197
and
 i–1 
 
N= –
  
 k   i 3-198
 k=1 

where N is the normalization factor,  is the total implantation dose, and xeff is the effective
depth calculated as follows. In the case of projected range scaling, xeff for the ith layer is:
i–1
tk k + 1
x eff =  --------
Rp
k p
-R 3-199
k=1
k
where R p is the projected range of the specified ion in the material of the kth layer. For the
case of the maximal range scaling, xeff is calculated as:
i–1
tk
-  Rp
k=1 

k+1
x eff = --------------------------------- + 3R p 3-200
k k  
Rp + 3  Rp
k=1
k
where  R p is the projected range straggling in the kth layer. In this approximation, the
estimated maximum ion range R p + 3  R p is taken as the measure of the ion penetration into
the corresponding material.
MOM.SCALE
In all three models described above, the range parameters in each layer are considered
independent of the presence of other layers. But obviously, the distribution of ions stopped in
the deeper layers may depend on the thickness and stopping characteristics of the upper layers
because each ion trajectory passes through these upper layers. The Moment Correction
Method set by the MOM.SCALE parameter of the IMPLANT statement partially accounts for this
effect. In the SCALE.MOM, method the projected range and range straggling in the layer are
normalized according to the probability for the ion to penetrating into the layer. The only
available measure of the probability is the portion of the dose accumulated in the specific
layer. Therefore, the corrected projected range Rpc and range straggling  R pc in the ith layer
are calculated as follows:
i–1

i k
R pc =  i Rp +  k Rp 3-201
k=1
i–1

i k
R pc =  i Rp +  k Rp 3-202
k=1
where
i–1
i =  –  k 3-203
k=1
You can use the SCALE.MOM parameter together with any of three depth matching methods.

3.5.3 Creating Two-Dimensional Implant Profiles

Convolution Method
Athena calculates 2D implant profiles using a convolution method described as follows. First,
it calculates the implantation direction within the simulation plane using the TILT  and
ROTATION  angle parameters specified in the IMPLANT statement.  is the angle between
the ion beam direction and y-axis,  is the angle between ion beam direction and the
simulation plane. For example,  = 0 and  > 0 correspond to an ion beam parallel to the
simulation plane and directed toward the lower-right corner of the simulation area. The case
of  = 90 and  > 0 correspond to an ion beam in the plane perpendicular to the simulation
plane and directed from behind the simulation plane. The effective implantation angle in the
simulation plane could be found from
tan    = tan    cos    3-204
When the FULLROTAT parameter is specified in the IMPLANT statement Athena calculates
superposition of 24 implants with rotation angles equal to (15n) and doses equal to  /24.
The implantation front (perpendicular to the  direction) is divided into a number of slices Ns
(usually > 100) of width a. The implant concentration in each grid point, i, with coordinates
(xi, yi) is calculated by the summation of contributions from each slice, k:
C  x i, y i  =  C k  x i, y i  3-205
1  k  Ns
The contribution from each slice Ck is calculated by integration of the point source 2D
frequency function F2D(x,y) (with the starting point at the intersection of the normal n to the
central of the slice with the structure surface) over slice width:
a2
C k  x i, y i  = 
 f 2D  d i, t i  dt 3-206
–a  2
where di is the depth along implant direction (i.e., distance between the starting point and the
projection of the point i on the vector n) and ti is the transversal distance (i.e., distance
between the point i and the vector n). See Figure 3-21.

Figure 3-21 Integration Geometry for the Convolution Method

Depth-Independent Lateral Distribution
The simplest type of the 2D frequency function is a product of longitudinal function fl(x),
which can be a Gaussian (Equation 3-178), Pearson (Equation 3-191), Dual-Pearson
(Equation 3-193), and depth-independent transversal function ft(y). See Equation 3-207.
f 2D  x, y  = f l  x f t  y  3-207
This approximation is used in Athena by default. Obviously, the function ft(y) must be
symmetrical and have a bell shape.
Gaussian Lateral Distribution Function
The traditional selection for this function is a Gaussian. Athena uses the Gaussian
approximation unless the transversal kurtosis  y (KURTT in the MOMENTS statement) is
specified to be different from its default value of 3.0. In this case, Equation 3-206 can be
easily integrated into the following equation:
ti – a  2 ti + a  2
C  x i, y i  = 1---  f l  d i  erfc ------------------
- – erfc -------------------
- 3-208
2 2Y 2Y
where Y is the transversal (lateral) standard deviation defined from:


  f2D  x, y y
2
Y = dx dy 3-209
–

Specification of Lateral Standard Deviation

You can specify Lateral Standard Deviation (LSTD.DEV or LDRP) together with other
moments in the MOMENTS statement (see “Specification of Implant Parameters in the Moments
Statement” on 206). You can also control it with the LAT.RATIO1 parameter in the IMPLANT
statement. LAT.RATIO1 is the ratio between Y and R p which is equal to 1.0 by default.
This means that if the lateral standard deviation and LAT.RATIO1 are not specified it will be
equal to projected range straggling R p . In the case of dual Pearson model for longitudinal
profile, corresponding parameters, SLSTD.DEV or SLDRP and LAT.RATIO2 are used. The only
difference is that the default for LAT.RATIO2 is 0.2. This is because the channelled portion of
a 2D profile is obviously very narrow.
Parabolic Approximation of Depth-Dependent Lateral Distribution
It has been shown [54], [55], [56], and [57] that in general, the transversal function ft(y) is not
independent of depth because there is considerable correlation between transversal and
longitudinal motion of the implanted ions. This correlation could be taken into account by
using a transversal function with the depth-dependent lateral standard deviation  y  x  . As it
was shown in [55] and [57], if the spatial moments up to fourth order are used, the best
approximation for  y  x  is the parabolic function:
2 2
 y  x  = c0 + c1  x – Rp  + c2  x – Rp  3-210
In order to find the coefficients of the function two additional spatial moments should be
used. These are mixed skewness:

  f2D  x, y   x – Rp y
2
 xy = dx dy 3-211
–
and mixed kurtosis:

2 2
 xy =   f2D  x, y   x – Rp  y dx dy 3-212
–
The c0, c1, and c2 parameters can be found by substituting Equation 3-207 into Equations 3-
209, 3-211, and 3-212 and taking into account in Equations 3-180—3-186, while integrating
over x. This results in the system of equations where you can find the following relations.
2
c0 =  Y  1 – B  3-213
2
Y   – B 
c 1 = ---------- 3-214
 R p xy
2
Y -B
c 2 = ------------- 3-215
2
R p

where:
 xy – 1 –  xy
B = ----------------------------------
- 3-216
2
–1–
This parabolic approximation for depth-dependent ft will be used if the FULL.LAT parameter
is used in the IMPLANT statement and when mixed spatial moments:  xy (SKEWXY parameter)
and  xy (KURTXY parameter) are non-zeros. In the case of the Dual Pearson longitudinal
function, the mixed spatial moments for the second Pearson, SSKEWXY and SKURTXY, can be
also specified. The values of spatial moments are not yet included in the default moments
tables and should be specified in the MOMENTS statement (see “Specification of Implant
Parameters in the Moments Statement” on 206).
Non-Gaussian Lateral Distribution Functions
Detailed Monte Carlo simulations [58] and [59] also show that in most cases, transversal
distribution function, ft, is not Gaussian. In other words, the transversal kurtosis  y is
calculated as

 f2D  x y y
4
y = dx dy 3-217
–
and is not always equal to 3.0 and also depends on depth. Several non-Gaussian transversal
distribution functions were examined in [49]. Their conclusions were as follows. The
symmetrical Pearson functions (type II for  y  3 and type VII when  y > 3) are acceptable,
providing an agreement with amorphous Monte Carlo simulations and have computational
advantage because they can be integrated over x in a closed form through incomplete beta
functions [60].
Another good alternative for transversal distribution function is the Modified Gaussian
Function (MGF) suggested in [58]. It is shown in [60] that it also can be integrated in the
close form through the incomplete gamma function. Selection of transversal distribution
function is subjective because it is based on comparison with the lateral cross-section of the
2D Monte Carlo distributions, which cause accuracy to diminish further away from its
maximum. The analysis of [59] based on the BCA simulation (see Section 3.5.4 “Monte
Carlo Implants”) showed that when  y  2.5  2.8 , which usually happens for random part
of the 2D distribution or for amorphous implants, the Pearson type II function slightly
underestimates concentrations obtained in the BCA calculations while the MGF slightly
overestimates these concentrations. Therefore, it was decided to use in Athena an average
between the Pearson type II and the MGF for all  y < 3 . When  y = 3 both functions reduce
to standard Gaussian. Finally, in the case of higher values of lateral kurtosis it was found [59]
that the MGF appears to be a better approximation, so it is used in Athena.
It is very difficult to find  y  x  as was done for  y  x  already mentioned, because the spatial
moments of fifth and sixth order would be needed to build analytical functions for  y  x  .
Therefore, Athena uses constant  y (the KURTT and SKURTT parameters for the first and
second Pearson functions, correspondingly) when you specify the FULL.LAT model in the

IMPLANT statement. The generic approximations [59] for  y  x  (instead of Equation 3-210)
and for  y  x  will be implemented in future when more complete tables of lateral parameters
will be generated using Monte Carlo simulations.
Specification of Implant Parameters in the Moments Statement
As mentioned previously, the analytical ion implantation simulations strongly depend on the
input parameters (moments). Athena provides several ways of implant parameter
specification. They are: look-up tables, user-defined look-up tables, and the MOMENTS
statement.
Two types of look-up tables are currently provided with Athena. The files containing the
tables are in ASCII format and can be found in the <install>/lib/athena/<>version>/
common/implant-tables directory. The first type are standard tables (std-tables file)
containing parameters for most ion-material combinations used in Athena. These are
longitudinal parameters for the single Pearson distribution in the energy interval 10keV to
1MeV. The energy interval is extended to 1keV8MeV for B, P, and As in silicon, silicon
oxide, polysilicon, and silicon nitride. These tables also include a limited set of parameters for
the dual-Pearson function (only for B and BF2 in the energy interval 10100keV, tilt angle 7,
rotation angle 30, and native oxide as well as with simple interpolation of the dose ratio
parameter  between different doses). Parameters for the FULL.LAT model are provided only
for B implants in silicon. They are based on the spatial moment calculations in amorphous
silicon as in [61], [62]. The auxiliary file userimp in the <install>/lib/athena/<version>/
common directory provides a template for specifying implant parameters in the format of
standard tables. The second type of look-up tables are SVDP tables described in the “Dual
Pearson Model” on 198. The format of these tables is much more flexible than the format of
the standard tables. It also allows parameters for lateral distribution to be added easily.
The SVDP tables are used by default. If no moments are found, Athena will search through
standard tables. If it cannot find parameters for a specified energy/ion/material combination, a
warning message is issued, which will tell you a very small projected range and straggling
will be used in simulation for this combination. The message will also suggest that you use
the Monte Carlo method in order to find the right moments. This is the sequence of Athena
actions in the case when no MOMENTS statement precede the current IMPLANT statement. The
MOMENTS statement serves to control the moment parameters/tables to be used in subsequent
IMPLANT statements. If you specify the STD_TABLES parameter, Athena will skip searching
through SVDP tables and proceed directly to the standard tables. If you specify the
USER_STDT or USER_SVDPT parameter, then the user-defined file specified with the
USER.TABLE=<filename> parameter will be used as the first choice. Of course, if the
moments are not found in the specified file, Athena will proceed to the standard tables.
Finally, the set of MOMENTS statement can be used to specify all spatial moments for any ion/
material/energy/dose combination. Athena will use parameters from this set before
proceeding to a standard search sequence. If the moments for certain implant conditions are
unavailable, then you can use the Monte Carlo simulation for these conditions.

Using PRINT.MOM for Extraction of Spatial Moments

The PRINT.MOM parameter in the IMPLANT statement prints the calculated (or extracted from
the tables) moments into output and also saves the moments in the standard structure file. The
last capability allows you to use the extract (EXTRACT statement) and substitution functions of
Deckbuild for automatic generation of the MOMENTS statement. If spatial (lateral and mixed)
moments need to be found from Monte Carlo calculation, use the IMPCT.POINT parameter
because it forces all trajectories to be started in one point. This not only allows the spatial
moments to be found, but also the building of a Monte Carlo calculated source-point 2D
distribution function, which can be useful for comparison purposes.
3.5.4 Monte Carlo Implants

The analytical models described in the previous section give very good results when applied
to ion implantation into simple planar structures (bare silicon or silicon covered with thin
layer of other material). But for structures containing many non-planar layers (material
regions) and for the cases, which have not been studied yet experimentally requires more
sophisticated simulation models. The most flexible and universal approach to simulate ion
implantation in non-standard conditions is the Monte Carlo Technique. This approach allows
calculation of implantation profiles in an arbitrary structure with accuracy comparable to the
accuracy of analytical models for a single layer structure.
Athena contains two models for Monte-Carlo simulation of ion implantation: Amorphous
Material Model and Crystalline Material Model. Both of them are based on the Binary
Collision Approximation (BCA) and apply different approximations to the material structure
and ion propagation through it.
Nature of the Physical problem
A beam of fast ions (energy range, approximately 50 eV/amu to 100 keV/amu) entering
crystalline or amorphous solid is slowed down and scattered due to nuclear collisions and
electronic interaction. Along its path, an individual projectile may create fast recoil atoms that
can initiate collision cascades of moving target atoms. These can either leave the surface (be
sputtered) or deposited on a site different from their original one. Together with the
projectiles being deposited in the substrate, this results in local compositional changes,
damage creation and finally amorphization of the target. Depending on the crystal orientation
or the direction of the beam or both, the implanted projectiles and the damage created by them
has different spatial distribution. With even more higher fluency, these phenomena will cause
collisional mixing in a layered substances, changes of the surface composition due to
preferential sputtering, and the establishment of a stationary range profile of the implanted
ions.

Method of Solution
The paths of the individual moving particles and their collisions are modeled by means of the
binary collision approximation for a crystalline, polycrystalline and amorphous substance,
using a screened Coulomb potential for the nuclear collisions and a combination of local and
non-local free-electron-gas approximation for the electronic energy loss. For each nuclear
collision, the impact parameter and the Azimuthal Deflection Angle are determined according
to the crystal structure using its translational symmetry. For amorphous materials, the impact
parameter and the azimuthal deflection angle are determined from random numbers. A proper
scaling is chosen so that each incident projectile (pseudo-projectile) represents an interval of
implantation dose. Subsequent to the termination of each pseudo-projectile and its associated
collision cascades, the local concentrations of the implanted species, created vacancies and
interstitials are calculated according to the density of the matrix.
Nuclear Stopping
As mentioned before, during their passage through matter ions interact not only with the
atoms from the lattice but also with the electrons. Figure 3-22 shows the scattering geometry
of two particles in the Laboratory Co-ordinate System. In the computational model, it is
assumed that ions from one deflection point to the next move along straight-line segments,
these being the asymptotes of their paths. At each collision, ion loses energy through
quasielastic scattering by a lattice atom and by an essentially separate electron energy loss
part.
Figure 3-22 The trajectories of the ion (projectile) and the lattice atom (recoil).
The scattering angles of the projectile and the recoil are as follows:
tan  1 = Af sin    1 + Af cos   3-218
tan  2 = f sin    1 + f cos   3-219

where:
f = 1 – Q  Er 3-220
Q is the energy lost by electron excitation.

A = M 2  M 1 is the ratio of the mass of the target (scattering) atom to that of the projectile
(implanted ion).
 is the barycentric scattering angle calculated as follows:

R ---------------
 =  – 2p 1 - dr 3-221
2
r gr
where:
2
gr = 1 – p------ – V  r -
----------
2 Er
r
where:
• p is the impact parameter,
• E r = AE 0   1 + A  is the relative kinetic energy,
• E 0 is the incident energy of the projectile,
• r is interatomic separation,
• V  r  is the interatomic potential,
• R is defined from equation g  R  = 0 .
In Athena, the intersections of the incoming and outgoing asymptotes are evaluated with the
hard core approximation of the time integral:
x 1 = p tan    2  3-222
x2 = 0 3-223
Interatomic Potential
Athena uses two-body screened Coulomb potentials with a screening function, which is a
numerical fit to the solution given by Firsov [63]. It also preserves the same analytic form as
for the isolated atom:
2
Z1 Z2 e
V  r  = -------------------   ------
r
3-224
r a 
0
where Z1 and Z2 are the atomic numbers of the two atoms and a 0 is the screening length
defined by

–1  3
a 0 = 0.8853a B Z 3-225
where Z is an ‘average’ atomic number of the two atoms calculated as
–1  3 –1
=  Z1
0.23 0.23
Z + Z2 M . 3-226
 
The main drawback of these two-body potentials is their relatively slow decay as r   . The
screening parameter, a 0 , is often regarded as an adjustable parameter for each two-body
combination, which can be matched either to self-consistent field calculations or to
experimental data. Athena uses the screening function in the form
4
 =  ai exp  –bi x  3-227
i=1
where a i and b i are taken from [64].
Electronic Stopping
Electronic stopping used in the simulation consists of two essentially separate mechanisms
for inelastic energy losses: local and non-local. These two types of electronic stopping are
quite different in nature and behavior -- they have different energy and spatial dependencies
[65]. The local inelastic energy losses are based on the model proposed by Firsov [66]. In this
model, the estimation of the electronic energy loss per collision is based on an assumption of
a quasi-classical picture of the electrons (i.e., the average energy of excitation of electron
shells, and electron distribution and motion according to the Thomas-Fermi model of the
atom).
In this quasi-classical picture, the transfer of energy,  E , from the ion, to the atom, is due to
the passage of electrons from one particle to the other. Thus, resulting in a change of the
momentum of the ion (proportional to its velocity ,  , and a rising of a retarding force acting
on the ion). When ions move away from the atom (despite being trapped by ions) electrons
will return to the atom. There is no transfer of momentum calculated back, because the
electrons fail in higher energy levels. The energy loss in the Firsov's Model is calculated as
follows:
53
0.05973   Z 1 + Z 2  E  M1
–  E = ----------------------------------------------------------------------------eV 3-228
13 5
 1 + 0.31  Z 1 + Z 2  R 0 
where:

• R 0 is their distance of closest approach in A , which is approximately equal to the
impact parameter in case of small-angle collisions.
• E is the energy of the moving atom (the ion) in eV.
• M 1 is its mass in a.m.u.

In a binary collision, the scattering angles are affected by the inelastic energy loss  E (see
Equation 3-228) through the parameter f .
The non-local electronic energy losses are based on the model proposed by Brandt and
Kitagawa [67]. Their stopping power, S = – dE

------- , of the medium for an ion is in the first
dx
approximation proportional to a mean-square effective ion charge. They derive the effective
stopping power charge of a projectile, Z1 from a given ionization state, q. If a fractional
effective charge of an ion with the given ionization state, q is defined as
Z S 12
 ½ ------- q
1 - = ------------
- 3-229
Z1 Sq = 1
where S q = 1 is the stopping power for bare nucleus. Brandt and Kitagawa theories produces
the following simple expression for the fractional effective charge of an ion:
2
  q + C  k F   1 – q  ln  1 +  2  v F a 0 v 0   3-230
where:
• q =  Z 1 – N   Z 1 is the fractional ionization,
• N is the number of electrons still bond to the projectile nucleus,
• a 0 and r 0 are Bohr's radius and velocity,
• k F and v F are Fermi wave vector and velocity.
For the screening radius  , Brandt and Kitagawa assume exponential electron distribution,
which becomes:
23
0.48N
 = -----------------------------------------
- 3-231
Z 1  1 – N   7N 1  
The only undefined quantity, C, is of about 0.5 and somewhat depends on the target. The
degree of ionization, q, can be expressed as
 – 0.92v r
q = 1 – exp  ------------------
23
- 3-232
 v0 Z1 
where v r   v 1 – v e  is the relative velocity between the projectile and the target electrons,
which are calculated as follows:
2
3v  2v 1  1  v 1 
v r = --------F 1 +  ----------
- – ------ ----- for v1 < vF 3-233
4  3v2  15  v F
F
 vF 
2
v r = v 1  1 + ----------
- for v1  vF 3-234
 5v 1 
2

Damage Accumulation Model

The present model includes dynamic processes of the transformation from crystalline to
amorphous state as ion implantation proceeds. Each pseudo-projectile in the simulation
represents a portion of the real dose  , where N is the number of projectiles.
-
 = --- 3-235
N
The deposited energy is accounted for each grid point of the target and accumulated with the
number of projectiles. As the implantation proceeds, deposited energy increases and the
crystalline structure gradually turns into an amorphous structure. This is quantified by the
Amorphization Probability Function as follows:
E r
f  r  = 1 – exp --------------- 3-236
Ec
Here,  E  r  is the energy deposited per unit volume at the grid point r, and E c is the critical
energy density, which represents the deposition energy per unit volume needed to amorphize
the structure in the relevant volume.
It is defined as:
 E 0  T – T   – 2
E c  T  = E c0  1 – exp ----------------------------  3-237
 2k B TT  
where E is activation energy, k B is Boltzmann's constant, and T  is the temperature at and

above which the infinite dose is required for crystalline to amorphous transition.
Some experimental values for E c  E c0  are given by F. L Vook [68]. In the BCA module, the
value f(r)=0.6 corresponds to a fully amorphized state and any additional energy deposited at
point r does not contribute to the amorphization process.
Implantation Geometry
Figure 3-23 shows the orientation of the ion beam, relative to the crystallographic orientation
of the substrate. There three major planes regarding ion implantation in crystalline materials,
mainly:
• the implantation plane ,
• the surface plane ,
• and the simulation plane .
The implantation plane is where the initial beam of incoming ions lays in. It equivocally
defines the direction of the incoming beam -- tilt and rotation. If the orientation of the surface
plane is [100], which is the only substrate orientation available currently in the Binary
Collision approximation implantation module (BCA or CRYSTAL parameters), the offset of the
rotation angle is the direction <101> on this plane. This means that the tilt angle, , specified
by the TILT parameter in the IMPLANT statement will be the polar angle in laying this plane,
while the rotation angle, , specified by the rotation parameter will be the difference of
azimuths of the line where the implantation plane, , crosses the surface plane, , and the
direction <101>. See Figure 3-23.

Note: Presently, the surface orientation (the ORIENT parameter in the INITIALIZE statement) does not have any
affect in the crystal Monte-Carlo module and the surface orientation is always {100}.
Figure 3-23 Implantation geometry

The simulation (projection) plane is where all data regarding the simulation is projected on,
which is what finally goes into Athena’s structure. The orientation of the simulation plane is
specified by the ROT.SUB parameter in the INITIALIZE statement.
In summary, the laboratory coordinate system used in the BCA implant simulation is right-
hand-sided -Y is depth, X is the other co-ordinate and Z is from the observer. The azimuth
angle is measured in the X-Z plane. ROT.SUB is relative to X. Simulation plane is X’-Y.
The simulation plane (where TONYPLOT displays results and Athena calculates) is always
perpendicular to the wafer surface and its orientation is specified by the ROT.SUB parameter
in the INITIALIZE statement relative to the wafer major flat. By default, it’s ROT.SUB= =
45°.
The implant calculation in bulk is 3D. Ray-tracing for BCA (i.e., calculation of ion impacts,
scattering from walls and re-implantation) is 2.5D. In other words, structure is infinitely
extended in the third dimension (along Z). All simulation results (doping, damage) are
projected on the simulation plane and appropriately scaled.

ROTATION is measured from the major flat and in Athena’s case from the simulation plane
because it is coupled with the wafer’s major flat.
For silicon (and other crystalline materials), you can think of TILT/ROTATION as always
relative to the simulation plane giving the same shadowing effects while ROT.SUB defines
which direction the simulation plane will slice the crystal structure through (in the laboratory
coordinate system).
Specifying different ROT.SUB will have an effect on channeling. But remember, that ion
propagation is three dimensional and there are some channeling patterns that remain the same
or become stronger/weaker because of favorable/unfavorable initial impact conditions (TILT/
ROTATION/ROT.SUB combination).
Implantation Geometry in SiC
Monte Carlo module of Athena allows accurate simulation of ion implantation in 4H-SiC and
6H-SiC [15]. It is known that channeling effects are very pronounced in SiC since its lattice
has several wide open channels (e.g., <0001>, <11 2 0>, and <11 2 3> for 4H-SiC). Therefore,
implant profiles strongly depend on wafer orientation and ion beam direction. Another
important aspect of ion implantation simulation in SiC is the fact that most of commercially
available standard <0001> wafers are offered with miscut of 3.5°–8.5° usually toward
<11 2 0> direction. The miscut is introduced primarily to facilitate epitaxial growth. In the
same time, the miscut alters ion implantation because implanting normal to the wafer surface
is equivalent to a tilt of the beam in the (11 2 0) plane.
In order to perform Monte Carlo implant simulation in Athena, you should first specify
simulation plane. The Monte Carlo module of Athena simulates ion trajectories in 3D and
then registers the stopping points and consequently doping concentration by projecting 3D
result onto simulation plane. The module also assumes that configuration of material
interfaces don’t change in direction normal to this simulation plane. The simulation plane is
always perpendicular to the wafer surface and its orientation is specified by the ROT.SUB
parameter in the INITIALIZE statement relative to the wafer major flat (see Figure 3-24).
Most commercial 4H-SiC and 6H-SiC wafers have major flat parallel to (1 1 00) plane as in
Figure 3-24.

Figure 3-24 SiC wafer orientation relative to simulation plane in Athena BCA Module.
Also shown are major plane projections on the wafer surface.
So if you want your Athena simulation plane to be parallel to the major flat, <1 1 00>, you
have to specify ROT.SUB=0. If you want your simulation plane to be parallel to <11 2 0>
crystallographic direction, you should specify ROT.SUB=90.
The ion beam direction is specified by the TILT and ROTATION parameters in the IMPLANT
statement. The TILT is the angle (in degrees) between the ion beam vector and the normal to
the wafer surface. The ROTATION is the angle between the ion beam vector plane
perpendicular to the wafer surface and the simulation plane. For example, if the ion beam
vector plane is parallel to < 2 110> crystallographic direction in Figure 3-24, you have to
specify ROTATION=30 (see Figure 3-25). Alternatively, if the ion beam vector plane is
perpendicular to simulation plane, you have to specify ROTATION=90.

Figure 3-25 3D SiC wafer and simulation plane orientation. ROT.SUB=0 for simulation plane parallel to (1 1 00)
The wafer miscut is specified by the MISCUT.TH and MISCUT.PH parameters. MISCUT.TH
specifies the angle between actual wafer surface and the XZ plane (see Figure 1). Parameter
MISCUT.PH specified the angle (measured counter-clockwise from the major flat direction)
toward which the wafer was cut. So if the wafer is cut toward <1100> (i.e., X direction in
Figure 3-24), you need to specify MISCUT.PH=0. This type of miscut is not used very often.
The majority of commercial wafers (see, for example, [16] and [17]) have miscut toward
<11 2 0>. So practically in all cases, you need to set MISCUT.PH=90 and MISCUT.TH=4 or
MISCUT.TH=8, depending of “off axis” angle in commercial wafer specifications (see [16]
and [17]).
So, a typical initialize of Monte Carlo implant statements will be as follows:
init sic_4h rot.sub=0
implant aluminum energy=50 dose=1e13 bca n.ion=10000 tilt=4
rotation=90 miscut.th=4 miscut.ph=90
This will correspond to implant to a wafer with standard miscut usually specified as “Off axis
4° toward <11 2 0> ± 0.5°” in commercial wafer specifications. The simulation plane will be
parallel to the major flat and ion beam will be tilted by 4° in the plane perpendicular to the
simulation plane.
The following examples closely demonstrate equivalent implant directions based on miscut
MISCUT.TH/PH or TILT/ROTATION specifications.

Amorphous Material Monte Carlo

In the doping of semiconductors, the rest distribution of the implantations is of principal
importance. The penetration of ions into amorphous targets is most simply described by using
a Statistical Transport Model, which is the solution of Transport Equations or Monte Carlo
Simulation. Among the two approaches, Monte Carlo is more convenient for multiple
components and two or three dimensional targets, which is partly possible because the Monte
Carlo method treats an explicit sequence of collisions, so the target composition can change
on arbitrary boundaries in space and time.
The rest of the distribution is built up from a vast number of ion trajectories and the statistical
precision of which depends directly on this number:  N . As the ion penetrates a solid, it
undergoes a sequence of collisions with the target atoms until it comes to rest. A simplified
model of this interactions is a sequence of instantaneous binary nuclear collisions separated
by straight line segments (free flight path lengths) over which the ion experiences continuous
(non-local) electronic energy loss. The collisions are separated (i.e., the state of an ion after a
collision depends solely on the state of the ion before the collision).
The model assumes that the arrangement of the target atoms is totally randomized after each
collision (i.e., the target has no structure and no memory). As a result, a sequence of collisions
is described by randomly selecting the location of the next collision partner relative to the
pre-flight location and velocity direction of the ion. This means that this model cannot
simulate the anomalous tail penetration observed for implanted ions into aligned single
crystal targets. The model adequately describes the ion penetration into multilayer non-planar
structures.
Crystalline Material Monte Carlo
The crystalline model used in Athena is based on the program CRYSTAL described elsewhere,
[69]. In order to calculate the rest distribution of the projectiles, Athena simulates atomic
collisions in crystalline targets using the Binary Collision Approximation (BCA). The
algorithm follows out the sequence of an energetic atomic projectiles (ions) launched from an
external beam into a target. The targets may have many material regions, each with its own
crystal structure, (crystalline or amorphous) with many kinds of atoms. The slowing-down of
the projectiles is followed until they either leave the target or their energy falls below some
predefined cut-off energy.
The crystal model is invoked with the MONTE parameter in the IMPLANT statement. Athena
will then choose which model to use depending on the predefined crystal structure of the
material. Specifying CRYSTALLINE has no affect on the implantation and the BCA parameter
is just a synonym for MONTE. You can manipulate the implantation module to consider all
materials amorphous by adding the AMORPHOUS parameter in the IMPLANT statement. At that
moment, the materials with predefined crystal structure are Si, Ge, GaAs, SiGe, InP as well as
three types of silicon carbides (3C-SiC, 4H-SiC, and 6H-SiC) and two types of
superconductors (Ba2YCu3O7 and Ba2NdCu3O7). All remaining materials in Athena are
considered amorphous.

Statistical Sampling
In order to reduce calculation time and improve statistical quality of simulated profiles,
Athena implements a three-dimensional rare event algorithm. An implantation profile can
differ significantly in concentration values across implantation depth. Low concentrations in
the profile are due to low probability of implanted species (rare events) to reach that point in
space. Therefore, the number of cascades simulated to get good statistics profile depends on
the desired number of orders of magnitude of accuracy. Even in real experiments, depending
on device size, implant distributions below some threshold concentration value could exhibit
significant statistical noise.
The algorithm uses trajectory splitting to achieve increased occurrence of the rare target
events by generating several independent sub-trajectories from less rare events. The original
idea [70], [71], and [72] was first developed into a refined simulation technique by Villién-
Altamirano et al., [73]. They call their version of this approach, Restart. The basic idea is to
identify subspaces from where it is more likely to reach the target subspace where the rare
event occurs. Each time these subspaces are reached, current events sequence are split in a
number of replicas all continuing from the splitting state. The number of rare events will then
increase, depending on the number of restart thresholds defined and the number of replicas
generated.
The trajectory splitting algorithm naturally fits into the problem of Monte Carlo simulation of
stopping and ranges (i.e., ion implantation). A similar method was first used by Phillips and
Price to simulate hot electron transport [74]. Yang et al. were the first ones to apply the rare
event algorithm was applied to simulation of transport phenomena of ions in matter [75].
Then, Beardmore et al. significantly refined this algorithm [76].
A brief but comprehensive review of trajectory splitting methods used in modelling of ion
implantation is given in [77].
With the rare event trajectory splitting technique, the speed-up is due to changes in the
statistical behavior so that rare events are provoked to occur more often. The rare event
algorithm in Athena achieves this by identifying subspaces from where it is more likely to
observe given collision event, followed by making replicas of the cascade sequences that
reach these subspaces.
Figure 3-26 illustrates the trajectory splitting and restart of events (replicas) as a new
threshold is reached. When applying splitting to collision cascades (or any other specific
system), the two things that need to be determined are when to split and how many sub-
trajectories to create when splitting.
There are different criteria that can be used to obtain the threshold states when splitting need
to occur. For example, Bohmayr et al. use a trajectory split method based on checking the
local dopant concentration at certain points [78]. Beardmore et al. rare event algorithm uses
the integrated dose as a criterion when to split [76].
Athena uses the same criterion to determine the splitting depths. Dose integration is carried
out along the radius vectors of ions' co-ordinates thus, roughly taking into consideration the
three-dimensionality of the ion distribution.

Figure 3-26 Restarting Collision Events by splitting at “m” thresholds

Due to the discrete nature of collision cascades, the number of sub-trajectories created at each
split depth should be an integer number greater or equal to two. Suppose Ti is the event at
each threshold state (i.e., this is the event of ion passing through a split depth). Suppose also
the probability of an ion being in state Ti-1 to reach the state Ti is pi=P(Ti|Ti-1). Then, the
recommended number of replications at each threshold (a split depth) is Ri=1/P. This relation
gives the link between the number of replications at each split and the criterion to identify the
threshold states (i.e., the split depths).
If Ri=2, then the number of ions passing through split depth will be twice smaller the number
of particles passing through split depth. In Athena, the criterion to determine the split depths
is the integrated dose along the radius vectors of stopped particles, di (i.e., split depths d1, d2,
d3, and so on will be at doses 0.5, 0.75, 0.875, and so on, where  is the total retained
implant dose).
In Athena, trajectory splitting is turned on with the SAMPLING parameter in the IMPLANT
statement. For more information this statement, see Section 6.28 “IMPLANT”.
Theoretically, sampling estimators are unbiased and consistent. In practice, however, the
estimate is obtained as the average of finite number of samples, as opposed to the theoretical
expectation, which is obtained from the whole ensemble of sample paths, including even the
very unlikely ones. Overbiasing can occur if the only goal in mind is to increase the
probability of the event that needs to be analyzed further as is the case of trajectory splitting
ion implantation simulations. The result of overbiasing is usually the underestimation of the
probability to be evaluated (dopant concentration in case of ion implantation). In fact, it has

been reported in [79] that when the splitting parameters are not consistent with the system's
large deviations behavior, the probability in question may be underestimated by several
orders of magnitude. This situation is almost present in ion implant simulators when treating
multi-layered targets and two-dimensional layouts. Therefore, use splitting with caution.
In conclusion, you should bare in mind that:
• increasing the probability of occurrence of the event to be analyzed is not always enough
to guarantee variance reduction.
• trajectory splitting should be used carefully.
• complex implantation geometries could lead to large deviations behavior of the system,
thus overbiasing and underestimating the relevant statistics.
3.5.5 Ion Implantation Damage

Ion implantation induced crystal damage can play an important role in the various
mechanisms related to diffusion and oxidation. Athena includes several different types of
damage formation, which can be used in a subsequent diffusion calculation. Implantation
induced damage results from cascades of atomic collisions. If these collisions cascades are
dense, it may result in the crystal lattice becoming locally amorphized. Accurate simulation
of collision cascades with simultaneous estimation of generating various types of point
defects, clusters, and spatial defects can be done only in elaborated Binary Collision
Approximation (BCA) or Molecular Dynamics (MD) simulators. Such simulations are
usually time consuming and impractical within general-purpose process simulators.
Generally, the amount of damage and distribution of defects associated with it depend on the
energy, species, and dose of implanted ions.
Athena includes several simple models that link various types of defect distributions with ion
implantation distributions calculated using any of the models described in previous sections.
The following types of defects can be estimated:
• Interstitial profiles
• Vacancy profiles
• {311} Clusters
• Dislocation Loops
You can describe the damage types to the simulator during the analytical ion implantation
process simulation step by scaling their distribution densities to the implanted profile.
Plus 1 Model
The first damage model is related to free point defects. Here, interstitials are scaled to the as
implanted dopant profile with the scaling parameter DAM.FACTOR=<n>. This model is
invoked with the UNIT.DAM flag on the IMPLANT line.
For example:
IMPLANT PHOS DOSE=5E14 ENERGY=45 UNIT.DAM DAM.FACTOR=0.001
This model is known as the Plus 1 model. In the case of low implantation doses, the value for
DAM.FACTOR has been suggested to be equal to unity. Although perhaps valid at low doses,
the related and subsequent diffusion mode METHOD FULL.CPL is not required in most cases.
Therefore, this combination is an impractical approach. Recent research on RTA diffusion
models (e.g., Stanford’s {311} Cluster model) has introduced other forms of damage. Thus,
lowering the dependency of free point defects being initially set at a Plus 1 scaled profile.

The DAM.FACTOR parameter, when used with the {311} Cluster model, should have a far
lower value in the order of 0.001. Note that this is an extremely sensitive parameter when
studying shallow junction formation so use it carefully.
{311}Cluster Model
The {311} Cluster model introduces a bulk injection source of interstitials in addition to any
other free point defects sources. Clusters are introduced during ion implantation, scaled to the
dopant, and within two user-defined concentration thresholds. For example, you can scale
clusters to 1.4 times the dopant concentration but only exists between the dopant
concentrations of 1e19 and 1e17 cm-3. This allows a scalable approach, where clusters will
follow implanted dopant as energies and doses vary (see Figure 3-27). The following syntax
to both switch on and control the cluster model damage scaling.
METHOD FULL.CPL CLUSTER.DAM
CLUSTER CLUST.FACT=1.4 MIN.CLUST=1e17 MAX.CLUST=1e19 PHOS
See “RTA Diffusion Modelling” on 145 on how to use the {311} clusters during RTA.
Figure 3-27 Cluster Damage Control

Dislocation Loops Model
Dislocation loops can also be scaled to the as-implanted dopant profile. Loops are introduced
as a simple static band to act as an interstitial sink. Here, interstitials will be recombined at an
enhanced rate according to
Rate = damalpha   C I – C I   3-238

 

Here, CI is interstitial concentration and C I is equilibrium interstitial concentration.
Loops are placed in a band scaled to dopant concentration with the following command
before implantation.
DISLOC.LOOP MIN.LOOP=1e16 MAX.LOOP=1e18 PHOSPHORUS

The recombination rate within the loop band is controlled as follows:

INTERSTITIAL SILICON DAMALPHA=1e8
C-Interpreter Model
The C-Interpreter capabilities shown in Appendix A “C-Interpreter” allows you to extend
control over the damage formation models described. The template for the implant damage
model function is also shown in Appendix A “C-Interpreter”. The function is introduced by
setting the DAM.MOD parameter in the IMPLANT statement. The user-defined damage model
introduced in the function will be used only within the current IMPLANT statement. All
subsequent implants will use the default damage models.
3.5.6 Stopping Powers in Amorphous Materials and Range Validation

Stopping powers in amorphous materials have been validated against available experiments.
Figure 3-28 shows a validation of boron and phosphorus ranges in amorphous silicon where
compiled experimental data are taken from [64].
Figure 3-28 Comparison of Monte Carlo simulated project ranges (lines) and measured ranges (dots) for
Boron and Phosphorus in Silicon. Experiments are from [64].

The solid lines were calculated with Athena’s Monte-Carlo Module. The spread of the
experimental points in Figure 3-28 is typical and cannot be avoided. For example, systematic
errors due to the depth calibrations and memory effects in SIMS measurements if accounted
improperly would yield less accurate (usually longer) ranges. Therefore, the Monte-Carlo
module in Athena is calibrated to give overall agreement with the available experimental
data. The figure also demonstrates that there can be a possible disagreement with individual
set of measurements. Similar stopping powers validations were performed for other important
materials. The accuracy of the calculated ranges in Athena is within 10% for majority of ion/
material combinations, which is close to the best possible achievements of today’s theory of
stopping and ranges.

Deposition Models SSuprem4 Models
3.6 Deposition Models

A deposition step is simulated by the DEPOSIT statement where the material and the thickness
(THICKNESS parameter) of the layer to be deposited must be specified. The deposited layer
is constructed by a simple algorithm that describes conformal deposition. In this algorithm,
the whole layer is divided into a number of sublayers with thicknesses equal to grid spacings
calculated according to the grid control algorithm (see Section 3.6.2 “Grid Control During
Deposit”). Each sublayer is deposited and triangulated separately.
More complete physically based models for deposition are available in the Elite module as
described in Chapter 4 “Elite Models”.
3.6.1 Deposition of Doped Layers

You can add the uniform or graded concentration of impurities or defects or both to each node
of the deposited material by using the C.BORON, F.BORON, C.INTERST parameters in the
DEPOSIT statement.
3.6.2 Grid Control During Deposit

You can control the grid inside deposited layer. The grid distribution along normal direction is
controlled by a number of divisions (the user-defined parameter, DIVISIONS) in a uniform
vertical grid. If non-uniform vertical grid is used, then specify the DY and YDY parameters in
the DEPOSIT statement.
DY specifies nominal spacing and YDY specifies the position where to apply the nominal
spacing. The spacings further from the nominal position YDY increase or decrease according
to geometrical series. The coefficients of the geometrical series are calculated so that total
number of the spacings will be equal to the DIVISIONS parameter.
When the conformal deposition algorithm fails to deposit next sublayer (which happens when
deposition occurs on the structure with narrow trenches or undercuts or both), the current
spacing divides into two and thinner layer is checked. This spacing division algorithm is
applied recursively because in these situations, the number of sublayers (divisions) actually
deposited can be higher than the number specified in DIVISIONS.
3.6.3 Epitaxy Simulation

SSuprem4 models high temperature deposition of single crystal silicon through the EPITAXY
statement. This statement combines deposit and diffusion steps and parameters. See Section
6.18 “EPITAXY” for more information.

Etching Models SSuprem4 Models
3.7 Etching Models

Although etching is an integral process step in silicon technology, SSuprem4 lacks a complete
physical description of etching steps. To circumvent this problem, SSuprem4 considers
etching simulation as a purely geometrical problem. Etching is simulated as a low-
temperature process. Impurity redistribution is ignored during the etching process.
Etch steps are simulated using the ETCH statement in which the material to be etched and the
geometrical shape of the etch region are specified. It is not necessary that material to be
etched be exposed or at the top surface of the structure. There are five different ways to define
an etch region:
1. A polygonal region may be defined by specifying the x and y coordinate of each vertex in
the polygon. Etching will confined to that polygon only.
2. A region to the left or right of a line segment may be defined by specifying the x and y
coordinates of the end points of the line segment. Etching will then proceed from the left
or right of the line segment to the edge of the structure.
3. A region between the top boundary of the structure and a line obtained by translating
exposed portion of the top boundary down in the y-direction may be defined by
specifying the DRY parameter in the ETCH statement. The THICKNESS parameter will
determine the distance to etch in the y-direction.
4. An extension of the DRY etch produces the etch region with slopped sidewalls and
undercuts under the mask. The UNDERCUT parameter specifies the top boundary’s
extension of the etch region under the mask. The ANGLE parameter defines the slopes of
sidewalls of the region. The bottom line of the etch region is defined by vertical
translation of the top boundary with undercut taken into account.
The ANGLE less than 90° results in trenches narrowing to the bottom. The ANGLE
greater than 90° produces retrograde sidewalls. The UNDERCUT length is measured
along the boundary line between etched material and masking layer.
In a special case when the etched material layer is sandwiched between two other
layers, the THICKNESS parameter is ignored and UNDERCUT is applied to both the
upper and lower boundaries of the etched material layer.
5. All regions of a particular material may be etched by specifying the ALL parameter of the
ETCH statement.
When a region is defined in one of the first three ways. By default, all materials in the defined
region will be etched. Specifying a material in the ETCH statement limits etching to only that
material within the defined region. For a complete description of physically based etch
models, see Chapter 4 “Elite Models”. Elite is a complete 2D topography simulator included
in the Athena framework.

Compound Semiconductor Simulation SSuprem4 Models
3.8 Compound Semiconductor Simulation

ATHENA allows you to simulate basic technological processes in compound semiconductors.
The set of standard compound materials includes: GaAs, AlGaAs, InGaAs, and InP.
Additional user-defined materials can be also specified. The following chemical elements are
typical dopants in these compound semiconductors: Si, C, Se, Be, Mg, Ge, and Zn.
3.8.1 Diffusion Models

The default diffusion model in compound semiconductors is the same as the Fermi Model
with electric field effect used for silicon in SSUPREM4 (see Section 3.1.2 “The Fermi Model”).
All diffusivity parameters from Table 3-2 can be specified for each dopant in all compound
materials. But, only reasonably calibrated set of diffusion parameters exist for GaAs, [80].
You should perform calibration or all other materials.
Note: More advanced diffusion models TWO.DIM and FULL.CPL can be potentially specified for compound
semiconductors. The point defects kinetics, however, is largely unknown which means that extensive
research and calibration is needed
It was determined in [80], that n-type dopants, Si, Se, and Ge in GaAs, diffuse through the Ga
Vacancy Mechanism, while p-type dopants, Be, Mg, Zn, and C, diffuse through the Ga
interstitial mechanism. This means that for donors in GaAs the diffusivity is calculated as
follows:
n 2
+ D AV  ----
x – n =
D donor = D AV + D AV ---- 3-239
ni n 
i
To look up these diffusivity terms, see Table 3-2.

The intrinsic carrier concentration ni is calculated by Equation 3-9 with the parameters NI.0,
NI.E, and NI.POW taken from [81]. The experimental data cited in [80] show that diffusivity
of Si and Se in GaAs can be considered as concentration independent and therefore only the
first term of this equation is non-zero for these impurities. It appears that diffusion of Ge in
GaAs is proportional to the second power of (n/ni).
The diffusivity for acceptors is the following:
2
x – p
D acceptor = D AI + D AI ----
=
+ D AI  ----
p
3-240
ni n 
i
Different terms are dominant for different acceptors. Carbon diffusivity is considered as
concentration independent. Be and Mg diffusivities are proportional to p/ni, while diffusivity
of Zn is proportional to (p/ni)2.
It’s important to know that some dopants in compound semiconductors are amphoteric and
can be either donor or acceptors under certain conditions. This means, you can use the DONOR
and ACCEPTOR parameters in the IMPURITY statement to specify the type of dopant.
Boundary and interface condition for impurities in compound semiconductors are specified
using the transport velocity parameters TRN.0 and TRN.E and the segregation coefficients
SEG.0 and SEG.E.

Compound Semiconductor Simulation SSuprem4 Models
The impurity activation in compound semiconductors is calculated using solid solubility

model with default value for solid solubility limit for all impurities set at 1019 cm-3.
3.8.2 Implantation Models

Ion implantation models for compound semiconductors are essentially the same as those for
silicon. The Pearson analytical approximation uses look-up tables derived from experiments
[82] and calculations [62]. These tables correspond to amorphous approximation and ignore
effects of material crystalline structure.
When accuracy of simulated as-implant profiles is important and ion implant channeling is
pronounced, use the Monte Carlo BCA implant model. The Monte Carlo model takes into
account both crystalline structure and composition of compound materials. The actual
composition and density of the default ternary compounds, InGaAs and AlGaAs, as well as
the user-defined materials must be specified in the MATERIAL statement.

SiGe/SiGeC Simulation SSuprem4 Models
3.9 SiGe/SiGeC Simulation

Several experiments (e.g., [83], [84], and [85]) revealed that boron diffusion in Si1-xGex or
Si1-xGexCy alloys may differ from diffusion in pure Si substrates. To simulate effects of Ge
concentration fraction, x, and carbon concentration fraction, y, on boron diffusion several
models were introduced in SSuprem4. To activate these models, set the MODEL.SIGEC
parameter in the METHOD statement.
3.9.1 Deposition of SiGe/SiGeC Epitaxial Layer

In this model, Si1-xGex or Si1-xGexCy alloys are considered as heavily doped with Ge or Ge
and C. Usually, layers with either constant or graded germanium content are formed by a
special epitaxy process.
You can simulate the formation of the Si1-xGexCy layer with a constant Ge and C content
using the DEPOSIT statement with a C.GERMANIUM parameter set equal to y*NSi, where NSi
is the atomic density of undoped silicon equal to 5.0 1022 cm-3. For layers with graded
germanium content, use an additional parameter, F.GERMANIUM, in the DEPOSIT statement.
The following statement:
DEPOSIT SILICON THICK=0.1 DIN=10 C.GERMANIUM=1e20 F.GERMANIUM=5e21
C.CARBON=1e19
will create the 0.1m layer of Si1-xGexCy with constant carbon concentration of 1.0 1019 cm-
3
and with germanium concentration varying from 1.0 1020 cm-3 at the bottom of the layer to
5.0 1022 cm-3 at the top of the layer. This corresponds to Ge content x varied from 0.2% to
10%.
3.9.2 Boron Diffusion in SiGe/SiGeC

The special model takes empirically into account experimental facts that boron diffusivity
apparently decreases with germanium and carbon content. The total boron diffusivity (see
Equation 3-10) decreases exponentially with the Ge content, x, and carbon content, y.
 Þ EAFACT  SiGe + y  EAFACT  SIGE
D B  Si 1 – x – y Ge x C y  = D B  Si  exp  – x---------------------------------------------------------------------------------------------------------- 3-241
 N Si kT 
Another effect taken into account in this model is variation of intrinsic carrier concentration ni
with x and y. It is presumed that ni increased with x and decreased with y.
 x  NIFACT  SIGE   NIFACT  SIC
n i  Si1 – x – y Ge x C y  = n i  Si   1 – ------------------------------------------------   1 + y------------------------------------------- 3-242
 N Si   N Si 
The user-defined calibration parameters (EAFACT and NIFACT) for the above equations are
specified in the MATERIAL statement for silicon.
As an alternative to Equations 3-241 and 3-242, you can use different dependencies for
diffusivity and intrinsic carrier concentration through C-Interpreter function specified by
SIGECDF.MOD and SIGECNI.MOD parameters in the METHOD statement.

SiGe/SiGeC Simulation SSuprem4 Models
3.9.3 Boron Transient Diffusion Suppression by Carbon Incorporation Models

There are experimental indications [85] that interstitials diffuse slowly and tend to disappear
or get trapped more intensively in SiGe layers with substitutional carbon. These effects result
in suppressing of the boron transient diffusion when carbon is incorporated into SiGe layer.
The following equations shows the diffusivity of interstitials, DI, is controlled by the
DCARBON.E parameter.
 y  DCARBON  E 
D I  Si 1 – x – y Ge x C y  = D I  Si   exp  – --------------------------------------------- 3-243
 N Si kT 
This model also introduces an additional sink for interstitials in the layers with high carbon
concentration. Intensity of the sink is proportional to the carbon concentration and is
controlled by the recombination parameters, KCARBON.0 and KCARBON.E, specified in the
“INTERSTITIAL SILICON” statement.
This means that the following recombination term will appear on the right-hand side of the
interstitial transport equation Equation 3-24.
R CARBON = KCARBON  0 exp  – ---------------------------------  -------

KCARBONE y -  C – C  3-244
kT NS i I I

Stress Models SSuprem4 Models
3.10 Stress Models

Athena provides several capabilities to calculate stresses generated during various process
steps. Simulation of the stresses generated during viscous oxidation is discussed in Section
3.3.3 “Viscous Model”. Three other sources of stresses could be simulated in Athena. These
stresses are generated by:
• the intrinsic stress in deposited films
• the stresses induced by local lattice distortion in presence of impurity atoms
• the stresses due to thermal expansion mismatch between material layers
To calculate the stresses generated in the current structure, use the STRESS statement. You can
do this at any moment of the processing sequence usually after deposition or etch or both of
the layer with specified intrinsic stresses or after formation of a highly doped layer or area by
implantation or epitaxy process. Alternatively, Athena could follow the stress history if you
specify the STRESS.HIST parameter in the METHOD statement. To calculate resulting stresses,
Athena solves the following system of equations using a finite element analysis.
 xx  xy
----------- + ----------- = 0 3-245
x y
 yy  xy
----------- + ----------- = 0 3-246
y x
 T2 
 u x u y 

E int int int
 xx +  y y = -------------------------------------  --------- + -------- – 2  1 +   LCTE  dt  +  xx +  yy + 2 3-247
 1 +    1 – 2    x y 
 T1 
u u x
 xx –  y y = ------------------  -------y- – --------- +  xx –  yy
E int int
3-248
 1 +   y x
u u
 xy = ---------------------  --------x- + -------y-
E
3-249
2  1 +    y x 
This system of equations corresponds to 2D formulation of the linear elastostatics problem

for an isotropic homogeneous media. This means that all forces applied to the elastic structure
sum to zero and displacements are not a function of time. Equations 3-245 and 3-246 are
equations of motion in equilibrium conditions and Equations 3-247 through 3-249 are
constitutive equations for elasticity formulated through displacements. Here, xx, yy, and xy
are calculated components of the stress tensor, ux and uy are displacements in x and y
directions.
Note: Displacements ux and uy are saved in the structure file as velocity X and velocity Y respectively and can be
used in EXTRACT commands or visualized in TonyPlot.

The elastic properties of each material in the simulation structure are specified through two
elastic moduli: Young modulus E and Poisson’s ratio . In the case of cyrstalline materials,
the values of E and  depend on the crystallographic plane defining material surface by the
parameters 100, 110, or 111 in the INITIALIZE statement) and on specific orientation within
the plane or angle measured from the wafer flat direction (which is specified by ROT.SUB in
the INITIALIZE statement). For the cubic lattice, these dependencies are as follows [140]:
for surface plane {100}
1
E = -------------------------------------------------------------------
s 11 – 2S  1 – cos2   cos2 
3-250
s 12 + 2S cos2   1 – cos2  
 = – -------------------------------------------------------------------
-
s 11 – 2S cos2   1 – cos2  

1
E = -----------------------------------------------------------------------
1
s 11 – --- S cos2   4 – 3 cos2  
2
3 3-251
s 12 + --- S cos2   1 – cos2  
2
 = – -----------------------------------------------------------------------
1
s 11 – --- S cos2   4 – 3 cos2  
2

4
E = -------------------------------------------
2s 11 + 2s 12 + s 44
1 3-252
s 11 + s 12 – --- s
2 44
 = – ------------------------------------------------
-

3 s 11 + s 12 + --- s
1 
 2 44
where s11, s12, and s44 are elastic constants (compliances) specified by the parameters
1
S11.ELAST, S12.ELAST, and S44.ELAST in the MATERIAL statement. S = s 11 + s 12 – --- s 44
2
and is the substrate rotation angle (ROT.SUB).
In case of amorphous material, the relations between E and  and constants are simplified as
follows: S=0, E=1/s11, and =-s12/s11. Therefore, the following formulae can be used for
obtaining elastic constants from Young modulus E and Poisson’s ratio :
s 11 = 1  E 3-253
s 12 = –  E 3-254
s 44 = 2   1 +    E 3-255

The second term of Equation 3-247 describes extrinsic stress generated by thermal expansion
mismatch between material layers, for example, substrate and overlying films. The linear
expansion coefficient, LCTE, is specified as a function of absolute temperature T in the
MATERIAL statement. This function is integrated between temperature T1 and T2, which are
either specified as TEMP1 and TEMP2 in the STRESS statement or correspond to temperatures
at the start and end of process steps if STRESS.HIST is used (see Section 3.10.2 “Stress
History”).
The term int in Equation 3-247 corresponds to the initial intrinsic stress in a deposited
material layer. This intrinsic stress in grown or deposited films depends on many factors
including thickness, deposition method, rate, temperature, and ambient.
Athena cannot simulate all these effects. You can, however, specify the initial intrinsic stress
value INTRIN.SIG in the MATERIAL statement.
3.10.1 Dopant-Induced Intrinsic Stresses

int int
The last terms  xx and  yy in Equations 3-247 and 3-248 represent local intrinsic stresses
induced by dopants. The size of impurity atoms can differ significantly from that of the
substrate silicon. Smaller dopants like B and P result in local lattice contraction, while bigger
atoms. For example, Ge in a substitutional site causes lattice expansion. The lattice expansion
or contraction increases with doping concentration. The local strain can be expressed as  =
a/aSi, where a is local change of lattice constant and aSi is the silicon lattice constant.
According to the Vegard’s law, the lattice constant is approximately proportional to the
concentration of the constituent elements in “alloyed” materials. To take into account a
possible deviations from the linear dependency, Athena allows quadratic dependence of
intrinsic stress on local impurity concentration as follows:
int 2
 xx = E xx = E  A xx + B xx x + C xx x  3-256
int 2
 yy = E yy = E  A yy + B yy x + C yy x  3-257
where x = CA/N, CA is local impurity concentration and N is atomic density of the material
(for example, 5 *1022 cm-3 for silicon). You can specify the parameters Axx, Bxx, Cxx, Ayy, Byy,
and Cyy as follows in the IMPURITY statement.
The dopant-induced intrinsic stresses are included only when you specify the
DOPANT.STRESS in the METHOD statement.
Note: If there are several local impurities in the structure, the intrinsic stresses are calculated by summation of
above equations for each impurity.
High concentration of impurity may also affect elastic properties of the bulk material. For
example, it is known that the Young modulus of SiGe is lower than that of pure Silicon and
that it can be approximated as follows:
E = E0 + Ay x 3-258
where E0 is the Young modulus of Silicon, Ay is linear coefficient specified by parameter
A.YOUNG in the IMPURITY statement, and x is Ge fraction in SiGe. Athena allows the same
linear dependence for all impurities.

3.10.2 Stress History

If the STRESS.HIST method is specified, Athena then calculates stresses when the simulation
structure changes after etching, deposition, epitaxy, and diffusion processes. The temperature
(including ramp) specified for current process step is used in the calculation. The room
temperature is used if the temperature isn’t specified in the ETCH or DEPOSIT statement. The
final stresses from the previous step is used as a initial condition for the subsequent step. If
the temperature is changed between the end of one step and the start of another, the stress
calculation with corresponding temperature ramp is automatically inserted.
When the STRESS.HIST specification is active, all oxidation steps will be simulated with
VISCOUS model and stresses inside silicon will be calculated in the end of each oxidation
step.
The STRESS.HIST will be switched off when stresses are calculated explicitly using the
STRESS statement or when material reflow is simulated by using the REFLOW parameter in the
DIFFUSE statement.
3.10.3 Stress-Induced Diffusion Enhancement/Retardation

Several studies have clearly demonstrated that mechanical stresses in silicon may affect
impurity diffusion. Most of those studies were related to diffusion in SiGe and strained
Silicon near the SiGe layers (for example, see [137]). Experiments and accompanying models
show diffusion retardation in case of compressive stresses. It was also shown that diffusivity
dependence on local strain could reasonably be approximated by an Arrhenius formula. Sheu
et.al. [138] suggested to use this approximation for any stresses generated during Si
processing.
To use the stress-induced diffusion model in Athena, specify the STRAIN.DIFF parameter in
the METHOD statement. The local impurity diffusion coefficient DS in presence of stresses is
calculated as follows:
E s   xx  x, y  +  yy  x, y  
D s  x, y  = D I exp – -------------------------------------------------------------
- 3-259
kT
Here:
• DI is the local diffusivity calculated without stresses.
• ES is the activation per volume change ratio (or strain), which is specified by the
E.STRAIN parameter in the IMPURITY statement.
• xx=xx/E and yy =yy/E are the components of strain.
Sheu et.al. [138] have found that diffusion appears to be retarded in the presence of
compressive stresses (or when xx + yy is negative). Therefore, the values for ES are
negative for B, P, and As impurities.

Chapter 4
Elite Models
Overview Elite Models
4.1 Overview
The Elite module of Athena allows the use of sophisticated models for deposition and etch
processes. These processes are modeled by defining a machine and invoking the machine to
perform either deposit or etch. Elite also includes a model for material reflow. Elite can also
be licensed with modules for Monte Carlo deposition, Monte Carlo etching, and Chemical
Mechanical Polishing (CMP).
In Elite, a number of default machines are defined so that specifying any process reasonably
close to the standard is especially simple. Process modifications or additions are easily
implemented by changing or adding individual machines without affecting the remainder of
the simulator.
For all models except Monte Carlo deposition and Monte Carlo Etching, Elite uses a string
algorithm to describe topographical changes that occur during deposition and etching
processes. This chapter describes the models and techniques used in Elite and the command
language used to access model parameters.

String Algorithm Elite Models
4.2 String Algorithm

The Elite simulation regime consists of a set of triangles that hold information on the
materials that are being simulated. The string algorithm treats each of these interfaces as a set
of segments that move in response to a particular process calculation.
As microfabrication technology becomes more complex, modeling each step of the
manufacturing process is increasingly important for predicting the performance of the
technology. Etching is a step that is universal in microfabrication. It may take place as the
dissolution of a photoresist by an organic solvent, the etching of an oxide by an alkali, or the
plasma etching of an electron resist. Whatever its physical details, the etching process can in
many cases be modeled as a surface etching phenomenon. Etching simulation starts from an
initial profile that moves through a medium in which the speed of etching propagation can be
a function of position and other variables that determine the final profile.
Two major assumptions limit the generality of the string algorithm in Elite. First, the pattern
to be etched is uniform in one dimension, so the problem can be solved using only two
dimensions. For most microfabrication problems, the important cases involve the cross-
sections of lines, so this model is directly applicable. In certain other cases, such as round
holes, the symmetry of some cross-sections is such that the algorithm is still valid. The
second major assumption is that the etch rate is a scalar function of position, and is
independent of the direction of local etch front motion and the history of the front. In some
real situations, this does not hold. PMMA, for instance, has been found to have a gel region at
the resist-solvent interface during development, so the etch rate is a function of the history of
the adjacent regions as well as of the exposure. Another case where the second assumption
does not hold is in the so-called “preferential etching” where etching proceeds more quickly
along certain crystal directions, making the etch anisotropic.
The algorithm described here is known as a “string algorithm” [86]. The etch front is
simulated by a series of points joined by straight line segments, forming a string. During each
time increment, each point advances perpendicularly to the local etch front as in Figure 4-1. A
major portion of the algorithm adjusts the number of segments to keep them approximately
equal in length. Other subroutines input the data and output the etch front.
Choosing suitable criteria for segment length was a major problem in developing the
algorithm. It seemed that segments must be short enough so that any curve that developed
would be well defined, i.e., there should be some maximum angle between adjacent
segments, perhaps 0.1 radians. This criterion, however, led to a proliferation of segments in
regions where the front was either expanding or contracting. The algorithms in Elite attempt
to maintain approximately equal segment lengths. This results in position errors of about one-
half segment length. The error can be reduced by decreasing the average segment length with
a proportional increase in computation time.

String Algorithm Elite Models
Figure 4-1 String Model approximation to the Etch Front

For the most cases of interest, the etch rate varies with position. This leads to some errors in
the position and in the direction of each point on the string. Errors in position arise from the
use of a rather simple integration algorithm. The local rate at the start of each time step is
assumed to be constant throughout the step. This can easily lead to position errors as large as
the distance covered in one step. Consider, for example, an etch front in a photoresist
approaching an unetchable substrate. A point, which is barely outside the substrate at the start
of the time step, will advance into the substrate at the rate associated with the resist. Thin
layers of alternating fast and slow etch rates could spawn errors in position. With too large a
time step, a point could jump over a slow region.
Errors in direction arise from non-uniform rates along the string and from certain boundary
conditions. During each step perpendicularity to the front, which is defined below, is assumed
to be constant in direction. If two adjacent points have greatly differing rates, however, the
quickly moving point cannot start turning towards the slower point until the end of the time
step. This mechanism tends to introduce relatively small errors in position because the error is
roughly proportional to the cosine of the angle error.

Deposition Models Elite Models
4.3 Deposition Models

Elite provides a set of deposition models that correspond to different physical deposition
techniques. Most of the models were first developed at UC Berkeley [87], [88], [89], and [90]
and were originally implemented in the topography simulator SAMPLE [91]. Any one of
these models can be selected to define a “machine” for simulating processes on the structure.
In addition, Elite provides a conformal deposition capability that can be used to define initial
structures.
In most integrated-circuit processes, at least one layer of interconnect is formed by depositing
and patterning an Al or Al alloy film. The trend toward lower temperature processing,
combined with the very steep edge profiles produced by anisotropic dry etching processes,
results in sharp step profiles, which are difficult to cover with a uniform film of metal.
4.3.1 Conformal Deposition

You can perform conformal deposition by specifying a material to deposit, a thickness, and a
number of vertical grid spacings on the DEPOSIT statement. The conformal deposition model
produces unity step coverage.
4.3.2 CVD Deposition

To use this model, specify the CVD parameter in the RATE.DEPO statement as well as the
material type, the deposition rate DEP.RATE, and step coverage, STEP.COV.
The local deposition rate R(x,y) for the CVD model is given by
R  x y  = DEP.RATE   1 – STEP.COV  cos  + STEP.COV  4-1
where  is the angle between the surface segment and the horizontal.
4.3.3 Unidirectional Deposition

To specify this model, specify the UNIDIRECT parameter in the RATE.DEPO statement.
As shown in Figure 4-2, the region of the substrate not shadowed sees the arrival of the vapor
streams in one direction only. The growth rate of the deposited film in the shadowed region is
equal to zero. According to these assumptions, growth rate on the substrate R(x,y) can be
expressed as:
R (x,y) = 0 , if point (x,y) is shadowed 4-2
R (x,y) = C sin  i + C cos  j 4-3
where:
•  is the angle between the y-axis and the direction of the vapor stream.
• i and j are the unit vectors in the x and y direction respectively.
• C is the growth rate of an unshadowed surface normal to the vapor stream.
• Angle  is specified as ANGLE1 on the RATE.DEPO command.

The following is a short description of the cosine law deposition. This is a simple model that
accounts for metallization, due to evaporation. The cosine law deposition model is based on
the following assumptions:
• The mean free path of atoms or particles is much larger than the distance between the
source and the substrate.
• The source to substrate distance is large compared to the surface topography.
• The film grows in the direction toward the vapor flux.
• Shadowing effects must be included.
The magnitude of the film growth rate follows the cosine distribution law, which says that
deposited film thickness grows at a rate proportional to cos(), where  is the angle between
the vapor steam and the normal surface.
The sticking coefficient can be used as a tuning parameter. It is assumed to be 1.0 for
deposition on the cold substrates (at 300 K).
Figure 4-2 Step Profile with a Unidirectional Source

4.3.4 Dual Directional Deposition

This model is invoked by specifying the DUALDIRECT parameter in the RATE.DEPO statement.
In this type of source, each point in the unshadowed region views the vapor streams arriving
from two different directions, and assumes the diffusion length of deposited material large
compared to the features (see Figure 4-3). Growth rate is given as:
R  x, y  = 0, if point  x, y  is shadowed 4-4
R  x y  = C sin  1 i + C sin  1 j or R  x y  = C sin  2 i + C sin  2 j 4-5
if point (x,y) is partially shadowed.

R  x y  = C  cos  1 + cos  2 i + C  sin  1 + sin  2 j 4-6
if point (x,y) is unshadowed, where 1 and 2 are the incident angles. 1 and 2 are specified
on the RATE.DEPO command by ANGLE1 and ANGLE2, respectively.
ω1 ω2
Figure 4-3 Step Profile with Dual Source

4.3.5 Hemispheric Deposition

To use this model, specify the HEMISPHERIC parameter in the RATE.DEPO statement
The flux of vapor is continuously distributed in a range of directions (see Figure 4-4). The
growth rate can be calculated as:
R  x y  = C  cos  1 – cos  2 i + C  sin  1 – sin  2 j 4-7
where 1 and 2 are the lower and upper bounds, respectively, of the incident angles of the
vapor streams set by parameters ANGLE1 and ANGLE2, respectively.
To avoid step-coverage problems, planar sputtering is often used to achieve better film
profiles. The ideal sputtering source is modeled by means of a hemispheric vapor source with
atoms impinging on the substrate from all angles.
ω1 ω2
Figure 4-4 Step Profile with a Hemispherical Vapor Source

4.3.6 Planetary Deposition

To use this model, specify the PLANETARY parameter in the RATE.DEPO statement.
Figure 4-5 illustrates the planetary evaporation system. By inspecting this system, you can be
convinced that the rotation of the planet along the system central axis has no effect on the
deposition rate. For simplicity’s sake, you can calculate the growth rate by holding the planet
stationary and by rotating only the source along the axis of the planet (see Figure 4-5). The
growth rate is derived according to the following equations:
2 2 2
 R – r – rL tan  + LW   – L sec      L tan  sin  – L cos   tan 
R x (x,y) =
 ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ d
2 2 2
 R – r + L – 2rL tan    R + W   R –  r + L tan   
2 2 2 2 2
4-8
2 2 2
 R – r – rL tan  + LW   – L sec      L tan  sin  – L cos  
R x (x,y) =  ----------------------------------------------------------------------------------------------------------------------------------------------------------- d
2 2 2
 R – r + L – 2rL tan    R + W   R –  r + L tan   
2 2 2 2 2
4-9
where:
•  is the incident angle of the vapor stream.
•  is the tilt angle of the planet plane.
• .
• r is the distance between the position of the wafer and the planet axis.
• R, L, and W are the parameters dependent on the system dimensions.
Figure 4-5 Illustration of Planetary Evaporator

Using the planetary model of Elite, you can observe asymmetries both in edge coverage and
the depth of cracks produced by the particular location and orientation of a specimen in a
planetary system.
Figure 4-6 shows the following planetary model parameters:
= ANGLE2, = ANGLE3,
r = DIST.PL, P.AXIS, C.AXIS, and ANGLE1
ANGLE1 is used to calculate DIST.PL = P.AXIS * tan(ANGLE1).
Figure 4-6 Geometric of Source to Substrate in a Conical Evaporator

4.3.7 Conical Deposition

To use this model, specify the CONICAL parameter in the RATE.DEPO statement.
The Conical model is a simplified version of the Planetary model with  and r = 0, the
substrate always sees a symmetrical cone source. In this type of configuration, the integral of
the above two equations can be evaluated analytically and expressed in the following simple
closed form:
2
– R  R + LW   2 2
R x  x y  = ---------------------------------------------------------  1 –  --- tan  max – 1 –  --- tan  min 
L L
2 2 R  R 
R + W  R + L   
2 2 2
4-10
If expression under either square root in this formula is less than 0 it is set to exact 0.
 
2
L  R + LW    L2 2 
R z  x y  = ---------------------------------------------------------  asin  --- tan  max –
L
asin  ------ tan  min   4-11
2  R   R2 
R + W  R + L  
2 2 2 2
 
In the Conical model, the parameter, ANGLE1=, and other parameters are C.AXIS and
P.AXIS as shown in Figure 4-6.

4.3.8 Monte Carlo Deposition

There are two models that are invoked by specifying the MONTE1 or MONTE2 parameters in the
RATE.DEPO statement. The parameters SIGMA.DEP, DEP.RATE, and ANGLE1. MONTE1 invokes
the Monte Carlo based deposition model, which you can use to model low-pressure chemical
vapor deposition (LPCVD) [92], [93].
Since the radicals are incident on the substrate with non zero thermal velocities, they may be
re-emitted from the surface before they react. Therefore, the probability of their sticking is
considered. You can define the sticking coefficient using STICK parameter in the RATE.DEPO
statement.
Figure 4-7 Deposition and Relaxation Model used in Ballistic Deposition model (MONTE2)

The model uses an analytical approach to calculate a surface diffusion through a normalized
gaussian distribution nd:
 x
2 
nd = exp  – ------------------------------------ 4-12
 SIGMA.DEP2
where x is the point of contact with the surface as shown in Figure 4-7.
MONTE2 invokes a ballistic deposition model, which simulates film growth by the random
irreversible deposition of hard two-dimensional discs launched with linear trajectories from a
random point at the top of the simulation area towards the structure surface.
At the point of contact with the growing film, the incident discs are relaxed to the nearest
cradle point with the highest coordination number (contacting the largest number of neighbor
discs) within a radius equal to SIGMA.DEP, which is four disc diameters by default.
The profile was initialized using a series of discs. In order to inhibit unrealistic “epitaxial”
growth from a closest-packed surface [94] and [95], the initial series of discs was spaced with
centers approximately 1.3 diameters apart.
This relaxation process simulates limited surface diffusion that occurs in films to reduce the
surface energy associated with areas of high curvature.
A prediction of the trends in local film density can be achieved. Plot with discs can be
obtained using the parameter OUTFILE=<filename> in the DEPOSIT statement. Figure 4-7
shows the vapor flux distribution arriving can be defined using the ANGLE1 parameter
describing the angle measured between the vertical from the source and the wafer normal.
To use multiple steps for both MONTE1 and MONTE2 models, set the DIVISION parameter in
the DEPOSIT statement. The number of incoming particles can be defined by the N.PARTICLE
parameter in the DEPOSIT statement.
4.3.9 Custom Deposition Models

Elite implements two slightly different custom deposition models. In both models, the angle
is d.

Etch Models Elite Models
4.4 Etch Models

Elite provides a set of etch models that correspond to different physical etching techniques [96],
[97], and [98]. Any one of these models can be selected to define a machine that can then be
invoked to perform processing on the structure. In addition, Elite provides a primitive etching
capability that can be used to define initial structures.
4.4.1 Isotropic Etch Model

To use the model, specify the WET.ETCH parameter in the RATE.ETCH statement. In wet etching
and simple plasma etching, the substrate is immersed in a fluid (liquid or gas), which chemically
reacts with the exposed surface. In wet etching, the liquid attacks the surface and removes soluble
products. This reaction produces volatile by-products, which are removed by a vacuum pump.
Physical reactions do not take place. A barrel plasma reactor achieves such conditions, usually at
low powers and moderate pressures. Due to the chemical reaction, isotropic profiles develop with
mask undercutting and circular cross-sections.
riso bisects the angle

between line segments.
r = riso
Figure 4-8 Segment Point in Case of Isotropic
4.4.2 RIE Model

In the Reactive Ion Etching (RIE) model, the etching process is divided into the two adjustable
components isotropic etching and anisotropic etching. Each of these components is characterized
by empirical etch rates (riso and rdir).
The ratio:
r dir
A = -------------------------
- 4-13
r riso + r dir
defines the measure of anisotropy.

The isotropic component riso models chemically reactive etching which results in profiles with
undercut and circular cross-sections. For A=0, the process is completely isotropic. Under
isotropic conditions, the string-points are advanced at the constant rate riso, in the direction of the
perpendicular bisector of the adjacent segments (see Figure 4-8).

The anisotropic etch rate component rdir is proportional to the cosine of the angle between the
flux direction and the surface normal (the perpendicular bisector of adjacent segments). For
A=1, the process is anisotropic yielding vertical sidewalls (see Figure 4-9).
Figure 4-10 illustrates the regions of significance for each component in the RIE model. The
shadowing effect is accounted for by the riso component in the shadowed area.
α
rdir
r = rdir + cos α
r = rdir
Figure 4-9 Point Advance due to Directional Influence
α α
Figure 4-10 Regions of Significance of rdir and risc

4.4.3 Dopant Enhanced Etching

Dopant enhanced etching is a feature included in Elite and allows the etch rate at any point on
the surface to be changed depending on the value of any solution variable present. The etch
rate at any point is then given by the formula
ER enh =  1 + enh ERM 4-14
where ERenh is the enhancement or retardation due the presence of particular dopant. All
impurities as well as interstitials, vacancies, and stress solutions Sxx, Sxy, and Syy can be
specified in the model.
This enhancement is calculated using the formula
enh = 0.5 ENH · MAX  tanh  ENH · SCALE  S – ENH · MINC  + 1   4-15
where ENH.MAX is the maximum value of enhancement or retardation, ENH.MINC gives the
solution value below which enhancement decays and ENH.SCALE gives the spread of the
enhancement over solution values. In other words, how quickly does the enhancement or
retardation factor reach its maximum. S is the dopant value. The positive value of ENH.MAX
corresponds to enhancement, while negative value corresponds to retardation.
For exponentially varying solutions, e.g., oxidation stress and dopant concentrations, both S
and ENH.MINC are taken to be log base 10 of their respective value. Parameters of the model
are specified in the RATE.DOPE statement.
4.4.4 Plasma Etch Model

The plasma etch model in Athena is based on a Monte Carlo simulation of the ion transport
from the neutral plasma or bulk, denoted by its glow, through the dark sheath surrounding the
electrodes and walls and isolating the plasma. Ions enter the sheath from the plasma and are
then accelerated through the sheath due the electrical potential drop between the plasma and
the electrodes. The Monte Carlo simulation follows a large number ions in their transport
through the sheath including collisions with other gaseous species present in the etch
chamber. The number of collisions encountered by a particular ion depends on both the ion
mean free path, a calculated quantity, and the sheath thickness, an user-specified quantity. To
reduce the computation time, ion trajectories are calculated independently and inter-ion
interactions are not considered in this version of the code.
In the current version of Athena, the simulated Monte Carlo distributions are used to calculate
an ion flux incident on the substrate surface. This flux is then used to calculate an etch rate by
integrating this flux over the window of visibility at each point on the surface. The window of
visibility is, for point on a flat surface, simply from 0 to 2  . But for more complicated
structures (e.g., trenches), points on the surface are shadowed and the “window of visibility”
is reduced. Currently, only a simple linear surface kinetic model for etching is supported.
See Chapter 3 “SSuprem4 Models” for a description of the RATE.ETCH parameters required
for plasma etch simulation.

4.4.5 Monte Carlo Etching Model

The shrinking critical dimensions of modern technology place a heavy requirement on
optimizing the etching of narrow mask opening. In addition, the aspect ratio of etches has
been increased, requiring deeper etches along with the small kc/s. The simulation of these
process requires more advanced techniques than the analytical rate-based etching models
described above. A more complete treatment involving calculation of the plasma distribution
and direct interaction of plasma particles with substrate materials is required.
The Monte Carlo etch module is implemented into Athena/Elite. The main application of the
module is simulation of plasma or ion assisted etching. The module can take into account the
redeposition of the polymer material generated as a mixture of incoming ions with etched
(sputtered) molecules of substrate material. Also, the module has interface to the C-
Interpreter, which allows simulation of several other processes such as wet etch and
deposition, ion milling and sputtering deposition of various materials. The Monte Carlo etch
module was successfully used by Toshiba researchers for simulations of reactive ion etching
of narrow deep trenches in oxide [99].
Simulation of Incoming Ions and Neutrals
Direct modeling of the plasma sheath is not included into this release and will be added later.
It is assumed that ions and neutrals fluxes leaving plasma sheath are represented by
bimaxwell velocity distribution function along the direction determined by user specified
incident angle:
   
f(     )  I  exp  – ------------ – ------ 4-16
 T  T 

where:
  is the ion velocity component parallel to the incident direction.
  is the ion velocity component perpendicular to the incident direction.
I ion (or neutral) current density specified by parameters, MC.ION.CU1 or MC.ION.CU2 in the
RATE.ETCH statement.
T  is the dimensionless parallel temperature specified by parameters, MC.NORM.T1 or
MC.NORM.T2.
T  is the dimensionless lateral temperature specified by parameters, MC.LAT.T1 or
MC.LAT.T2.
The incident angles are specified by the MC.ANGLE1 and MC.ANGLE2 parameters.

Calculation of Ion and Neutral Fluxes

During each time step, the simulation consists of the three stages. These stages are as follows:
1. Calculation of ion, neutral, and polymer fluxes
2. Calculation of etch, polymer ejection and redeposition rates
3. Surface movement
On the first stage, the fluxes of incoming and reflected ions and neutrals are calculated on the
each segment of the surface. Computation of the ion fluxes is done by tracing the user-defined
number of particles (Figure 4-11, model a). Each particle is generated at random positions on
top of the simulation area, with normal and lateral velocities randomly determined from the
bimaxwell distribution function (Equation 4-1). Then, each particle trajectory is traced until
the ion is either absorbed by the surface or back scattered out of the simulation area.
Figure 4-11 Diagram of Plasma Flux algorithm: (a) including zoom-in of ion reflection models (a & b)

The interaction of the ion with material surface is governed by two factors. The first is the
reflection coefficient Prefl, which is specified by the MC.ALB1 and MC.ALB2 parameters for
two types of plasma particles. The second is MC.PLM.ALB for polymer particles and
roughness of the surface R, which is specified by the MC.RFLCTDIF parameter. Both factors
depend on the surface material and the type of ion. Reflection coefficient is the probability of
the particle to be reflected from the surface. Roughness determines how the ion is reflected. If
R = 0 the reflection is specular (Figure 4-11, model b). If R = 1, the reflection is random with
uniform angular distribution (Figure 4-11, model c). In general, the velocity v refl of the ion
after a collision with a surface segment could be presented as follows:
v refl = 0(ion is absorbed) , if x  P refl 4-17
v refl = v sp   1 – R  + v rand  R if x  P refl 4-18
where:
• v sp is the ion velocity after specular reflection.
• v rand is the ion velocity after random reflection.
• x is a random number.
• v sp = v rand = v i , where v i is the velocity of incident ion.
Each absorbed ion is used to compute the incoming flux F i at the surface segment. The
following characteristics describe the flux.
N norm = N abs  I  N traj 4-19
where:
• N norm is the normalized number of absorbed particles,
• N abs is the number of absorbed particles,
• N traj is the number of trajectories specified by the MC.PARTS1 and MC.PARTS2 parameter
for each type of plasma particles, and by the MC.POLYMPT parameter for polymer
particles,
• normalized normal v and tangential v  abs velocity components of the absorbed
abs
particle before the encounter with the surface:
 v
I
v abs = ------------  4-20
N traj
N abs
 v
I
v  abs = ------------  4-21
N traj
N abs
• normalized kinetic energy of absorbed particles:
v
2 I 2
v abs = ------------  4-22
N traj
N abs

Calculation of Polymer Fluxes

After ion and neutral fluxes are determined, the fluxes of the polymer particles are calculated
as follows. As the result of ion flux interaction with the surface segment, the polymer
particles are generated. The angular distribution of the polymer particles is uniform and the
current density of these particles is determined by the etch model (see “Linear Etch Model”
on 253) and the sum of the fluxes from incoming ions, neutrals, and from polymer particles
ejected from other surface segments. Obviously, the latter flux needs to be pre-calculated.
This flux is computed as follows. First, the configuration (or geometrical) factors, are
calculated. These factors are the fractions of the number of particles ejected from one
segment and absorbed by the other one. These are calculated using the same trajectory tracing
algorithms, which are described above for the incident ions and neutrals with the only one
difference: starting points are not at the upper boundary of the simulation area but at the
surface segments. Then, an iteration process is initialized. At the first iteration, only the
incoming ion and neutral fluxes are used for calculation of the ejection rates from each
surface segment. Knowing the current densities of ejected particles and the configuration
factors, the polymer fluxes are calculated. At subsequent iterations, the polymer fluxes
calculated at the previous iteration are used to update the etch and ejection rates. The
iterations are repeated until etch and ejection rates converge.
Calculation of Rates
The second stage involves calculation of the etching rates as well as ejection and redeposition
rates of the polymer particles. During each time step, the two processes simultaneously take
place on each surface segment. The first is redeposition of the polymer with the rate equal to
the polymer flux. The second is etching by incoming ions and neutrals. The combination of
these two processes can be treated as deposition of a virtual polymer layer with subsequent
etching of the two-layer structure. If the etch rate of polymer by incoming ions and neutrals is
less than the polymer deposition rate, the result is the redeposition of a polymer layer on the
surface. If the etch rate of polymer by incoming ions and neutrals is larger than the polymer
deposition rate, the result is actual etch of the underlying material.
Linear Etch Model
In the case of the linear model, the etch rate ER(m) of each material m is calculated as
ER  m  =
 EP  m i   vabs 4-23
n
where n is the number of plasma ion types specified by the parameter ION.TYPES (n could be
equal to 1 or 2), EP(m,i) is the etch parameter for material m, and ion type i specified by
parameters MC.ETCH1 and MC.ETCH2, v abs is the ion velocity as calculated in Equation 4-22.
If calculated, ER(polymer) is less than the polymer flux (redeposition rate), PF the actual etch
rate, and ER is negative which corresponds to redeposition.
ER = ER  polymer  – PF  0 4-24
The corresponding ejection rate EJR is equal to the etch rate of polymer.
EJR = ER  polymer  4-25

When calculated ER(polymer) is larger than polymer flux, the actual etch rate is positive.
EP  m i 
ER = ER  m  – PF 
 EP
---------------------------------------
 polymer i 
4-26
n
The corresponding ejection rate is calculated as follows.

EJR = PF + ER 4-27
C-Interpreter
You can use the C-Interpreter to introduce different etch and ejection models. The following
parameters are passed to the C-Interpreter file and can be used for implementing the models:
number of ion types, the four characteristics of ion fluxes for each ion type (Equations 4-19-
4-22), PF, and surface material, m. Returned parameters are ER and EJR.
For example, you can simulate the wet etching by setting the etch rate to a constant positive
value depending only on the surface material. In this case, the trajectory tracing part of the
model is not needed. The number of trajectories can be set to one.
Uniform deposition can be simulated by the setting of a negative constant etch rate and by
specifying the redeposited material other than polymer in the ETCH statement. If the fluxes are
not used, as in the wet etching simulation, the void formed will eventually be filled with the
deposited material, because inside the C-Interpreter there is no way to determine if the current
surface segment belongs to the void or not. This obstacle can be overcome by simulating ion
fluxes and by setting the etch rate to zero if the flux on the surface segment is less than some
small threshold value.
Surface Movement
A sophisticated string algorithm is used to move all segments, according to the rates (positive
or negative) calculated at each time step. If the rate is negative, the surface moves outside and
the area is filled with redeposited material (by default, polymer). If the rate is positive, the
surface moves inwards and the area is filled with vacuum.

Reflow Model Elite Models
4.5 Reflow Model

A two-dimensional viscous reflow capability is included in Elite. The vitreous silica (e.g.,
oxide, BPSG,) are modeled as the viscous incompressible fluids, which are dynamically
deformed under the driving force of surface tension. The finite-element method is used to
solve the creeping flow equations for the chosen materials. With a 7-node triangle element as
the basic discretization unit, arbitrarily shaped 2D regions and surface curvatures are
automatically described. Using the built-in user defined material capability, you can simulate
multiple material combinations. The flow equation solver can be coupled with impurity
diffusion to simulate the impurity redistribution and oxide growth.
The reflow is invoked by setting the reflow flag in the DIFFUSE statement and by setting the
REFLOW flag in the MATERIAL statement to choose a specific material. Physical parameters
that are specific for various materials, such as viscosity and surface tension, are also given in
MATERIAL statement. Reflow will proceed according to the time and temperature given in the
DIFFUSE statement. The finite element solver are invoked by specifying the VISCOUS
parametrs and various numerical control parameters in the METHOD statement.
The viscous creep flow equations solved are as follows (see [100]):
2
 V = P 4-28
  V = –  ---------------- P
1 – 2
4-29
  
E -
 = -------------------- 4-30
21 + 
where V is the velocity, P the pressure,  the viscosity,  the Poisson’s ratio, and E the
Young’s modulus. The parameters v and E are specified through elastic constants as described
in Section 3.10 “Stress Models”.

Chemical Mechanical Polish (CMP) Elite Models
4.6 Chemical Mechanical Polish (CMP)

Chemical Mechanical Polish (CMP) is a module in Athena. To run CMP, you need to have the
license to use Elite. CMP is used to model wafer planarization using polishing pad and
chemical slurry characteristics. CMP is also used to circumvent two major problems. First,
the depth of focus of high numerical aperture lithography systems. Second, metal thinning
that can occur over non-planar topographies
The CMP module that is incorporated into Athena has two distinct models. The first is the
Hard Polish or Buzz Saw Model [101]. The second is the Soft Polish Model based on the
work of J.Warnock [102]. To access these models, use the Athena statements: RATE.POLISH
and POLISH. These statements are similar to those used for the Elite deposition, RATE.DEPO
and DEPOSIT, and the Elite etching, RATE.ETCH and ETCH. The RATE.POLISH statement sets
up the parameters for a particular machine, while the POLISH statement executes the actual
polishing step using the machine.
4.6.1 Hard Polish Model

The Hard Polish Model [101] simulates the grinding down of the topography based on a rate
calculated as a function of the pattern factor Pf of the surface. The higher the pattern factor,
the lower the polishing rate. Use the following formula.
R  x y  = MAX.HARD   1 – Pf  + MIN.HARD  Pf 4-31
The hard polish model parameters are MAX.HARD and MIN.HARD. Pf = 1 corresponds to a flat
surface. Pf is calculated from the topography by the formula:
 
Pf = 
   xi – xi – 1  y  Y   X total
max – Y
4-32
i
where (xi,yi) are points on the polished material surface, y denotes both yi and yi-1, Ymax is the
highest point of the structure and Y is the rate effective height calculated by the previous
rate multiplied by the current time step value. For Figure 4-12, a pattern factor will be as
follows:
 X1 +  X2
Pf = ---------------------------
- 4-33
X total
The rate for points at height Ymax are equal to the R calculated in Equation 4-31. Points below
Ymax have a rate cause the structure to polish to the y coordinate Ymax  Y. Therefore, the
structure becomes more planar as shown in Figure 4-13.

Figure 4-12 Illustration of the Hard Polish Model: Structure before Planarization
Figure 4-13 Illustration of Hard Polish Model: Structure after Planarization

A total amount of Y is always removed at each time step in the above fashion.
You can mix the hard polish model with the soft polish model and isotropic etch component
by specifying the ISOTROPIC parameter of the RATE.POLISH statement.

4.6.2 Soft Polish Model

The soft polish model is based on the work of J.Warnock [102]. It has four parameters: SOFT,
LENGTH.FAC, HEIGHT.FAC, and KINETIC.FAC. SOFT is the polish rate on a flat surface.
LENGTH.FAC is the horizontal deformation scale in microns. It is a measure of the polishing
pad’s flexibility. It describes the distance at which shadowing will be felt by a “tall” feature.
HEIGHT.FAC is the vertical deformation scale in (microns). This measures how much the
polishing pad will deform with respect to the height of the feature. KINETIC.FAC increase the
vertical polish rate as the surface becomes more vertical.
The following formula gives the polishing rate.
Ki Ai
P i = ----------
- 4-34
Si
Ki is the kinetic factor or horizontal component of the polish removal rate at point i. Ai is the
accelerating factor of point i and is large for points that are higher and shadow other points. Si
is the shadow factor and decreases the polish rate as a function of the points that are above
point i. For a flat surface, Ki Ai/Si = 1. Following the work of Warnock, these three factors are
calculated using the following set of equations.
The shadow factor is the one for flat surfaces. But it is generally calculated basedon one or
two points that shadow point i and is given by the equations below.
 zi
S i = exp  -------------------------------------- 4-35
 HEIGHT.FAC
zi >= 0.0, so Si > 1. zi is obtained by integration over the surrounding topography.
i right
ri
z i  cosh  ---------------------------------------
 zi =   LENGTH.FAC
4-36
i left
In these equations, ileft and iright refer to the two points that can possibly shadow point i. The
effect of these shadow points depends on the two parameters LENGTH.FAC and HEIGHT.FAC
as shown in the equations. The variable zi is the vertical distance between the point i and the
point ileft/iright. The variable ri is the horizontal distance between the point i and the point ileft/
iright.
The acceleration factor, Ai, is given by the equations below. Ai is calculated for the two points
that shadow point i. In this manner, multiple shadowing effects are taken into account through
the term Aileft/Airight. This is the acceleration factor for the point(s) that shadow. If point i
shadows some other point j in the system, it will increase Ai by a similar factor. This increase
is then passed on to Aileft.
A ileft = A ileft + A i  B i   1 – 1  S i  4-37
A iright = A iright + A i  B i   1 – 1  S i  4-38

The constant Bileft/Biright is a weighing factor based on the amount of shadowing at point i due
to shadow point ileft/iright.
The kinetic factor is based on the following equation.
K i = 1 + KINETIC.FAC  tan  i 4-39
This shows the effect of the parameter, KINETIC.FAC, on the polishing rate. The angle i is
the local angle that is tangent to the polished surface. The maximum allowable angle i is
89.9544° (1.57 radians) will avoid calculation errors.
Figure 4-14 demonstrates three regions where each of the components of the polishing rate
would be large.
Figure 4-14 Soft Polishing Model: Areas where different components dominate

Chapter 5
Optolith Models
Overview Optolith Models
5.1 Overview
The Optolith module of Athena allows the use of sophisticated models for imaging,
photoresist exposure, photoresist bake, and photoresist development. Optolith includes a
library of photoresists with default characterizations for development and optical properties.
These default characterizations can easily be tuned to adjust for variations that very typically
occur from one facility to another. This chapter describes the models and capabilities of
Optolith.

The Imaging Module Optolith Models
5.2 The Imaging Module

Optolith includes an imaging module that utilizes the Fourier series approach.
The theoretical resolution (RES) and Depth Of Focus (DOF) of a microlithographic exposure
system are approximated by:

RES = k 1  -------- 5-1
NA
and

DOF = k 2  --------------- 5-2
2
 NA 
where is the wavelength of the exposing radiation, NA is the Numerical Aperture of the
imaging system, and k1 and k2 are process dependent constants. Typical values for k1 are 0.5
for a research environment and 0.8 for a production process; the value usually assigned to k2
is 0.5.
We shall discuss the basic assumptions upon which the model rests. Next, we shall derive the
principal equations used for calculation of the image irradiance distribution for objects
illuminated by partially coherent light.
The treatment presented here assumes the radiation incident on the object to be quasi-
monochromatic, which means that the spectral bandwidth is sufficiently narrow so that
wavelength-dependent effects in the optics or in diffraction angles are negligible. The source
is of a finite spatial extent so that the advantages of spatial incoherence are realized in
imaging.
The mask is completely general in that phase and transmission are variable, but it must be
composed of rectangular features.
The calculation of the diffraction phenomena is based upon the scalar Kirchhoff diffraction
theory. Since the dimensions of the mask are almost the same as the illumination wavelength,
we can ignore any polarization taking place as the radiation propagates through the mask.
We assume scalar diffraction, which means neglecting the vector nature of the radiation. This
is acceptable if all convergence angles are small.
According to Watrasiewicz [103], who experimentally investigated the limiting numerical
aperture, the breakdown of the scalar theory occurs at angles of convergence greater than 30°,
which corresponds to a numerical aperture of 0.5. Similar results were published by Richards
and Wolf [104], who used theoretical calculations to investigate the electromagnetic field
near the focus produced by an aplanatic system working at a high convergence angle. They
also found appreciable departures from scalar theory for convergence angles larger than 30°.
Since the convergence angles are calculated in air, we can assume that the accuracy of this
model is even better inside the photoresist, where angles are reduced in accordance with
Snell’s law.
Consequently, it can be stated that the scalar diffraction theory gives a reliable limit for
imaging system numerical apertures of 0.5.

The approach used for calculating the image irradiance distribution is based on the work of
Hopkins [105] and [106], which showed the partially coherent illumination of the object
structure can be simulated by the incoherently illuminated exit pupil of the condenser. The
exit pupil serves as an effective source, which produces the same degree of coherence in the
illuminated object plane as the actual condenser system. The degree of coherence in the
object plane is therefore determined by the shape and angular size of the effective source. The
condenser system is assumed to be diffraction limited, that is, free of aberrations. Residual
aberrations of the illuminator do have an appreciable influence on the final image for Koehler
type illumination systems as shown by Tsujiuchi [107].
Figure 5-1 shows a schematic diagram of a generalized optical system. The actual source and
the condenser system are replaced by the equivalent effective source having an irradiance
distribution of g (x0, z0). The effective source for the object plane U is taken to lie in the exit
pupil reference sphere of the condenser lens. This means that directing from arbitrary points
(x0, z0) on the effective source, plane waves propagate towards the object plane U having
irradiance values of  (x0, z0).
source
(condensor) reticle plane Imaging system Image plane
X X U′
P P′
h h′
α α′
α0
h0
P0
E E′
X U
Figure 5-1 Schematic Diagram of a Generalized Optical System

The reduced coordinates [106] on the object plane are defined as follows:
2
u = ------  n  sin    5-3

2
v = ------  n  sin    5-4

where  and  are the Cartesian coordinates of the object plane, 2/ is the absolute value of
the wave vector, and n-sin is equal to the numerical aperture (NA) of the imaging system.
Primed quantities indicate the corresponding coordinates and angles in the image space of the
projection system. The fractional coordinates on the object pupil spheres are defined as
follows:
x = --- 5-5
h
z = 
--- 5-6
h
where h is the radius of the pupil. The fractional coordinates of the exit pupil of the condenser
are given by
x 0 = --x- 5-7

z 0 = --z- 5-8

In these equations:
n 0  sin  0
 = -------------------------
- 5-9
n  sin 
where 0 and  are angular semi-apertures of the condenser and the objective respectively. n0
and n are the refractive indices in the image space of the illuminator and the object space of
the imaging system, usually both are set to one. The ratio  is the radius of the effective source
referred to the aperture of the objective and governs the degree of spatial coherence in the
object plane. The limits  0 and   correspond respectively to coherent and incoherent
illumination.
The object is taken to be infinitely thin. Therefore, a complex amplitude transmission
function can describe the object, which gives the change in magnitude and phase produced on
the radiation passing through it. The object has the complex transmission A(u,v). Its real part
is given by
  A  u v   =  1 in transparent areas 5-10

0 in opaque areas

The complex amplitude of the Fraunhofer diffraction pattern on the entrance pupil reference
sphere at E of the imaging system is given, apart from a constant factor, by
  A  u v   exp  –i  ux + vz   dudv
1
a  x z  = ------ 5-11
2
which is the inverse Fourier transform of the complex amplitude transmission of the object. If
not stated otherwise, integration ranges from to –  to + .
If the object is illuminated by an element dx0, dz0 of the effective source at (x0, z0) with its
amplitude proportional to   x0 z 0  , the object spectrum a(x,z) is then shifted by a
corresponding amount. In this instance, the complex amplitude distribution on the entrance
pupil sphere of the objective is
  x 0 z 0   a  x – x 0 z – z 0  5-12
The complex amplitude on the exit pupil reference sphere at E’ will be given by
a   x z  =   x 0 z 0   a  x – x 0 z – z 0 f  x z  5-13
In this equation, f(x,z) denotes the pupil function of the optical system. If the system has an
annular aperture, where the central circular obstruction has the fractional radius , the pupil
function has the form:

 0 2
x +z 
2 2

f  x z  =    x z   exp  i  k  W  x z   2
x +z 1
2 5-14

 1 2
x +z 1
2

t(x,y) is the pupil transmission, which is usually set to one, and W(x,z) denotes the wave-front
aberration. For an entirely circular aperture,  becomes zero. Note that the approach taken
here is somewhat similar to the one used in the investigations on phase contrast microscopy
[108].
The function W(x,z) gives the optical path difference between the real wave-front and the exit
pupil reference sphere. Commonly, the wave-front aberration is expanded into a power series
[106], giving
2 n
 Wl m n  
2 2 2
W  x z  = +   +  x + z + x + z  5-15
l m  n
for a particular position (x,z) in the exit pupil.  and  denote the fractional coordinates of the
image field. The values of l, m, and n describe the order of aberrations, while the coefficients
W(l,m,n) determine the magnitude of the aberrations.
The aberration coefficients up to the ninth order of aberration are specified in the
ABERRATION statement.

For third order aberrations l, m, and n take the values:

l=0 , m=0 , n=2 : spherical aberration
l=0 , m=1 , n=1 : coma
l=0 , m=2 , n=0 : astigmatism
l=1 , m=0 , n=1 : field curvature
l=1 , m=1 , n=0 : distortion
l=0 , m=0 , n=1 : defocus
where isoplanatism is assumed for the particular section of the image field for which the
irradiance distribution is calculated. The coefficient W001 can be determined from
2
 n   sin  
W 001 =   -------------------------------- 5-16
2
where  refers to the distance of the defocused image plane to Gaussian image plane.
The resulting amplitude in the image plane due to a wave coming from the point x0, z0 of the
effective source is

1
A   x 0 z 0 ;u  v   = ------    x 0 z 0   a   x – x 0 z – z 0   exp  i  u  x + v  z  dxdz 5-17
2
where (u,v) refers to a point in the image plane. The irradiance distribution associated with
the illuminating wave of the effective source will then be represented by
2
dI   x 0 z 0 ;u  v   = A   x 0 z 0 ;u  v   dx 0 dz 0 5-18
Since, by definition, the effective source is equivalent to a self-luminous source, the total
irradiance at (u,v) can be obtained by integrating over the entire source 
2
I   u  v   =   A   x 0  z 0 ; u  v   dx 0 dz 0 5-19

where  indicates the area of the effective source for which (x0,z0) has non-zero values.
For this purpose, Equation 5-19 is put into the form:
    x 0  z 0     x 0  z 0 ; u  v  
2
I   u  v   = d x 0 dz 0 5-20

where:
  a   x –x0 y – y0   f  x y   exp  i  u  x + v  y  dxdy

1
  x 0 y 0 ;u  v   = ------  5-21
2
  x 0 z 0 ; u v 
is proportional to the intensity at the point (u,v) due to a wave of unit
irradiance passing through (x0, z0) of the effective source.

In the case of an annular shaped source, x0, z0 has the form:
 2 2 2
 0 for x 0 + z 0   0

  x 0 z 0  =  1 for x 2 + z 2  1 5-22
 0 0
 2 2
 0 for x 0 + z 0  1
where 0 is the fractional radius of the centered circular obstruction in the exit pupil of the
condenser lens. For a circular exit pupil, 0 becomes zero.
Equation 5-20 is the principle relation of a generalized Abbe theory, where the image
formation under partially coherent illumination of the object is accounted for by a
combination of coherent imaging processes for perpendicular and obliquely incident
illuminating plane waves on the object. Since only the image irradiance is of interest, it can be
determined without using of coherence theory [106]. For the computation, the whole source is
divided into a number of luminous point sources considering the imaging due to each source
as an independent coherent image formation process. The contributions from each point
source do not interfere, so the net image irradiance is the sum of the irradiance from each
source point.
The normalization used throughout this investigation is that the mask is illuminated with unit
irradiance so that the ideal image has unit irradiance, where unit magnification is assumed.
Therefore, the brightness of the source decreases as its size increases. Equation 5-20 is the
principle equation of the algorithm, which is used for studying the influence of annular
apertures.
The object spectrum (see Equation 5-11) is calculated analytically, and the coherent image
(see Equation 5-18) is calculated using a Fourier Series approach.
The shape of a single mask feature must be rectangular. This is because the Fourier transform
for a rectangular feature is calculated based on an analytical formula.
Since the Fourier transform is linear, you can compose arbitrary shaped mask features from
the rectangular components. The object spectra of the single mask features (components) are
simply added up. The treatment can then be considered as appropriate, and no numerical
discretization errors in the size and placement of the mask features can occur.
Note: You may use MaskViews to create or import masks of any arbitrary shape. The mask layout will be sliced
(divided) on rectangular elements when it is imported into Optolith. Optolith can import masks containing
any number of mask elements.

Optical System Optolith Models
5.3 Optical System

Figure 5-2 shows the optical system used by Optolith. The meshes in the Fourier and Image
planes are totally independent. There is no mesh in the object or reticle plane.
α’ α
source condensor reticle projection aperture projection image plane

lens stop lens
Figure 5-2 The Generated Optical System
5.3.1 Discretization Errors

The size of the window in the reticle plane is determined by the number of mesh points in the
projector pupil, the numerical aperture, and by the chosen wavelength:
CW = NP  lambda  NA 5-23
where:
• CW is a computational or sampling window (mask or image cell) in the object or reticle
plane.
• NP is the number of mesh points in the projector pupil.
• NA is the numerical aperture of the stepper.
• lambda is the chosen wavelength.
For an i-line stepper with NA = 0.54, the size of the sampling window is the square whose
side length is equal to 6.8 m (10 · 0.365/0.54). No mask feature should exceed this
dimension.
You can increase the size of the sampling window for this particular stepper to any size by
increasing the number of mesh points in the projector pupil. This will be done automatically
to accommodate the mask and image windows that were specified.
Mask features cannot be placed outside of the sampling window. As mentioned earlier, the
image mesh is totally independent of the mesh in the Fourier plane. This allows you to
arbitrarily specify the number and distance of image points.

Optical System Optolith Models
5.3.2 Mesh
The size of the computational window is determined by Equation 5-23) and the position of
the mask points. By positioning this window so that the mask cell in the object plane is
covered, multiple image cells can be calculated.
5.3.3 Computation Time

To increase computation, first use a very coarse mesh for screening-type simulations and then
refine the mesh as you approach specific points of interest. Computation time is linearly
dependent on the number of source points, which is determined by the coherence factor.

The Exposure Module Optolith Models
5.4 The Exposure Module

The Exposure Module computes the intensity distribution in the photoresist through the
numerical solution of the Helmholtz equation (Equation 5-24).
2 2 2
 E+k n E = 0 5-24
Here, E is the electric field, n(x,y) is the complex refractive index of media, and k is the wave
number. The specific solution of this general equation is determined by a set of boundary
conditions. Generally, the target (substrate coated with the resist film) consists of an arbitrary
number of different materials with n in Equation 5-24 and interfaces between material
regions. Also, material refractive index can depend on the absorbed dose (i.e., n can vary with
the exposure time).
According to the electromagnetic theory, any field distribution can be represented by unique
set of plane waves. Generally, such a set has an infinite number of terms. But, the main
contributions are the incident wave and reflections from interfaces. Therefore, the calculation
of the field inside the complex structure can be divided into several subsequent calculations
of the plane wave propagation through the structure. The final field distribution is the sum of
distributions obtained for all separate plane waves. The simulation algorithm is outlined as
below.
1. Simulation of the incident field propagation through the target (resist + substrate)
2. For each segment of each interface:
• compute the direction of the reflection,
• align the coordinate system and simulation domain with the direction of the
reflection,
• obtain the field distribution over the segment,
• recursion of the propagation simulation procedure for the field generated over the
surface segment.
3. Summation of the distributions obtained with all recursive steps.
The reflection from each segment of interface is computed, using the preliminary obtained
field on the segment as the initial (incident) field. The recursion depth can be specified as a
simulation parameter.
The Beam Propagation Method (BPM) [109] is used to simulate the field propagation. The
BPM can be used for different types of radiation (e.g., UV, EUV, X-ray) as well as for multi
exposure processes and multilayer and non-linear resists.
There are three reason why we choose this method. The first reason is because the diffraction
of the field along the propagation is automatically taken into account. The second reason is
because it includes a capability to simulate non-linear effect of the intensity distribution on
the local optical properties of the resist material. The third reason is because it provides a
good “accuracy to run time” ratio.
The BPM is used to solve the Helmholtz equation for electromagnetic field inside the
structure. During the simulation, the field distribution is formed as the superposition of
incident light with all the reflections from all elements of the resist-substrate interface and
secondary reflection(s) from the upper resist surface. [109] shows the formal descriptions of
the BPM.Papers [110], [111], and [112] also describe some applications using BPM.

In this model, the Helmholtz equation (Equation 5-24) for the electric field, E, in the media
with complex refractive index, n(x,y), is solved in two main stages. The first stage, the
diffraction over a small spatial step along the propagation is calculated. Thus, obtaining the
new field amplitude distribution without absorption taken into account. Then, the actual field
distribution is computed as a product of this amplitude distribution and the distribution of the
complex absorption over the step. Let the wave propagate along the y-axis.
We find the solution as a quasi-plane wave E=A(x,y)exp(inky) with a slowly varied amplitude
A. A is then modified with y, which is slower than phase term inky. In this case, the Fourier
image of current distribution A in the plane y=y0 is defined as follows.
FA  kx  =
 A  x y0  exp  –ikx x  dx 5-25
After propagating over a small step, y, each component of FA obtains additional phase shift
corresponding to the value of k y  k  k x . Thus, the amplitude distribution at new y

2 2
(without accounting for absorption) can be written as follows.
A  x y 0 + y  =  FA  kx  exp  iky y  exp  ikx x  dkx 5-26
Since actual material optical properties differ from the properties of vacuum, the field at the
new plane is computed simply as follows.
E  x y 0 + y  = A  x y  exp  ik  n  x x  – 1 y  5-27
The algorithm is repeated recursively step-by-step over all simulation domains. The same
calculations are then applied to reflections from all interface segments. The current intensity
distribution is calculated from the field distribution as
2
I  x y  = E  x y  5-28
During the exposure, the resist structure is modified. So, the dissolution inhibitor is converted
to the photo reaction product. The initial normalized concentration of photoactive compound
(PAC) is defined by local intensity magnitude as:
M PAC = exp  – CI  x y   5-29
where I(x,y,) is the current intensity distribution, and C is Dill’s C-parameter. Accordingly, the
optical properties of the resist (complex refraction index n, which includes both refractive and
absorption indices) are modified too. The capability to take into account is the effect of dose
(intensity) on the refraction index n, which is implemented into the module and the following
approach is realized. Non-linear dependence of n on dose is defined through the PAC
concentration as
= n un exp osed + n  1 – M PAC  x y   5-30

Here, nunexposed is the complex refraction index of the unexposed resist. n=nexposed -
nunexposed is the difference between values of n for completely exposed and unexposed resists.
Current intensity distribution is calculated after each simulation of direct propagation and all
the reflections from interfaces with BPM. Then, current MPAC and n(x,y) are calculated using
(Equation 5-29) and (Equation 5-30) respectively for each point of the resist volume. The new
values of n(x,y) are used during next recursion of the field and intensity simulations. Thus, the
resulting intensity distribution is obtained as an accumulation of intermediate results.
You can specify the optical properties of the simulated material in the OPTICAL statement.
You can specify the refraction and absorption indices for unexposed resist (and/or for any
other material) using the REFRAC.REAL and REFRAC.IMAG parameters respectively. To
specify the difference of the refraction index for the completely exposed resist from the
unexposed one, use the DELTA.REAL and DELTA.IMAG parameters. If this difference isn’t
specified, the effect of intensity on the resist refraction index will not be taken into account
during the simulation. The number of recursions to obtain the intensity distribution in both
cases with and without taking into account dose-to-n effect is specified in the EXPOSE
statement as NUM.REFL. The maximum dose that corresponds to completely exposed resist is
specified with DOSE parameter.
The examples below show how to use the OPTICAL and EXPOSE statement to specify
parameters for the exposure module.
Example1: Resist exposure with accounting dose effect
OPTICAL NAME.RESIST=CURREN_RESIST I.LINE REFRAC.REAL=1.4
REFRAC.IMAG=0.02 DELTA.REAL=-0.2 DELTA.IMAG=0.01
EXPOSE DOSE=200 NUM.REFL=5
Example2: Resist exposure without accounting dose effect
OPTICAL NAME.RESIST=CURREN_RESIST I.LINE REFRAC.REAL=1.4
REFRAC.IMAG=0.02
EXPOSE DOSE=200 NUM.REFL=5

Photoresist Bake Module Optolith Models
5.5 Photoresist Bake Module

Post Exposure Baking (PEB) of the photoresist has been demonstrated to dramatically reduce
standing wave fringes of the developed resist image resulting from optical interference of
monochromatic illumination. This effect is generally accepted to be a result of bulk diffusion
of the PAC and photo reaction products.
The simple physical model, which is adopted here to describe the PEB, is that just one
chemical constituent of the resist diffuses. This constituent is generally assumed to be PAC or
the dissolution inhibitor, which diffuses according to the diffusion equation with the diffusion
constant D being independent of time, concentration and location.
dM
-------- =    D  M  5-31
dt
M is the PAC concentration and t is the PEB time. For a more general discussion, see [113]. M
is calculated by solving the two dimensional diffusion (Equation 5-31).
The diffusion length can be related to the bake time t and the diffusion coefficient D.
2
2tD =  5-32
For a PEB of 60 seconds at 125ºC, a diffusion length in the range of 0.04 <  < 0.06 microns
would be appropriate.
PEB can also be specified with parameters temperature and time. The diffusivity D is given
by the equation:
D = D o exp  – D E  kT  5-33
You can specify the D0 and DE parameters in the RATE.DEVELOP command.

“Reflective” boundary conditions at the air/resist interface and at the resist/substrate interface
must be incorporated to ensure that the total amount of dissolution inhibitor in the resist is
conserved. M(x,y,t) is extended into regions outside the resist by reflection at the planar
interfaces.
A post development bake is also available. It models a physically based reflow of the
photoresist.

The Development Module Optolith Models
5.6 The Development Module

The development model is based on the knowledge of the PAC distribution or dissolution
inhibitor in the resist layer after exposure and post-exposure bake. In classical Novolac
resists, the dissolution inhibitor and the PAC are usually part of the same molecule. In
chemically amplified resists, the reaction kinetics are more complicated. The inhibitor
concentration still, however, is considered to be the key quantity for the development process.
In positive tone Novolac resists, to determine the inhibitor concentration from exposure
simulations, use Dill’s model as previously described. This model applies when the resist
material undergoes a transition between two chemical states during the exposure step. The
actual development process is treated as a surface limited etching process, which is dependent
on the particular resist-developer chemistry and on the local concentration of the dissolution
inhibitor at the surface of the resist that has been decomposed to a degree during the exposure
step. If the resist developer chemistry is held constant, the dissolution rate is assumed to be a
function of the inhibitor concentration only.
The rate function r(x,y) is determined experimentally and usually fitted by an empirical
function to experimental development rate data as a function of the remaining PAC
concentration M(x,y).
You can use one of the following models to simulate the development process for the specific
resist-developer combination.
• Dill
• Kim
• Mack
• Trefonas
• Hirai
Each model assumes a specific rate function type to describe the rate-inhibitor concentration
relation. These models are described in the following sections.
5.6.1 Dill’s Development Model

The Dill model [114] uses the parameters E1, E2, and E3. Surface induction effects are not
considered. The bulk development is given by:
2
R  x y  = exp  E 1 + E 2  M  x y  + E 3  M  x y   5-34
and for M(x,y)  0.4:

R  x y  = exp  E 1 + E 2  0.4 + E 3  0.16  5-35

5.6.2 Kim’s Development Model

The Kim model [115] describes the development rate through the function
dy
R Induction = 1 –  1 –  R 5 –  R 5 – R 6 M  x  y z     exp – ---------- 5-36
R4
1
R Bulk  x y  = --------------------------------------------------------
- 5-37
1 – M  x  y M   x y -
----------------------------- + -------------------

R1 R2
M   x y  = M  x y   exp  – R 3  1 – M  x y    5-38
R  x y  = R Induction  R Bulk 5-39
R  x y  = R Induction  R Bulk 5-40
where RBulk is the bulk development rate and RInduction is the surface induction factor. The
limiting development rate values are R1 and R2 respectively for completely exposed and
unexposed resist. The function RInduction(x,y) is an empirical relationship describing the
reduced dissolution rate at the surface of a resist layer and is a function of the normal distance
from the original surface of the resist d(y) and the amount of remaining PAC, M(x,y). The
parameter R4 is the characteristic length along this path for the induction effect. The
parameters R5 and R6 are respectively the ratio of the surface rate to the bulk rate for a
completely exposed resist, and the ratio of surface rate to bulk rate for an unexposed resist.
5.6.3 Mack’s Development Model

The Mack model [116] describes the development rate through the function
n
a + 1   1 – M  x y   + R
R  x y  = R max  ----------------------------------------------------- 5-41
n min
a +  1 – M  x y  
n+1 n
a = -------------   1 – M th  5-42
n–1
where the parameter n is a selectivity parameter describing the sensitivity of the developer to
the exposed photoresist. The Mth parameter is the threshold PAC concentration. The Rmax
parameter is the development rate of a completely exposed resist. The parameter Rmin is the
development rate of totally unexposed resist.
5.6.4 Trefonas’ Development Model

The Trefonas development rate model [117] requires only two parameters.
q
R  x y  = R 0   1 – M  x y   5-43
where R0 is the development rate for unexposed photoresist and q is sensitivity.

5.6.5 Hirai’s Development Model

The development rate model by Hirai [118] is very similar to the one by Trefonas. The rate
function of the Hirai model is given by

R  x y  = R 0   1 – M  x y   + R C 5-44
where R0 is the development rate for fully exposed photoresist, RC is the rate for unexposed
resist material and  is a reaction constant.

Proximity Printing Optolith Models
5.7 Proximity Printing

Alongside with the “standard” projection imaging simulation, Optolith includes an additional
module for simulation of proximity printing. The proximity 11 printing (i.e., the imaging
without any reduction lens) is used to print relatively big features of a micron scale. This
method is still practical and in some cases could be cost effective. Internally, projection and
proximity printing modules use different simulation techniques. A common syntax and user
interface, however, allow you to switch easily from one module to another. Some specifics of
the proximity lithography and a brief theory used in the proximity printing module are
described below.
5.7.1 General Description of Proximity Lithography

Proximity lithography is used to print images of masks without expensive projection systems.
It has internal resolution limits but still applicable for relatively large objects. Figure 5-3
shows the optical scheme of proximity lithography.
Figure 5-3 Scheme of proximity optical system.

Light illuminating the mask creates a diffraction image in resist film, placed on some distance
(gap) from the mask. Due to diffraction effects in the gap, the image in the resist film is a
distorted mask image. Distortion depends on the gap size and on the radiation wavelength and
size of printed features. The reliability and longevity of the mask also depends on the gap.
Therefore, identification of the optimal printing conditions is a very important technology
design task.

5.7.2 Theory of Proximity Printing

A formal description of diffraction on a gap can be found in any classical book on physical
optics. The Fresnel diffraction is applicable here. Usually, the Fresnel diffraction integral is
obtained by applying Huygens diffraction approximation. It can be derived from the “first
principles” by applying specific limitation to common wave equation. The goal is to obtain
distribution of the light intensity over the resist surface. The real resist exposure with the light
absorption into resist, reflections from materials interfaces, and following photo-chemical
resist modifications are considered the same way as for the projection printing.
The common wave equation is
2 2
2 2  E + ---------
 E + k E = -------- E + ---------
E + k2 E = 0 5-45
2 2 2
x y z
where E is the electromagnetic field and k is the wave-vector.

Because the normal incidence on the mask is considered and propagation of light from the
mask to resist surface is to be calculated, it is convenient to represent
E  r  = A  r  exp  ikz  5-46
where A is amplitude of electromagnetic field and z is direction along the light propagation.
Substitution from Equation 5-46 into Equation 5-45 gives
2 2 2
---------
A + ---------
A + --------- A
A + 2ik ------ = 0 5-47
x
2
y
2
y
2 z
Obviously, A is modified with z weakly. Therefore, you can neglect the third term in Equation
5-47. As the result, Equation 5-47 can be rewritten as
2 2
---------
A + --------- A
A + 2ik ------ = 0 5-48
x
2
y
2 z
Now Equation 5-48 looks as a diffusion equation with complex diffusion coefficient, where z
replaces diffusion time. It means that the propagation along the z-direction can be formally
considered as a “diffusion” of complex amplitude A in the x,y-plane with a complex
“diffusion coefficient”.
The known -solution of traditional diffusion equation
2 2
 f + --------
 f – -----
1- f
-------- + 2ik ---- = 0 5-49
x
2
y 
2 2 t
is as follows
1  x 2 + y 2
f 0 = --- exp  – -----------------
- 5-50
t  4 2 t 
The substitution in Equation 5-48 results in:

2 2
A 0 = --- exp  – ik ------------------
1 x +y
5-51
z  2z 

The solution for given “initial conditions” (i.e., amplitude A immediately after (thin) mask
should be equal to the mask transparency T(x,y)) can now be expressed as a convolution of Ao
with T:
2 2
  x – Xm  +  y – Ym  

1
A  x y z  --- T  X m Y m  exp  – ik ----------------------------------------------------- dX m dY m 5-52
z  2z 
This expression exactly corresponds to the Fresnel diffraction.
5.7.3 Simulation Method

The Beam Propagation Method (BPM) is very efficient for tasks where a “one-way”
propagation of electromagnetic field is considered. Common description of BPM can be
found in [119], [120]. A simplified approach applicable to proximity printing conditions (i.e.,
a gap space filled with homogeneous matter, so refractive index does not depend on
coordinates in the gap) is described here.
The initial distribution of amplitude A in x and y (in the plane perpendicular to the
propagation direction z) can be obtained easily as a complex mask transparency function:
A 0 = A  x y z = 0  5-53
Then by applying the 2D Fourier transform over x and y, you can obtain the “angular”
distribution:
A  k x k y k z = 0  =
 A  x y z = 0  exp  i  kx x + ky y   dx dy 5-54
In this angular-spectrum domain, the relation between A and its z-derivative from Equation 5-
48 is given by the following expression [121]:
A 2 2 2
------  k x k y z  = i k – k x – k y A  k x k y z  5-55
z
It means that for each particular “direction”, a specific phase shift has to be applied to get the
value of A at a new z-position:
A  k x k y z +  z  = A  k x k y z  exp  ik z  z  5-56
In this equation, the z-component of the wave vector is derived by:

2 2 2
kz = k – kx – ky 5-57
Now the inverse Fourier transform will produce the distribution of the field amplitude over x,
y plane at z=z (i.e., at the gap distance from the mask).
Finally, the required intensity distribution is the square of the module of the field amplitude A.

Chapter 6
Statements
Overview Statements
6.1 Overview
Athena executes a file that describes the process, meshing, and models to be used in a
simulation. The contents of the file are statements, each of which prompts an action or sets a
characteristic of the simulation. This chapter is a reference to the command language that can
be used to control Athena.
Throughout this manual, we will refer to commands, statements, and parameters. A line in an
input file is referred to as a statement (or statement line).
An Athena statement is specified in the general format
<COMMAND> <PARAMETERS>=<VALUE>
where <COMMAND> is the command name, <PARAMETER> is the parameter name, and <VALUE>
is the parameter value. Four types of parameters are used in Athena — Real, Integer, Logical,
and Character. The space character is used to separate parameters from a command or from
other parameters.
Table 6-1 Types of Parameters
Type Description Value Required Example
Character An alphabetic, alphanumeric, or Yes OUTFILE=MOS.STR

numeric string
Integer Any whole number Yes DIVISIONS=10
Boolean A true or false condition No OXIDE or OXIDE=f
Real Any real number Yes C.BORON=1.5e14
Any parameter that does not have a logical value must be specified in the form PARAM=VAL,
where PARAM is the name of the parameter and VAL is the value of the parameter. Boolean
parameters must be separated from other parameters or commands with a space.
For example, in the statement line:
DEPOSIT NITRIDE THICK=0.35
the NITRIDE parameter has a Boolean value (true) and the THICK parameter has a value of
0.35(real).
Many parameters are provided default values. If a parameter is not specified, its default value
will be used. Table 6-1 explains the different types of parameters which may be used when
preparing an Athena input deck. The command language of Athena is not case sensitive and
can be entered using either upper case or lower case letters.
6.1.1 Abbreviations
It is not always necessary to input the entire statement or parameter name. Athena only
requires that you input enough letters to distinguish that command or parameter from other
commands or parameters. For example, DEPO can be used to indicate the DEPOSIT
command.

Overview Statements
6.1.2 Continuation Lines

Since it may be necessary for a statement line to contain more than 256 characters, Athena
allows you to specify continuation lines. If a statement line ends with a backslash (\), the next
line will be interpreted as a continuation of the previous line.
6.1.3 Comments
Comments are indicated by the COMMENT statement or a number sign (#). All characters on a
line which follow a comment indicator (COMMENT or #) will not be processed by Athena. The
comment symbol $ is not supported anymore. The $ should be avoided for use as a character
in strings since it is used as part of shell capabilities included in DeckBuild.
6.1.4 General Syntax Description

An Athena statement is a sequence of words starting with a statement name and followed by
some or all of the statement’s parameters. This manual describes the syntax for each
statement in the following way:
STATEMENT NAME
DESCRIPTION OF PARAMETER 1
DESCRIPTION OF PARAMETER 2 . . .
Parameters are described in the following form:
PARAM=<n>... a real valued parameter
PARAM=<c>... a string valued parameter
PARAM ... a Boolean parameter
Boolean parameters are those that recognize the Boolean values TRUE and FALSE as valid
values. In Athena, Boolean parameter values are automatically set to true if the name of the
Boolean parameter appears by itself in a statement. A Boolean parameter can be set to false
using the syntax: PARAM=FALSE or PARAM=F.
A mutually exclusive choice among parameters is indicated by parentheses around the
parameters and vertical bars between each parameter (PAR1|PAR2). Only one parameter in
such a group may be specified at a time. Specifying more than one parameter in a mutually
exclusive group is an invalid operation and will generally prompt a warning or error message.
Parameters that are optional to a statement are enclosed by brackets [ ]. Most parameters are
assigned default values and so defining them is optional. All parameters and parameter
values, however, should be checked in the context of the actual process that will be simulated
before relying on the results of any simulation.
String-valued parameters can be specified as a single word, e.g., INFILE=FILE1, or as a
sequence of words surrounded by double quotes, e.g., MATERIAL="Nickel Silicide".
Real-valued parameters can be specified as expressions involving numbers, numerical
constants, the operators +, -, *, /, and the functions listed in Table 6-2. If an expression
contains spaces, then enclose it in parentheses.

Overview Statements
Table 6-2 Functions
Function Description
abs5 Absolute value

active Active portion of the specified dopant
erf Error function
erfc Complimentary error function
exp Exponential
gradx Computes the approximate gradient in the x direction
grady Computes the approximate gradient in the y direction
log Logarithm
log10 Logarithm base 10
mat1@mat2 Returns the y value of the interface between mat1 and mat2 along a
vertical slice at the given location
mat1|mat2 Returns the x value of the interface between mat1 and mat2 along a
horizontal slice at the given location
scales Scales the value given by the maximum value
sqrt Square root
xfn Takes y and z and finds a matching x
yfn Takes x and z and finds a matching y
zfn Takes x and y and finds a matching z
Examples:
PAR1=<n>
PAR1 is a required numeric valued option.
PAR1=( 4.0 * EXP( -2.0 / (8.62E-5 * 1173.0) ) )
PAR1 is a required numeric valued option, assigned a real number expression.
[PAR2=<c>]
PAR2 is an optional character variable.
For further examples of expressions, see SET and EXTRACT in DeckBuild User’s Manual.

Overview Statements
6.1.5 Command Line Parsing

Athena supports expressions on the command line. For example:
DIFFUSE TIME=10/60 TEMP=1000
Be careful when using parentheses as the precedence of arithmetic operators, as in
programming languages, is not guaranteed in all cases.

Athena Statements List Statements
6.2 Athena Statements List

This chapter contains a complete description (in alphabetical order) of every statement used
by any of the Athena products. The following documentation is provided for each statement:
• The statement name
• A list of all of the parameters of the statement and their type
• A description of each parameter or group of similar parameters
• An example of the correct usage of each statement
The Athena command language encompassed by this document describes each of the
modules of Athena, namely: Elite, Optolith, SSuprem4 and their submodules. Depending on
which of the Athena modules have been purchased, some of the capabilities described may
not be available as part of the Athena installation.
Note: You can print a summary of statement names and parameters by using the HELP statement.
The following list classify Athena statements and provide their brief description and use.
6.2.1 Structure and Grid Initialization Statements

These statements define the dimensions, boundary conditions, grid density, and material type
of the initial structure. Typically, only LINE and INITIALIZE statements are required.
• BASE.MESH specifies parameters of the base mesh used for initial grid generation.
• BOUNDARY specifies which lines in a rectangular grid are exposed to gas.
• INITIALIZE sets up the initial grid and specifies background doping concentrations and
material type.
• LINE specifies the positioning of x and y grid lines for a rectangular mesh.
• REGION specifies corresponding sections of the rectangular mesh and material.
6.2.2 Structure and Mesh Manipulation Statements

These statements manipulate the geometry or attributes of the structure or create output files.
• ADAPT.MESH enables the adaptive meshing algorithm.
• ADAPT.PAR specifies adaptive meshing parameters.
• BASE.PAR defines adjacent mesh characteristics of an automated base mesh.
• ELECTRODE names electrode regions.
• GRID.MODEL defines a template file containing adaptive meshing commands.
• PROFILE causes Athena to read in an ASCII file of depth and doping data.
• RELAX loosens the grid within a user-specified area.
• STRETCH allows changes in structure geometry by stretching at a horizontal or vertical
line.
• STRUCTURE writes the mesh and solution information into a file. This is the main
output statement for generating program data to be plotted.

6.2.3 Simulation Statements

These statements apply physically based models for processing operations to the structure.
• BAKE performs post-exposure or post-development photoresist bake.
• DEPOSIT deposits a material layer.
• DEVELOP performs photoresist development.
• DIFFUSE performs a time/temperature step on the wafer and calculates oxidation and
diffusion of impurities.
• EPITAXY performs high temperature silicon epitaxial growth.
• ETCH performs a geometric or machine type etch on the structure.
• EXPOSE models photoresist exposure.
• IMAGE calculates a 2D or 1D aerial image.
• IMPLANT models ion implantation.
• POLISH simulates chemical mechanical polishing in the Elite module.
• STRESS computes the thermal elastic stresses.
• STRIP removes photoresist or another user specified material.
6.2.4 Model Statements

These statements are used to change model parameters and coefficients. The parameters are
described in the statement descriptions. When starting-up, Athena executes the model
statements in the file named athenamod located in the $SILVACO/lib/athena subdirectory
corresponding to the version number and system type of Athena that you are running. This
file contains the default parameters for most model statements.
• ABERRATION defines aberration parameters of the optical projection system.
• CLUSTER specifies parameters of {311} cluster model.
• ILLUMINATION describes the photolithographic illuminating system.
• ILLUM.FILTER defines filters used in the illumination source for photolithography.
• IMPURITY sets the coefficients of impurity kinetics.
• INTERSTITIAL sets the coefficients of interstitial kinetics.
• LAYOUT describes the mask reticle for imaging.
• MATERIAL sets the coefficients of various materials.
• METHOD sets the numerical options or models for solving the equations.
• MOMENTS specifies moments for Pearson implant model.
• OPTICAL specifies the coefficients of reflection and refraction.
• OXIDE specifies oxidation coefficients.
• PROJECTION defines the photolithographic projection system.
• PUPIL.FILTER defines filters in the pupil plane.
• RATE.DEPO specifies deposition rates for machine type deposits.
• RATE.DEVELOP specifies development rates and other photoresist parameters.
• RATE.ETCH specifies the etch rate for machine etches.
• RATE.POLISH specifies polishing parameters for definition of a polishing machine.
• SILICIDE sets the coefficients for silicidation reactions.
• TRAP sets the coefficients of trap kinetics.
• VACANCY sets the coefficients of vacancy kinetics.

6.2.5 Special DECKBUILD Statements

These statements invoke special operations when running under DeckBuild. For more
information on these statements, see the DeckBuild User’s Manual.
• AUTOELECTRODE defines layout-based electrodes.
• EXTRACT extracts parameters.
• GO indicates interfacing between simulators.
• MASK performs photoresist deposition and etching through the MaskViews interface.
• SET sets the value of a user-defined variable.
• SYSTEM allows execution of any UNIX C-shell command within an input file.
• TONYPLOT creates a plot using TonyPlot.
6.2.6 Post-processing Statements

Starting from version 4.0, all internal plotting capabilities of former SUPREM-IV have been
depreciated. Enhanced superior capabilities are available through TonyPlot and other VWF
Interactive Tools. Only the following two post-processing statements remain.
• PRINT.1D is used to print the values (data points and profile information).
• SELECT allows a variable to be chosen as the z coordinate for the PRINT.ID command
to follow.
6.2.7 Execution Control Statements

These statements control some execution capabilities. Some of them are useful only in a batch
mode when Athena is run outside DeckBuild.
• COMMENT is used to document the input file.
• CPULOG instructs Athena to output CPU statistics.
• FOREACH/END specifies the command looping facility.
• HELP prints summary of statement names and parameters.
• OPTION specifies the level of run-time output.
• PRINTF parses a string or expression and places result into standard output.
• QUIT terminates execution of Athena.
• SETMODE sets execution mode parameters.
• SOURCE causes Athena to read statements from the specified file.
• UNSETMODE sets execution mode parameters to false.
6.2.8 Obsolete Statements

The following statements existed in earlier versions of Athena. Their capabilities were
substituted either by superior capabilites of VWF Interactive Tools or included in other more
advanced or generic statements.
• ANTIMONY is substituted by the I.ANTIMONY parameter in the IMPURITY statement.
• ARSENIC is substituted by the I.ARSENIC parameter in the IMPURITY statement.
• BORON is substituted by the I.BORON parameter in the IMPURITY statement.
• COLOR - all plotting capabilities are now provided by TonyPlot.
• CONTOUR - all plotting capabilities are now provided by TonyPlot.
• ECHO is a synonym of the PRINTF statement

• DEFINE is substituted by the SET capability in DeckBuild.

• INDIUM is substituted by the I.INDIUM parameter in the IMPURITY statement
• LABEL - all plotting capabilities are now provided by TonyPlot
• PAUSE is substituted by the Pause button in DeckBuild.
• PHOSPHORUS is substituted by I.PHOSPHOR parameter in the IMPURITY statement
• PLOT.1D - all plotting capabilities are now provided by TonyPlot
• PLOT.2D - all plotting capabilities are now provided by TonyPlot.
• PLOT.3D - all plotting capabilities are now provided by TonyPlot.
• UNDEFINE is substituted by the SET capability in DeckBuild.
• VIEWPORT - all plotting capabilities are now provided by TonyPlot.
6.2.9 Standard and User-Defined Materials

Different materials can be specified as parameters in various statements. Both standard and
user-defined materials are available in Athena. The generic name MATERIAL appeared in a
statement syntax description signifies that you can only specify one of standard material
names from the list below or user-defined material. The generic name MATERIALS appeared
in a statement syntax description signifies that you can specify one or several standard and
user-defined materials.
The following shows the standard material names currently available in Athena.
Semiconductors
SILICON, POLYSILICON, GAAS, ALGAAS, INGAAS, SIGE, INP, GERMANIUM, SIC_6H, SIC_4H,
SIC_3C
Note: Since Athena parser doesn’t recognize parameter names that begin with numerals, non-standard names are
used for Silicon Carbides: SIC_6H, SIC_4H, and SIC_3C. Standard names: 6H-SiC, 4h-SiC, and
3C-SiC are used for these materials outside Athena (e.g. TonyPlot and Atlas).
Insulators
OXIDE, OXYNITRIDE, NITRIDE, PHOTORESIST
Metals
ALUMINUM, TUNGSTEN, TITANIUM, PLATINUM, COBALT
Silicides
WSIX (Tungsten Silicide), TISIX (Titanium Silicide), PTSIX (Platinum Silicide), COSIX
(Cobalt Silicide)
Special Materials
GAS is used only in the IMPURITY, INTERSTITIAL and VACANCY statements to specify some
parameters (i.e., segregation) at exposed boundaries.
BARRIER is a fictitious material. It can be specified only in DEPOSIT and ETCH statements and
serve as a masking material.

User-Defined Materials
User-defined materials can be specified by MATERIAL = <c>, where <c> could be a single word
MATERIAL=OXIDE1 or any string in double quotes as MATERIAL="MY INSULATOR". The user-
defined material with the names exactly corresponding to Silvaco standard material names are
saved in Silvaco Structure Files as those standard materials and will be recognized as such by
other tools, e.g., DevEdit and Atlas. The following is the list of those Silvaco standard materials
that can be used in Athena as previously described:
“Fictive GaAs”, AlInAs, AlAs, "Alpha Si 1", "Alpha Si 2", "Alpha Si 3","Alpha Si 4",
"AlxGa1_xAs_x_0.25", "AlxGa1_xAs_x_0.5", "AlxGa1_xAs_x_0.75", "InxGa1_xAs_x_0.50 Unstr",
"InxGa1_xAs_x_0.33 Str GaAs", "InxGa1_xAs_x_0.75 Str InP", "AlxIn1_xAs_x_0.50", Diamond,
AlP, AlSb, GaSb, GaP, InSb, InAs, ZnS, ZnSe, ZnTe, CdS, CdSe, CdTe, HgS, HgTe, PbSe, PbTe,
SnTe, ScN, GaN, AlN, InN, BeTe, InGaP, GaSbP, GaSbAs, InAlAs, InAsP, GaAsP, HgCdTe,
CdZnTe, InGaAsP, AlGaAsP, AlGaAsSb, SiN, Si, CuInGaSe, InGaN, AlGaN, InAlGaN, InGaNAs,
InGaNP, AlGaNAs, AlGaNP, AlInNAs, AlInNP, InAlGaAs, InAlGaP, InAlAsP, Pentacene, Alq3,
TPD, PPV, Organic, InAlP,CdZnTe, Tetracene, ZnO, InGaSb, InAlSb, GaAsSb, AlAsSb,
InGaAsSb, InAlAsSb, InPSb, InPAsSb, AlPSb, AlPAs, AlGaP, IGZO, NPB, InAlN,CuPc, NPD,
CBP, Irppy, BAlq, CdO, CDZnO, MgCdO, MGCdZnO, Sapphire, Vacuum, TEOS, BSG, BPSG,
PMMA, SOG, Polyimide, "Cooling package material", Ambient, Air, Insulator, Polymer, ITO,
HfO2, HfSiO4, ZrO2, TiO2, Al2O3, PML, Lens, SixNyHz, In2O3, Tio, SnO2, Gold, Silver, AlSi,
Palladium, Molybdinum, Lead, Iron, Tantalum, AlSiTi, AlSiCu, TiW, Copper, Tin, Nickel, WSix,
NiSix, TaSix, PdSix, MoSix, ZrSix, AlSix, Conductor, Contact, Ba2YCu3O7, Ba2NdCu3O7,
TiON, GST, GsSBTe, and PCM.
The generic name /MATERIAL specifies the second material in those statements, which
specify parameters related to the boundary between two materials.
6.2.10 Standard Impurities

Different impurities can be specified as parameters in various statements. The generic name
IMPURITY appeared in a statement syntax description signifies that you can only specify one
of impurity names from the list below. The generic name IMPURITIES appeared in a
statement syntax description signifies that you can specify one or more impurity names from
the list below in the statement simultaneously. The impurity names below can appear as is or
as a part of a parameter name, e.g. I.BORON, C.BORON and F.BORON. The following is the list
of standard impurity names currently available in Athena:
ALUMINUM, ANTIMONY, ARSENIC, BERYLLIUM, BORON, CARBON, CHROMIUM, FLUORINE,
GALLIUM, GERMANIUM, GOLD, HELIUM, HYDROGEN, INDIUM, MAGNESIUM, NITROGEN, OXYGEN,
PHOSPHORUS, SELENIUM, SILICON, and ZINC

ABERRATION Statements
6.3 ABERRATION
ABERRATION defines aberration parameters of the optical projection system.
Note: This statement is not available in Athena 1D.
Syntax
ABERRATION
[X.FIELD=<n>][Z.FIELD=<n>] [SPHERICAL=<n>] [COMA=<n>]
[ASTIGMATISM=<n>] [CURVATURE=<n>] [DISTORTION=<n>]
[FIFTH|SEVENTH|NINTH]
C1=<n>][C2=<n>][C3=<n>][C4=<n>][C5=<n>]
[C6=<n>][C7=<n>][C8=<n>][C9=<n>][C10=<n>][C11=<n>]
[C12=<n>][C13=<n>][C14=<n>][C15=<n>][C16=<n>][C17=<n>]
[C18=<n>][C19=<n>][C20=<n>]
Description
This statement specifies the aberration coefficients in the power series expansion of the wave
aberration function. Each coefficient is entered in fractions of a wavelength in the range 0  C
 0.5.
X.FIELD and Z.FIELD define or change the position in the image field for which the
irradiance distribution is to be computed. Note that the position is expressed in fractional field
coordinates, so that the values for the x and z directions vary between -1.0 and 1.0.
SPHERICAL specifies 0C40, the amount of third order spherical aberration present in the
power series expansion of the wave aberration function of the optical projector.
COMA specifies 1C31, which is the amount of third order coma present in the power series
expansion of the optical projector.
ASTIGMATISM specifies 2C22, which is the amount of third order astigmatism present in
the power series expansion of the optical projector.
CURVATURE specifies 2C20, which is the amount of third order field curvature present in
the power series expansion of the optical projector.
DISTORTION specifies 3C11, which is the amount of third order distortion present in the
power series expansion of the optical projector.
FIFTH, SEVENTH, and NINTH specify the aberration order. Coefficients for only one
aberration order can be specified on a single statement.
C1, C2, C3, C4, C5, C6, C7, C8, C9, C10, C11, C12, C13, C14, C15, C16, C17, C18, C19,
and C20 are described in Table 6-3. Coefficients for fifth, seventh, and ninth order aberrations
must be entered in separate ABERRATION commands for each order. Each of these
parameters represents a particular aberration coefficient depending on the order specified by
parameters FIFTH, SEVENTH, or NINTH.

ABERRATION Statements
Table 6-3 Aberration Coefficients
Parameter Fifth Seventh Ninth
C1 4C20 6C20 8C20
C2 2C40 4C40 6C40
C3 0C60 2C60 4C60
C4 5C11 0C80 2C80
C5 3C31 7C11 0C100
C6 1C51 5C31 9C11
C7 4C22 3C51 7C31
C8 2C42 1C71 5C51
C9 3C33 6C22 3C71
C10 4C42 1C91
C11 2C62 8C22
C12 5C33 6C42
C13 3C53 4C62
C14 4C44 2C82
C15 7C33
C16 5C53
C17 3C73
C18 6C44
C19 4C64
C20 5C55
Examples
If high order aberrations are to be studied, they must be entered on a separate command line
for each order.
ABERRATION X.FIELD=.5 SPHERICAL=.25
ABERRATION FIFTH C1=.25 C2=.5
ABERRATION SEVENTH C1=.3 C4=.4
For more examples, see IMAGE, ILLUMINATION, PROJECTION, ILLUM.FILTER,
PUPIL.FILTER, LAYOUT.

ADAPT.MESH Statements
6.4 ADAPT.MESH
ADAPT.MESH enables the adaptive meshing algorithm.
Syntax
ADAPT.MESH
[SMOOTH] [SMTH.COUNT=<n>] [ADAPT] [ADAPT.COUNT=<n>]
[ADD.I.LINE=<n>]
[SENSITIVITY] [MATERIAL] [/MATERIAL]
Description
This statement runs the adaptive meshing algorithm or the smoothing algorithm in standalone
mode.
SMOOTH flag to do mesh smoothing.
SMTH.COUNT specifies the number of smooth loops during the smooth operation. The
default is 1.
ADAPT flag to do stand alone mesh adapting. Specifies that a stand alone adaptive meshing
step should be performed to refine or relax the current mesh based on the material/impurity
specification given on ADAPT.PAR command The default is False.
ADAPT.COUNT specifies the number of adapting loops during the stand alone adaptive
meshing operation The default is 1.
ADD.I.LINE depth of the shadow interface mesh line in microns. Add the mesh line at the
interface between two materials as defined by the booleans MATERIAL and /MATERIAL. The
line is added in MATERIAL at a distance ADD.I.LINE from /MATERIAL.
SENSITIVITY specifies sensitivity of adaptation algorithm. The lower value leads to grid
with more triangles. The default is 1.0.
MATERIAL one of standard materials or user specified material (see Section 6.2.9
“Standard and User-Defined Materials” for the list of materials).
/MATERIAL one of standard materials or user specified material (see Section 6.2.9
Examples
The following statement will add a set of lines close to the silicon/silicon dioxide interface.
For this to work, the existing grid spacing at the interface must be greater than 0.005 microns.
Note that since the lines are added only between existing mesh lines and the interface the
lines must be specified in this order (i.e., getting closer to the surface)
ADAPT.MESH SILICON /OXIDE ADD.I.LINE=0.005
For more examples, see ADAPT.PAR.

ADAPT.PAR Statements
6.5 ADAPT.PAR
ADAPT.PAR specifies adaptive meshing parameters.
Syntax
ADAPT.PAR
[MATERIALS] [I.IMPURITIES] [I.INTERST] [I.VACANCY] [DISABLE]
[MAX.ERR=<n>] [MIN.ERR=<n>] [CONC.MIN=<n>] [AREA.MIN=<n>]
[AREA.MAX=<n>]
[EDGE.MIN=<n>] [EDGE.MAX=<n>] [MIN.ADD=<n>]
[MAX.POINT=<n>] [MAX.LOOP=<n>] [IMPL.SMOOTH] [DIFF.SMOOTH]
[IMPL.SUB]
[DOSE.ERR=<n>] [DOSE.MIN=<n>] [DIFF.LENGTH=<n>] [ANISOTROPIC]
Description
ADAPT.PAR specifies parameters used during adaptive meshing enabled by the
ADAPT.MESH statement.
MATERIALS specify standard materials or user specified material regions in which mesh
adaptation takes place (see Section 6.2.9 “Standard and User-Defined Materials” for the list
of materials). One or several materials can be specified at a time.
I.IMPURITIES specify the impurities to be used for the grid adaptation (see Section 6.2.10
“Standard Impurities” for the list of impurity names that can be used, e.g. I.BORON). You can
specify one or several impurities at a time.
I.INTERST specifies that interstitials to be used for the grid adaptation.
I.VACANCY specifies that vacancies to be used for the grid adaptation.
DISABLE specifies that the materials/impurities combinations given are disabled to be
effective on mesh adapting or smoothing. Default is false.
MAX.ERR specifies the maximum error allowable before adding points to the mesh,
(unitless). Error calculated above this value cause points to be added.
MIN.ERR specifies the minimum error below which points can be deleted from the mesh,
(unitless). Error calculated below this value will remove points. Both MAX.ERR and MIN.ERR
are calculated using the Bank-Weiser error estimator.
CONC.MIN specifies the minimum impurity concentration below which adapting will stop.
Units are cm-3. Default is 1.0 1014cm-3.
AREA.MIN specifies the minimum triangle area below which adding points will stop. Units
are cm2. Default is 1.0 10-15
AREA.MAX specifies the maximum triangle area below which deleting points will stop.
Units are cm2. Default is 1.0 10-11.
EDGE.MIN specifies the minimum edge length below which adding points will stop. Units
are cm. Default is 1.0 10-6.

ADAPT.PAR Statements
EDGE.MAX specifies the maximum edge length below which deleting points will stop.
Units are cm. Default is 1.0 10-5.
MIN.ADD percent criteria to turn off implant adapt loop. MIN.ADD stops point addition in
IMPLANT when the number of points added in the current loop is less than MIN.ADD * (total
number of points). The default value for MIN.ADD = 0.05.
MAX.POINT specifies the maximum number of points above which adapting will stop.
Default is 20000.
MAX.LOOP specifies the maximum loop count above which adapting will stop. This is only
effective with implant. Default is 10.
IMPL.SMOOTH specifies which annealing algorithm to use after each adaption step.
Currently, IMPL.SMOOTH=0 corresponds to no annealing during IMPLANT. IMPL.SMOOTH=1
corresponds to Laplacian smoothing and dose conservation interpolation algorithm. The
default is IMPL.SMOOTH=1.
DIFF.SMOOTH specifies which annealing algorithm to use after each adaption step.
Currently, DIFF.SMOOTH=0 corresponds to no annealing during DIFFUSE. DIFF.SMOOTH = 1
corresponds to Laplacian smoothing and dose conservation interpolation algorithm. The
default is DIFF.SMOOTH=0.
IMPL.SUB flag to do grid subtracting in implant adapt. IMPL.SUB is a boolean flag that
stops point removal during IMPLANT adaptive meshing. The default value for
IMPL.SUB=false signifies that points are not being removed.
DOSE.ERR specifies dose error for the refinement/unrefinement.
DOSE.MIN specifies minimum of dose level for grid refinement during adaptation.
DIFF.LENGTH used to limit the activity of adaptation of grid during the simulation of
dopant diffusion. This parameter will allow the mesh to adapt only after a given diffusion
length for a given dopant and will override any other adaptation triggers based upon gradient
error estimates. This is a useful control to limit the number of time steps. Units are microns.
ANISOTROPIC is the flag used to maintain the mesh to be anisotropic. The flag is material
dependent.
Examples
The following is an example of setting the adaptive meshing parameters during diffusion for
Boron.
IMPLANT BORON DOSE=1E15 ENERGY=60
ADAPT.PAR DIFF.LEN=0.1 SILICON I.BORON
DIFFUSE TEMP=1000 TIME=100 NITROGEN
For more examples, see ADAPT.MESH.

BAKE Statements
6.6 BAKE
BAKE performs post-exposure or post-development photoresist bake.
Syntax
BAKE
[DIFF.LENGTH=<n>] [TEMERATURE=<n>] [REFLOW]
[TIME] [SECONDS|MINUTES|HOURS]
[DUMP=<n>] [DUMP.PREFIX=<c>]
Description
This command runs a bake process using the diffusion length as the parameter that
incorporates the bake temperature and bake time.
DIFF.LENGTH specifies the diffusion length for the post exposure bake. Default is 0.05
micrometers
TEMPERATURE specifies the temperature of the bake process in oC.
REFLOW specifies that material flow should be calculated during the bake process.
Default is False.
TIME specifies the amount of time for the bake step in specified units. Default is MINUTES.
HOURS, MINUTES, and SECONDS specify the units of the TIME parameter.
DUMP and DUMP.PREFIX specify that a structure file be output at every DUMPth time step.
The files are readable with the STRUCTURE statement or can be displayed using TonyPlot. The
names will be of the form DUMP.PREFIX<time>.str, where <time> is the current total time
of the simulation.
Examples
The BAKE command is entered with the user-specified diffusion length for post exposure
bake:
BAKE DIFF.LENGTH=0.05
BAKE can also be entered with time and temperature parameters for post exposure bake:
BAKE TIME=45 TEMP=120
For photoresist reflow (post development bake), the above command is entered with the
REFLOW parameter:
BAKE REFLOW TIME=45 TEMP=120
For more examples, see DIFFUSE and RATE.DEVELOP.

BASE.MESH Statements
6.7 BASE.MESH
BASE.MESH specifies parameters of the base mesh used for initial grid generation.
Syntax
BASE.MESH
[SURF.LY=<N>] [SURF.DY=<N>] [ACTIVE.LY=<N>]
[ACTIVE.DY=<N>] [EPI.LY=<N>] [EPI.DY=<N>]
[SUB.LY=<N>] [SUB.DY=<N>] [BACK.LY=<N>] [BACK.DY=<N>]
Description
This command specifies parameters of the base mesh used for initial grid generation.
SURF.LY specifies the location of surface. The default is y=0.0 m.
SURF.DY specifies the local grid spacing in y-direction at SURF.LY.
ACTIVE.LY, EPI.LY, SUB.LY specify another three base line location at some critical
region of the device structure to be fabricated.
ACTIVE.DY, EPI.DY, SUB.DY specify the local grid spacing at the ACTIVE.LY, EPI.LY,
SUB.LY. The units are all in microns.
BACK.LY and BACK.DY are the location of bottom in the structure to be fabricated and the
BACK.DY is the local grid spacing in the BACK.LY location.
Examples
The following example assigns the initial base line for the substrate materials. It places the
base line at y=1.0, y=2.0, y=10.0 with a local grid spacing of 0.01 m, 0.5 m, 1.0 m and 10
m separately. The device dimension in the y-direction is specified as y.main=0.
y=max=100. This depth, however, is only for reference. The real depth and width of the
device structure will be assigned in the INITIALIZE command.
BASE.MESH SURF.LY=0.0 SURF.DY=0.01\
ACTIVE.LY=1.0 ACTIVE.DY=0.5 EPI.LY=2.0 EPI.DY=1.0\
SUB.LY=10.0 SUB.DY=10.0 BACK.LY=500 BACK.DY=100
For more examples, see BASE.MESH and INITIALIZE.

BASE.PAR Statements
6.8 BASE.PAR
BASE.PAR runs the base mesh for generating the initial grid.
Syntax
BASE.PAR
[MATERIAL] [GRAD.SPACE] [RATIO.BOX]
Description
MATERIAL one of standard materials or user specified material (see Section 6.2.9
GRAD.SPACE specifies the gradient of the adjacent grid spacing in the y-direction of this
material. Default is 1.5.
RATIO.BOX specifies the approximate aspect ratio of triangle element after base mesh
generation in this material. Default is 2.0.
Examples
The following example generates a good quality base mesh for each related material region.
BASE.PAR OXIDE GRAD.SPACE=5 RATIO.BOX=2
BASE.PAR SILICON GRAD.SPACE=1.5 RATIO.BOX=2
BASE.PAR POLYSILICON GRAD.SPACE=5 RATIO.BOX=2
BASE.PAR OXIDE GRAD.SPACE=5 RATIO.BOX=2
BASE.PAR SILICON GRAD.SPACE=1.5 RATIO.BOX=2
BASE.PAR POLYSILICON GRAD.SPACE=5 RATIO.BOX=2
For more examples, see BASE.MESH.

BOUNDARY Statements
6.9 BOUNDARY
BOUNDARY specifies boundary conditions for the initial material.
Note: For most typical boundary conditions, Athena has defaults that eliminate the need for BOUNDARY
statements. The BOUNDARY statement can be used to modify the treatment of the surfaces for special
purpose simulations.
Syntax
BOUNDARY
[REFLECTING|EXPOSED|BACKSIDE]
[XLO=<c>] [YLO=<c>)] [XHI=<c>] [YHI=<c>]
Description
EXPOSED surfaces correspond to the top of the wafer. Only exposed surface have
deposition or oxidation on top of them. A surface created by etching will also be exposed
unless the ETCH NO.EXPOSE syntax is used.
REFLECTING surfaces correspond to the sides of the device and are also applicable to the
backside as long as defects are not being simulated. All surfaces default to REFLECTING..
BACKSIDE surfaces are physically identical to the reflecting surface with special meaning
only when backside electrode is specified in the ELECTRODE statement.
XLO, YLO, XHI, and YHI set the left, right, top, and bottom bounds of the rectangle being
specified. The value string should be one of the tags specified in one of preceding line
statements.
Examples
The following lines define the top of the mesh to be an exposed surface and the bottom to be
the backside.
BOUNDARY EXPOSED XLO=LEFT XHI=RIGHT YLO=SURF YHI=SURF
BOUNDARY BACKSIDE XLO=LEFT XHI=RIGHT YLO=BACK YHI=BACK
For more examples, see REGION and INITIALIZE.

CLUSTER Statements
6.10 CLUSTER
CLUSTER specifies parameters of {311} cluster model.
Syntax
CLUSTER
[I.IMPURITY] [MATERIAL]
[CLUST.FACT=<n>] [MIN.CLUST=<n>] [MAX.CLUST=<n>]
[TAU.311.0=<n>] [TAU.311.E=<n>]
Description
This command specifies the scaling of {311} clusters during a subsequent IMPLANT step and
the time constant for the dissolution of clusters into free interstitials.
Note: This command will only work if you switch on the {311} cluster model with the METHOD CLUSTER.DAM
command.
I.IMPURITY specifies an impurities to be used for the {311} cluster scaling (see Section
6.2.10 “Standard Impurities” for the list of impurity names that can be used, e.g. I.BORON).
MATERIAL specifies a material in which the scaling takes place (see Section 6.2.9
“Standard and User-Defined Materials” for the list of materials). Default is SILICON.
MIN.CLUST and MAX.CLUST define two values of implanted dopant concentration.
Clusters will be placed between these two dopant concentration levels only. These parameters
are used to control the scaled position of clusters during ion implantation. Typically,
MIN.CLUST is the background doping level. MAX.CLUST is the dopant concentration required
to amorphize the substrate
CLUST.FACT specifies the ratio between the concentration of clustered interstitials and the
implanted dopant concentration.
TAU.311.0=<n> and TAU.311.E=<n> specify the time constant in seconds for the
dissolution of clusters into free interstitials. TAU.311.0 is the pre-exponential linear
coefficient and TAU.311.E is the exponential coefficient used to control temperature
dependence.
Examples
The following command introduces clusters during ion implantation. The clusters will have
an effective interstitial concentration of 1.4 times the concentration of implanted boron. The
clusters will lie in the region where Boron is between 1015 and 1019 cm-3.
METHOD CLUSTER.DAM
CLUSTER I.BORON SILICON MIN.CLUS=1e15 MAX.CLUST=1e19 CLUST.FACT=1.4
IMPLANT DOSE=1e14 ENERGY=50 BORON
The example goes on to define the cluster dissolution time and a short thermal cycle. Results
for each timestep of the diffusion cycle will be stored in files RTA_*.
CLUSTER I.BORON SILICON TAU.311.0=10 TAU.311.E=0.24
DIFFUSE TEMP=1000 TIME=10/60 NITRO DUMP=1 DUMP.PREF=RTA_
For more examples, see METHOD, DISLOC.LOOP, INTERSTITIAL, and VACANCY..

COMMENT Statements
6.11 COMMENT
COMMENT is used to specify character strings for documenting the input deck and Athena
output.
Syntax
COMMENT
#
or
$
Description
The COMMENT statement, # or $ are used to document the input file. You can insert them in the
beginning of any line of the input deck.

CPULOG Statements
6.12 CPULOG
CPULOG instructs Athena to output CPU statistics.
Syntax
CPULOG
[LOG] [CPUFILE = <c>]
Description
The CPULOG statement logs the CPU time used in various internal operations. The CPU time
information appears in the standard output or in the DeckBuild Text Subwindow unless it is
re-directed into CPUFILE.
LOG enables logging of CPU usage when true, and disables CPU logging when false. The
default is true.
CPUFILE specifies a name of the file to which CPU log is written. The default is the
standard output.
Examples
The following example enables Athena to gather CPU statistics and store it in the file
timeusage.out.
CPULOG LOG CPUGFILE=timeusage.out
Note: The accuracy of time statistics depends on the computer and operating system. It is usually around 0.01
sec.

DEPOSIT Statements
6.13 DEPOSIT
DEPOSIT deposits a layer of specified material. DEPOSITION is a synonym for this statement.
Note: Unless the Elite module is used, all deposition steps in Athena are 100% conformal. This means deposition
on all surfaces with a step coverage of 1.0.
Syntax
DEPOSIT
MATERIAL [NAME.RESIST=<c>] THICKNESS=<n>
[SI_TO_POLY] [TEMPERATURE=<n>]
[DIVISIONS=<n>] [DY=<n>][YDY=<n>] [MIN.DY=<n>] [MIN.SPACE=<n>]
[C.IMPURITIES=<n>] [F.IMPURITIES=<n.] [C.INTERST=<n>]
[F.INTERST=<n>]
[C.VACANCY=<n>] [F.VACANCY=<n>] [X.COMP=<n>] [FX.COMP=<n>]
[C.FRACTION=<n>]
[F.FRACTION=<n>] [Y.COMP=<n>] [FY.COMP=<n>] [GR.SIZE=<n>]
[F.GR.SIZE=<n>]
[MACHINE=<c>] [TIME=<n>] [HOURS|MINUTES|SECONDS]
[N.PARTICLE=<n>] [OUTFILE=<c>] [SUBSTEPS=<n>][VOID]
Description
This statement is used to simulate deposition of specified material on the exposed surface of
the current structure.
MATERIAL specifies the material to be deposited (see Section 6.2.9 “Standard and User-
Defined Materials” for the list of materials).
NAME.RESIST specifies the type of photoresist to be deposited.
THICKNESS specifies the deposited layer thickness in microns.
SI_TO_POLY specifies that crystalline silicon will be deposited only over crystalline silicon,
while polysilicon will be deposited elsewhere.
TEMPERATURE specifies deposition temperature used by STRESS.HIST model. The
temperature is also used for surface diffusion simulation during Elite deposition.
Grid Control Parameters
DIVISIONS specifies the number of vertical grid spacings in the layer. In some cases, it is
important to control the number of grid points in a conformally deposited layer since this also
controls the accuracy of subsequent processes. SPACES is an alias for DIVISIONS. The default
is 1.
Note: The default for DIVISIONS is 1. This typically needs to be increased for all deposition steps. If
DIVISIONS is set too low to maintain grid integrity in a non-planar deposition, Athena will attempt to
recover by increasing DIVISIONS automatically. Athena will echo the number of DIVISIONS finally
used to the run-time output.
DY specifies the nominal spacing in the layer. Units are microns.

DEPOSIT Statements
YDY specifies the depth where the nominal spacing will be applied. YDY is calculated
relative to the top of the newly deposited layer. Units are microns.
MIN.DY specifies the minimum spacing in microns allowed between grid lines in the y
direction in the new material. The default is 0.001 microns (10 Angstroms).
MIN.SPACE specifies a minimum spacing between points on the surface of each sub-layer.
Increasing this parameter will reduce the number of points on arced deposited surfaces. Units
are microns. ARC.SPACE is a synonym for this parameter.
Parameters Specific to Depositing Doped Layers
C.IMPURITIES specify the concentration of the impurity in the deposited layer in cm-3.
You can specify more than one of these parameters to define materials doped with multiple
impurities.
Note: There are several additional methods to specify impurity concentration in deposited material. These
methods are introduced to provide TSUPREM4 compatibility and described in Section C.8 “Changes in the
DEPOSIT statement”.
F.IMPURITIES can only be specified together with the corresponding C.IMPURITY (e.g.,
F.BORON and C.BORON). This parameter generates the linearly graded concentration of the
specified impurity in the deposited layer, where C.IMPURITY specifies concentration at the
bottom of the layer and F.IMPURITY specifies concentration at the top of the layer. Units are
cm-3.
C.INTERST specifies the concentration of interstitials in deposited layer. Units are cm-3.
F.INTERST can only be specified together with C.INTERST. This parameter generates the
linearly graded interstitial concentration in the deposited layer, where C.INTERST specifies
concentration at the bottom of the layer and F.INTERST specifies concentration at the top of
the layer.
C.VACANCY specifies the concentration of vacancies in deposited layer. Units are cm-3.
F.VACANCY can only be specified together with C.VACANCY. This parameter generates the
linearly graded vacancy concentration in the deposited layer, where C.VACANCY specifies
concentration at the bottom of the layer and F.VACANCY specifies concentration at the top of
the layer. Units are cm-3.
X.COMP specifies the composition fraction of the cation listed first in ternary or quaternary
compound semiconductor to be deposited. For example, Al is usually first cation of AlGaAs.
The full list of compound semiconductors that can be specified in Athena and the rules of
their naming can be found in Appendix B of Atlas User’s Manual. For backward
compatibility, C.FRACTION is the alias for X.COMP.
FX.COMP can only be specified together with X.COMP. This parameter allows you to
generate the compound semiconductor layer with linearly graded composition fraction of the
cation. If both X.COMP and FX.COMP are specified, the cation composition fraction will be
X.COMP at the bottom of the deposited layer and FX.COMP at the top of the new layer. For
backward compatibility, F.FRACTION is the alias for FX.COMP.
Y.COMP specifies the composition fraction of the anion listed first in quaternary compound
semiconductor to be deposited. For example, As is first anion in InGaAsP. The full list of

DEPOSIT Statements
compound semiconductors that can be specified in Athena and the rules of their naming can
be found in Appendix B of Atlas User’s Manual.
FY.COMP can only be specified together with Y.COMP. This parameter allows you to
generate the compound semiconductor layer with linearly graded composition fraction of the
anion. If both Y.COMP and FY.COMP are specified, the anion composition fraction will be
Y.COMP at the bottom of the deposited layer and FY.COMP at the top of the new layer.
GR.SIZE specifies grain size in deposited polysilicon layer. This parameter is recognized
only when POLY.DIFF model is specified in the METHOD statement. Units are microns.
F.GR.SIZE can only be specified together with GR.SIZE. This parameter deposits
polysilicon layer with grains linearly graded with their sizes, where GR.SIZE specifies grain
size at the bottom of the layer and F.GRAIN.SIZE specifies grain size at the top of the layer.
Units are microns.
Parameters Specific to Elite Depositions
MACHINE specifies the name of the machine to be run for Elite deposits. The machine
name must be specified in a previous RATE.ETCH statement.
TIME sets the time in specified units the etch machine will be running.
HOURS, MINUTES, and SECONDS specifies the units of the TIME parameter. Default is
MINUTES.
N.PARTICLE specifies the number of particle trajectories to calculate for the Monte Carlo
deposit model.
OUTFILE specifies the name of the file to be written with Monte Carlo particle positions.
SUBSTEPS specifies the number of timesteps made for each division of the deposit in the
Elite module.
VOID specifies that the voids formed during deposition are to remain unfilled with deposit
material.
Conformal Deposition Example
The following statement deposits a conformal layer of silicon dioxide, 1000 Angstroms thick,
on the surface of the simulation structure. It will contain 4 vertical grid points.
Example Depositing Doped User-defined Material
The following deposits a layer of a user defined material BPSG doped with boron and
phosphorus.
DEPOSIT MATERIAL=BPSG THICKNESS=0.1 DIV=6 C.BORON=1e20 C.PHOS=1e20
Grid Control Example
The following statement deposits a conformal layer of silicon nitride with a thickness of
0.3m. The grid spacing at the bottom of the layer is 0.01m and the layer will include 10
vertical sublayers.
DEPOSIT NITRIDE THICK=0.3 DY=0.1 YDY=0.3 DIVISIONS=10

DEPOSIT Statements
Elite Machine Deposition Example

The following statements define a machine named MOCVD and use it to deposit tungsten with a
thickness of 0.1m on planar areas and step coverage of 0.75.
RATE.DEPO MACHINE=MOCVD DEP.RATE=.1 u.m STEP.COV=.75 TUNGSTEN
DEPOSIT MACHINE=MOCVD TIME=1 MINUTE
For more examples, see RATE.DEPO.

DEVELOP Statements
6.14 DEVELOP
DEVELOP runs the development module in Optolith.
Syntax
DEVELOP
[MACK|DILL|TREFONAS|HIRAI|KIM|EIB]
[TIME=<n>] [STEPS=<n>] [SUBSTEPS=<n>]
[DUMP=<n>] [DUMP.PREFIX=<c>]
Description
This command runs the development module and enables the use of the option to select a
development model.
MACK, DILL, TREFONAS, HIRAI, KIM and EIB specify the development model to be
used.
TIME, STEPS, and SUBSTEPS are related parameters that control the string algorithm in
development. TIME is the total development time in seconds. STEPS gives the number of
times ETCH is to be performed. SUBSTEPS controls string movement. Each substep or string
movement has a time duration of TIME/STEP*SUBSTEPS.
DUMP determines whether a structure is saved after each step of the development is
completed.
DUMP.PREFIX specifies the prefix name for the structure file to be saved. The number of
steps will be equal to the number of output files. The files are readable with the STRUCTURE
statement or can be displayed using TonyPlot. The names of the files will be of the form
DUMP.PREFIX***.***.str, where ** is the current development time.
Examples
The following example dumps out five structure files to show the evolution of development
using the KIM development model.
DEVELOP KIM DUMP=1 TIME=60 STEPS=5
For more examples, see RATE.DEVELOP.

DIFFUSE Statements
6.15 DIFFUSE
DIFFUSE runs a time temperature step on the wafer and calculates oxidation, silicidation and
diffusion of impurities. DIFFUSION is a synonym for this statement.
Syntax
DIFFUSE
TIME=<n> [HOURS|MINUTES|SECONDS]
TEMPERATURE=<n> [T.FINAL=<n>|T.RATE=<n>]
[DRYO2|WETO2|NITROGEN|INERT] [HCL.PC=<n>] [PRESSURE=<n>]
[F.02=<n>|F.H2=<n>|F.H20=<n>|F.N2=<n>|F.HCL=<n>]
[C.IMPURITIES=<n>]
[DUMP][DUMP.PREFIX=<c>] [TSAVE=<n>] [TSAVE.MULT=<n>]
[B.MOD=<c>] [p.MOD=<c>] [AS.MOD=<c>] [IC.MOD=<c>] [VI.MOD=<c>]
[NO.DIFF] [REFLOW]
Description
This command specifies diffusion and/or oxidation/silicidation steps. Any impurities present
in the wafer are diffused if they have non-zero diffusivities. The oxidation and diffusion
control parameters are contained in the associated METHOD, OXIDE, and SILICIDE statements.
Default coefficients are in the ATHENAMOD file available from the DeckBuild Commands
menu under Models. To change model coefficients, refer to the appropriate IMPURITY
statement for information.
Parameters to Define the Diffusion Step
TIME specifies the amount of time for the diffusion step in specified units. If TIME is set to
zero, only stress simulation will be performed when STRESS.HIST is specified .
HOURS, MINUTES, and SECONDS specify the units of the TIME parameter. Default is
MINUTES.
TEMPERATURE specifies the ambient temperature in oC. This temperature should fall
within the range between 700 and 1200oC. Outside of this range, the diffusion coefficients
may be inaccurate and numerical difficulties may occur during simulation. For ramped
thermal step, a synonym T.START can be used.
T.FINAL specifies the final temperature for ramped thermal steps. Synonym is T.STOP.
T.RATE specifies the ramp rate in oC/minute for ramped thermal steps.
Parameters to Define the Diffusion Ambient
DRYO2, WETO2, INERY and NITROGEN specify the type of ambient during the
diffusion step. DRYO2 specifies that ambient is dry oxygen, WETO2 specifies that ambient is
wet oxygen, NITROGEN specifies that ambient is inert. INERT is a synonym for NITROGEN.
HCL.PC specifies the percentage of HCl in the oxidant gas stream.
PRESSURE specifies the partial pressure of the active species in atmospheres. Units are
atmospheres. The default is 1.

DIFFUSE Statements
F.02, F.H2, F.H20, F.N2, and F.HCL specifies the relative flow rate of the components of
oxygen, hydrogen, water, nitrogen, and HCl in the ambient. If these parameters are used, the
DRYO2, WETO2, NITROGEN set or HCL.PC should not be specified.
C.IMPURITIES specifies concentration of the impurities in the ambient gas. (see Section
6.2.10 “Standard Impurities” for the list of impurity names that can be used, e.g. I.BORON).
Units are atoms/cm3. You can define multiple impurity parameters for ambients with multiple
impurities. You can only use boron, phosphorus and arsenic if you specify the advanced
diffusion model (PLS) in the METHOD statement.
Note: There are several additional methods to specify impurity concentration in ambient gas. These methods are
introduced to provide TSUPREM4 compatibility and described in Section C.9 “Changes in the DIFFUSE
statement”.
Parameters Related to File Output

DUMP and DUMP.PREFIX specify that a structure file be output at every DUMPth time step.
The files are readable with the STRUCTURE statement or can be displayed using TonyPlot. The
names will be of the form DUMP.PREFIX<time>.str, where <time> is the current total time
of the simulation in minutes.
TSAVE and TSAVE.MULT specify that intermediate structure files be output when the
advanced PLS diffusion model is used. The structure files named DUMP.PREFIX<time>.str
will be output at time = TSAVE * TSAVE.MULTk, k=0,1,2,... where time is in seconds. The
default value for the parameter DUMP.PREFIX is “at”. The parameter TSAVE.MULT should be
greater than 1.0.
Parameters Related to the Model Files for Advanced Diffusion Models
B.MOD, P.MOD, AS.MOD, IC.MOD, and VI.MOD specify direct paths to boron.mod,
phosphorus.mod, arsenic.mod, i.mod, and defect.mod files correspondingly. By
default, these files are in $SILVACO/lib/athena/<version_number>/common/pls
directory. You can modify your own *.mod files inside directories specified by these
parameters.
Miscellaneous Parameters
NO.DIFF specifies that impurity diffusion be neglected during the calculation. This can be
used to observe oxidation or silicidation geometry without unnecessary timesteps related to
impurity diffusion
REFLOW specifies that a surface tension based reflow of the material is to be performed
during the diffusion step.
Predeposition Example
The following statement specifies a 1000°, 30 minute boron pre-deposition.
DIFFUSE TIME=30 TEMP=1000 C.BORON=1.0E20
Oxidation Example
The following statement instructs the simulator to grow oxide for 30 minutes in a dry oxygen
ambient.
DIFFUSE TIME=30 TEMP=1000 DRYO2

DIFFUSE Statements
Gas Flow Example

The following command performs diffusion with a mixed ambient with relative components
of oxygen, hydrogen, and HC1 of 10, 10, and 0.1 respectively.
DIFFUSE TIME=10 TEMP=1000 F.02=10 F.H2=10 F.HCl=.1
Hydrogen and Oxygen are combined in a ratio 2:1 to form the ambient WETO2. Any excess
hydrogen is considered inert. Any excess oxygen is considered as the ambient DRYO2. Since
the total pressure of the gas flow is defined (or defaults to one atmosphere) the partial
pressure of WETO2 will be reduced if any excess hydrogen or oxygen is present.
File Output Example
The following commands perform diffusion in dry oxygen ambient for 30 minutes at 1000
o
C. After every second timestep a structure file is written with a name prefix TEST. Following
the diffusion, the TonyPlot statement plots each timestep output file in a manner suitable for
creating a diffusion movie. A SYSTEM command is used to execute a UNIX command prior
to the diffusion step to remove all TEST*.str files from previous runs.
SYSTEM rm -rf TEST*.str
DIFFUSE TIME=30 TEMP=1000 DRYO2 DUMP=2 DUMP.PREFIX=TEST
TONYPLOT -st TEST*.str
Advanced Diffusion Model Example
The following command performs the boron pre-deposition at 950°C for 1 hour. The boron
concentration in the ambient gas is 1020cm-3. As a result, the output files predep1.str,
predep10.str, predep100.str, and predep1000.str will be saved.
METHOD PLS
DIFFUSE TIME=1 HOUR TEMP=950 C.BORON=1E20 TSAVE=1 TSAVE.MULT=10
DUMP.PREFIX=predep
For more examples, see IMPURITY, INTERSTITAL, MATERIAL, METHOD, OXIDE, TRAP, and
VACANCY.

DISLOC.LOOP Statements
6.16 DISLOC.LOOP
DISLOC.LOOP defines the scaling parameters and position of dislocation loops.
Syntax
DISLOC.LOOP
MATERIAL I.IMPURITY MIN.LOOP.CO =<n> MAX.LOOP.CO=<n>
Description
This command specifies the scaling of dislocation loops during a subsequent IMPLANT step.
Dislocation loops are used as interstitial sinks whose recombination rate can be determined
with the INTERSTITIAL DAMALPHA=<n> command.
Note: This command will only work if you switch on the dislocation loop model with the METHOD
I.LOOP.SINK command.
MATERIAL specifies material for which dislocation loops parameters are set (see Section
6.2.9 “Standard and User-Defined Materials” for the list of materials). Default is SILICON.
I.IMPURITY specifies an impurities to be used for the dislocation loop scaling (see Section
6.2.10 “Standard Impurities” for the list of impurity names which can be used, e.g. I.BORON).
MIN.LOOP.CO and MAX.LOOP.CO define the upper and lower bounds of the dopant
concentrations where the loops are placed.
Dislocation Loop Generation Example
The following example switches on the loop model and then places loops in the position
where indium concentrations lie between 1e16 and 1e15 cm-3.
METHOD I.LOOP.SINK
DISLOC.LOOP MIN.LOOP.CO=1e15 MAX.LOOP.CO=1e16 I.INDIUM SILICON
IMPLANT INDIUM DOSE=1e15 ENERGY=45
For more examples, see METHOD, CLUSTER, INTERSTITIAL, VACANCY, DIFFUSE, and
IMPLANT.

ELECTRODE Statements
6.17 ELECTRODE
ELECTRODE defines electrodes and names for Atlas or other device simulation.
Syntax
ELECTRODE
NAME=<c> [X=<n> | Y=<n> | BACKSIDE | LEFT | RIGHT]
Description
This statement defines a whole material region as an electrode.
NAME gives a name to the electrode that can be plotted or referenced in TonyPlot or Atlas.
X specifies the horizontal location or x-coordinate of the region, which will be defined as an
electrode.
Y specifies the vertical location or y-coordinate of the electrode being defined. If no value of
Y is specified, the top of the structure is assumed.
BACKSIDE specifies that a flat (zero height) electrode will be placed on the bottom of the
simulation structure. This is the one exception to whole regions being defined as electrodes. If
a metal region is present on the bottom of the structure, this parameter will not be used and
the XY coordinates used instead. BOTTOM is a synonym for this parameter.
LEFT specifies that the top left region of the structure will be defined as an electrode.
RIGHT specifies that the top right region of the structure will be defined as an electrode.
Note: The ELECTRODE statement recognizes the regions made of polysilicon, standard metals (see.Section
6.2.9 “Standard and User-Defined Materials” for the list of standard metals) or user-defined materials with
the following standard names: Gold, Silver, AlSi, Palladium, Molybdenum, Lead, Iron, Tantalum, AlSiTi,
AlSiCu, TiW, Copper, Tin, Nickel, NiSix, TaSix, PaSix, MoSix, ZrSix, AlSix, Conductor, Contact.
Note: Atlas contains syntax that makes use of the common electrical names for highly preferred terminals. These
are anode, cathode, emitter, base, collector, gate, source, drain, bulk and substrate.
Metal Region Electrode Definition Example

The following gives the name source to the metal or polysilicon region at location x=1
micron on the top of the current structure.
ELECTRODE X=1.0 NAME=SOURCE
Substrate Definition Example
The following gives the name well to a flat electrode along the bottom edge of the current
structure There is no metal required at this location.
ELECTRODE BACKSIDE NAME=WELL
For more examples, see STRUCTURE.

EPITAXY Statements
6.18 EPITAXY
EPITAXY specifies an epitaxial deposition process step.
Syntax
EPITAXY
TIME=<n> [HOURS|MINUTES|SECONDS]
TEMPERATURE=<n> [T.FINAL=<n> | T.RATE=<n>]
[THICKNESS=<n> | GROWTH.RATE=<n>]
[C.IMPURITIES=<n>] [F.IMPURITIES=<n.] [C.INTERST=<n>]
[F.INTERST=<n>]
[C.VACANCY=<n>] [F.VACANCY=<n>]
[DIVISIONS=<n>][DY=<n>][MIN.DY=<n>][YDY=<n>][SI_TO_POLY]
Description
This statement simulates the epitaxial deposition of silicon. This model is limited to silicon on
silicon applications and should not be used when other materials are present. The model is
inherently 1D and isn’t suitable for selective epitaxial deposition processes.
Parameters to Define the Epitaxial Step
TIME specifies the amount of time for the epitaxial step in specified units.
HOURS, MINUTES, and SECONDS specify the units of the TIME parameter. Default is
MINUTES.
TEMPERATURE specifies the ambient temperature, in oC. This temperature should fall
within the range between 700 and 1200oC. Outside of this range, the diffusion coefficients
may be inaccurate and numerical difficulties may occur during simulation. For ramped
thermal step, a synonym T.START can be used.
T.FINAL specifies the final temperature for ramped thermal steps. Synonym is T.STOP.
T.RATE specifies the ramp rate in oC/minute for ramped thermal steps.
THICKNESS specifies thickness of epitaxially grown layer. Units are microns.
GROWTH.RATE specifies epitaxial growth rate. It is applicable only when THICKNESS is
not specified. Units are m/minute.
Doping Related Parameters

C.IMPURITIES specify the concentration of the impurity in the epitaxially grown layer in
cm-3. You can specify more than one of these parameters to define materials doped with
multiple impurities.
F.IMPURITIES can be specified only together with the corresponding C.IMPURITY (e.g.
F.BORON and C.BORON). This parameter generates the linearly graded concentration of the
specified impurity in the epitaxially grown layer, where C.IMPURITY specifies concentration
at the bottom of the layer and F.IMPURITY specifies concentration at the top of the layer.
C.INTERST specifies the concentration of interstitials in the epitaxially grown layer in cm-3.

EPITAXY Statements
F.INTERST can be specified only together with C.INTERST. This parameter generates the
linearly graded interstitial concentration in the epitaxially grown layer, where C.INTERST
specifies concentration at the bottom of the layer and F.INTERST specifies concentration at
the top of the layer.
C.VACANCY specifies the concentration of vacancies in the epitaxially grown layer in cm-3.
F.VACANCY can be specified only together with C.VACANCY. This parameter generates the
linearly graded vacancy concentration in the deposited layer, where C.VACANCY specifies
concentration at the bottom of the layer and F.VACANCY specifies concentration at the top of
the layer.
Gridding Parameters
DIVISIONS controls the number of vertical grid points in the resulting epitaxial layer. This is
an optional parameter since it will be generated automatically by default and is related to the
surface grid spacing of the original simulation structure before epitaxial process. SPACES is an
alias for this parameter. The default is 10.
DY specifies the nominal spacing in microns in the epitaxial layer.
YDY specifies the depth where the nominal spacing will be applied. YDY is calculated
relative to the top of the newly grown epitaxial layer.
MIN.DY specifies the minimum spacing allowed between grid lines in the y direction in the
new material. The default is 0.001 microns (10 Angstroms).
SI_TO_POLY specifies that the crystalline silicon layer will be grown only over Silicon,
while Polysilicon will be grown elsewhere.
Deposition Rate Example
The following statement will simulate the growth of boron doped silicon on top of silicon at a
rate of 0.5 m per minute. The deposit thickness is time x rate=5 m.
EPITAXY TIME=10 TEMP=1150 C.BORON=5E14 GROWTH.RATE=0.5
Time and Temperature Example
The following statement will deposit 6m of epitaxial silicon on top of silicon over 10
minutes. Phosphorus is out-diffused during the processing. The number of vertical grid points
in the completed epitaxial layer is set with the DIVISIONS parameter. The syntax is similar to
the DEPOSIT statement.
EPITAXY THICK=6 TIME=10 TEMP=1180 C.PHOS=1.5E14 DIVISIONS=20
Non-uniform Grid Control Example
The following statement performs epitaxial process with a non-uniform vertical grid spacing.
The vertical grid spacing will be 0.5 m at a distance of 5 m below the final surface. The
epitaxial layer will be subdivided into 40 sublayers:
EPITAXY THICK=10 TIME=30 TEMP=1100 DY=.5 YDY=5.0 DIVISIONS=40
For more examples, see DEPOSIT and DIFFUSE.

ETCH Statements
6.19 ETCH
ETCH simulates an etch process.
Syntax
ETCH
[MATERIAL] [NAME.RESIST]
[ALL|DRY] [THICKNESS=<n>] [ANGLE=<n>] [UNDERCUT=<n>]
[LEFT|RIGHT|ABOVE|BELOW] [P1.X=<n>] [P1.Y=<n>] [P2.X=<n>]
[P2.Y=<n>]
[START|CONTINUE|DONE] [X=<n>][Y=<n>]
[INFILE=<c>] [TOP.LAYER] [NOEXPOSE] [DIRECT.ANGLE=<n>]
[MACHINE=<c>] [TIME=<n>] [HOURS|MINUTES|SECONDS]
[DT.FACT=<n>] [DT.MAX=<n>] [DX.MULT=<n>]
[MC.REDEPO] [MC.SMOOTH = <n>] [MC.DT.FACT = <n>] {MC.MODFNAME =
<c>]
Description
Athena provides two different etch simulation methods. The first is geometrical etching
available within any Athena module. The second is physical etching available only in Elite.
Parameters used for Geometrical Etching
MATERIAL specify the material to be etched (see Section 6.2.9 “Standard and User-Defined
Materials” for the list of materials). If a material is specified, only that material is etched even
if other materials lie within the etch region. If no material is specified, all materials in the etch
region are removed.
NAME.RESIST specify the type of photoresist to be etched.
ALL specifies that all of the specified materials are removed.
DRY indicates that the resulting surface will replicate the exposed surface and will simply be
lowered by a fixed depth of THICKNESS microns below the exposed surface. If ANGLE or
UNDERCUT or both is specified, the shape of DRY etched region is modified accordingly.
TRAPEZOI is a synonym for this parameter.
THICKNESS specifies the thickness to be etched for the dry etch type. Units are microns.
ANGLE specifies sidewall slopes in degrees (90° corresponding to vertical slope is the
default).
UNDERCUT specifies the distance in microns that ETCH extends under a mask when dry
etch is performed (the default is 0).
LEFT, RIGHT, ABOVE, and BELOW provide a quick means of etching with a trapezoidal
cross section. The etch region will be to the specified side (left/right/above/below) of the line
specified by the coordinates given in P1.X, P1.Y and P2.X, P2.Y.
P1.X, P1.Y, P2.X, and P2.Y allow you to specify a line for left/right/above/below
etching. The P1 parameters are always required if left/right/above/below are used. The
P2 parameters are required when the etch angle is non-vertical. Units are microns.

ETCH Statements
START, CONTINUE, and DONE specify an arbitrarily complex region to be etched. You
can combine several lines to specify the several points that make up the region. See the
examples.
X and Y specify a point in the start/continue/done mode of etch region specification.
Units are microns.
INFILE specifies that the etch profile will be taken from the filename specified by the
INFILE parameter. The specified file must have the following format:
X1 Y1
X2 Y2
X3 Y3
...
Xn Yn
This will etch the region enclosed by the boundary coordinates within the file. You can define
any number of coordinates within the file. This command is often useful for inputting data
from digitized experimental profiles or external programs. The closing line is automatically
drawn from the final coordinate point to the initial point.
TOP.LAYER specifies that only the top layer of the etched material should be etched.
NOEXPOSE specifies that the new surface is not exposed for subsequent oxidation or
deposition after geometrical etch. Use this parameter to remove a part of the structure from
the bottom or side of simulation.
DIRECT.ANGLE specifies angle (in degrees) of a directional etch. This angle is measured
from the vertical. Default is 0 degrees.
Parameters used only with physical etching in the Elite module
MACHINE specifies the name of the etch machine that is to be run.
TIME specifies the time the etch machine is to be run.
HOURS, MINUTES, and SECONDS specify the units of the TIME parameter.
Parameters used only with RIE, WET.ETCH, and PLASMA models
DT.FACT is used with Elite type etch calculations. By default, the movement of a string node
is limited to less than or equal to one quarter of the median segment length. This is a good
compromise between simulation speed and the danger of loop formation. The optimization
factor DT.FACT must not exceed 0.5. You can, however, decrease it if necessary for more
accuracy.
DT.MAX is used to limit timesteps size. By default, the upper limit for the maximum
timestep is one tenth of the total etch time specified. This is a good compromise between
calculation accuracy and calculation time. But sometimes, it is useful to adapt this value to
the specific simulation problem. Allowing the time steps to become greater gives a higher
simulation speed but the accuracy may suffer. For smaller time steps, the simulation speed
will decrease but the accuracy may be greater.
DX.MULT is the accuracy multiplier for Elite etches. The discretization size used for the etch
calculation will be multiplied by DX.MULT. For improved accuracy at the cost of extra
simulation time, decrease the value of DX.MULT.

ETCH Statements
Parameters used only with MC.PLASMA model

MC.REDEPO specifies that redeposition of polymer should be simulated. Default is true.
MC.SMOOTH specifies level of smoothing of the surface.
MC.DT.FACT specifies time step control for Monte Carlo etching and redeposition.
MC.MODFNAME specifies name of the C-Interpreter file with user-defined Monte Carlo
etching and redeposition models.
Simple Geometrical Etch Example
The following command etches all the nitride to the left of a vertical line located at x=0.5:
ETCH NITRIDE LEFT P1.X=0.5
Arbitrary Geometrical Shape Etch Example
The following set of commands etch the oxide in the square defined at (0,0), (1,0), (1,1), and
(0,1):
ETCH OXIDE START X=0.0 Y=0.0
ETCH CONTINUE X=1.0 Y=0.0
ETCH CONTINUE X=1.0 Y=1.0
ETCH DONE X=0.0 Y=1.0
Be careful when using this style of syntax that the list of coordinates forms a regular polygon.
The closing line from the last coordinate pair to the initial point is automatically added.
Anisotropic Geometrical Etch Example
The following command finds the exposed surface and lowers it straight down 0.1 microns.
This line will be the new surface.
ETCH DRY THICK=0.1
Physical Etch Example
The following sequence defines an etch machine named PLASMA1 that performs reactive ion
etching of silicon. The machine is applied to etch the current structure for 10 minutes.
RATE.ETCH MACHINE=PLASMA1 SILICON U.M RIE ISOTROPIC=0.1 DIRECT=0.9
ETCH MACHINE=PLASMA1 TIME=10 MINUTES
Note: The program can be sensitive to grid placement. It often helps to prepare the initial grid by having a vertical
grid line exactly at the etch coordinate for geometric etches.
For example, see RATE.ETCH.

EXPOSE Statements
6.20 EXPOSE
EXPOSE runs the exposure module of Optolith.
Syntax
EXPOSE
[INFILE=<c>] [PERPENDICUL|PARALLEL] [X.CROSS|Z.CROSS]
[CROSS.VALUE=<n>] [DOSE=<n>] [X.ORIGIN=<n>
[FLATNESS=<n>] [NUM.REFL=<n>]
[FRONT.REFL=<n>] [BACK.REFL=<n>] [ALL.MATS=<n>]
[MULT.EXPOSE] [POWER.MIN=<n>]
Description
This command defines the parameters associated with and performs two dimensional
exposure.
INFILE is the name of an input file that contains a user aerial image cross section data file.
This file has the form:
<wavelength in microns>
<number of data pairs>
<x location in structure> <relative image intensity>
PERPENDICUL and PARALLEL specify TE mode or TM mode respectively.
PERPENDICUL is the default.
X.CROSS and Z.CROSS specify that the cross section is parallel to the x-axis (z=constant)
and parallel to the z-axis (x=constant), respectively. X.CROSS is the default.
CROSS.VALUE specifies the x or z coordinates of the cross section of the aerial image. The
default will be centered in the image window. Units are microns.
DOSE specifies the exposure dose in mJ/cm2.
X.ORIGIN locates the beam relative to the structure. This allows the aerial image to be
shifted if necessary. Units are microns. The default is 0.0.
FLATNESS specifies the accuracy of the change in surface topography in degrees. A value
of zero specifies that all grid points will be calculated. The default value is 0.25. In any case,
maintain the limits 0 FLATNESS  1.
NUM.REFL specifies the number of reflections to be considered.
FRONT.REFL specifies that front surface reflection should be considered in the calculation.
The default is no front reflection.
BACK.REFL specifies the back surface reflection. The default is no back reflection.
ALL.MATS specifies that intensity be displayed in all materials. The default is photoresist
only.

EXPOSE Statements
MULT.EXPOSE is used to make multiple exposures. MULT.EXPOSE is specified on the

second EXPOSE command for addition of exposures. If MULT.EXPOSE is not specified,
previous exposures will be erased.
POWER.MIN sets the minimum power accounted for in multiple reflections. POWER.MIN is
used in a multiplicative format. In other words, if power attenuation due to 10 reflections is
less than POWER.MIN it will not be counted for calculation.
Examples
The following statement loads a cross section of an aerial image that you can input. It then
runs the exposure module. The number of reflections increases calculation time when it is set
to a value greater than one.
EXPOSE INFILE=CROSS.SECT NUM.REFL=3
The following command runs the exposure module for the Z.CROSS section of a two
dimensional aerial image that has been previously generated. The x value of the cross section
is 0.1.
EXPOSE Z.CROSS CROSS.VAL=0.1
Note: The DEFOCUS parameter on the IMAGE statement must be used to do defocus exposure calculations. The
image command must be used in conjunction with the EXPOSE command for a defocussed bulk image.
For more examples, see INITIALIZE and IMAGE.

EXTRACT Statements
6.21 EXTRACT
The EXTRACT command is used to analyze the current structure or a previously saved file. It
can extract important parameters such as material thickness, junction depth and peak doping
levels. It also includes electrical extractions such as sheet resistance, threshold voltage and
CV curves.
Note: The EXTRACT statement is supported under DeckBuild and is fully documented in the DeckBuilD
User’s Manual.

FOREACH Statements
6.22 FOREACH
FOREACH specifies the command looping facility.
Syntax
FOREACH(NAME)(LIST)
(COMMANDS)
END
Description
This command is used to specify input loops. FOR is equivalent to FOREACH. When the loop
executes, NAME will consecutively take on each value in LIST and exit the loop after assuming
the last value. COMMANDS will be executed once for each value in LIST. NAME is set to a value
in LIST using the shell define function.
LIST is a set of strings separated by commas or spaces. The values in LIST can be delimited
by either commas or spaces. LIST can also take the following numerical operator form:
START TO END STEP VAL
where START is a numerical start value, END is the last value, and VAL is the amount to
increment at each iteration.
Examples
The following statement will increment val from 1.0 to 10.0 in steps of 0.5. This loop will be
executed 19 times.
FOREACH VAL ( 1.0 TO 10.0 STEP 0.5 )
ECHO VAL
END
Note: Command line continuation using the backslash character “\” indicator is not supported in the FOREACH
statement.
Note: FOREACH, ECHO, and END statements are supported only in the batch mode. DeckBuild provides much
more powerful looping capabilities, which is recommended.

GO Statements
6.23 GO
GO starts the simulator. Each Athena input file should begin with a GO statement.
Note: The GO command is executed by DeckBuild and documented in the DeckBuild User’s Manual.
Examples
Two useful features of the GO command are shown here.
This command initializes Athena with a specified version number
go athena simflags="-V 4.3.0.R"
This command initializes Athena with a model file <install>/lib/athena/<version>/
common/athenamod.97a.
go athena simflags="-modfile 97a"
Note: If DeckBuild encounters a GO statement and there is no change in the version or model file, Athena will
continue running.

HELP Statements
6.24 HELP
HELP prints summary of statement names and parameters syntax.
HELP [<command>]
or
? [<command>]
Description
HELP lists the parameters of the specified statement and provides a short description of each.
If there is no statement name given, HELP will show an introductory help message and will
list all statements.
Examples
The following will print a list of valid Athena commands to the standard output.
HELP
The following will print a description of the DIFFUSE command and its parameters.
HELP DIFFUSE

ILLUM.FILTER Statements
6.25 ILLUM.FILTER
ILLUM.FILTER specifies the illumination source shape and illumination source filtering in
Optolith.
Syntax
ILLUM.FILTER
[CIRCLE|SQUARE|GAUSSIAN|ANTIGAUSS|SHRINC]
[GAMMA=<n>] [RADIUS=<n>] [ANGLE=<n>] [SIGMA=<n>]
[IN.RADIUS=<n>] [OUT.RADIUS=<n>] [PHASE=<n>]
[TRANSMIT=<n>] [CLEAR.FIL]
Description
This statement specifies the illumination source options as well as illumination source
filtering.
CIRCLE, SQUARE, GAUSSIAN, ANTIGAUSS, and SHRINC define or change the shape
of the exit pupil of the illumination system. SHRINC can be used to define the illumination
system only, not annular filters.
GAMMA defines or changes the GAMMA value for GAUSSIAN or ANTIGAUSS source
transmittance. GAMMA is a parameter that defines the truncation of the GAUSSIAN by the pupil.
In the limit of GAMMA0, the source will be uniform.
RADIUS specifies the radius of a single source if you define the SHRINC illuminator
concept. This parameter must be entered in fractions of unity.
ANGLE specifies the angular location for the SHRINC illuminator.
SIGMA defines or changes the filling factor for the combination of the illumination and
projection systems. The value of SIGMA is expected to vary, but it will not be reset. Also,
specifies the radius of a single source if you specify the SHRINC illuminator concept. This
parameter must be entered in fractions of unity (assuming a unit pupil radius).
IN.RADIUS and OUT.RADIUS define or change the intensity transmittance and phase
transmittance of an annular zone inside the exit pupil of the illumination system. This
qualifier is used to simulate spatial filtering techniques. IN.RADIUS and OUT.RADIUS are
used to define an annular zone in the exit pupil having the pupil transmittance equal to
TRANSMIT and producing the phase angle equal to PHASE. Radius values are specified in
fractions of unity, and phase is specified in degrees. Note that the annular zones should not
overlap. The outer radius of an inner zone must be smaller than the inner radius of an outer
zone.
PHASE specifies the phase shift in degrees produced by the illumination source filter (-180o
PHASE  180o).
TRANSMIT specifies the intensity transmittance produced by the illumination filter (0 
TRANSMIT  1).
CLEAR.FIL resets the illumination source filter list.

ILLUM.FILTER Statements
Example
The following example defines a SHRINC illumination source, where the quadruple circular
illumination sources are located at 45 to the x-axis with the center at a radius of 0.2 from the
origin and a circle radius of SIGMA=0.2.
ILLUM.FILTER SHRINC RADIUS=.2 SIGMA=.2 ANGLE=45

ILLUMINATION Statements
6.26 ILLUMINATION
ILLUMINATION specifies the basic illumination parameters in Optolith.
Syntax
ILLUMINATION
[I.LINE|G.LINE|H.LINE|KRF.LASER|DUV.LINE|ARF.LASER|F2.LASER|
LAMBDA=<n>]
[X.TILT=<n>][Z.TILT=<n>]
[INTENSITY=<n>]
Description
I.LINE, G.LINE, H.LINE, KRF.LASER (alias DUV.LINE), ARF.LASER, and
F2.LASER specify that the standard wavelengths of the illumination to be used. The
corresponding wavelengths are 0.365, 0.436, 0.407, 0.268, 0.193, and 0.157 microns.
LAMBDA defines or changes the source wavelength. Only monochromatic sources are
assumed for simulation, that is only one wavelength can be specified. The units are microns.
X.TILT and Z.TILT specify the tilt of the illumination system with respect to the optical axis
of the projection system. All values are to be entered in degrees.
INTENSITY defines or changes the absolute value (usually set to one) of the complex
amplitude, that is the intensity in the mask or reticle plane.
Examples
The following statement defines i.line illumination with X and Z tilt of 0.1 and an intensity
of 1.
ILLUMINATION I.LINE X.TILT=0.1 Z.TILT=0.1 INTENSITY=1
For more examples, see IMAGE, PROJECTION, ILLUM.FILTER, PUPIL.FILTER, ABERRATION,
LAYOUT.

IMAGE Statements
6.27 IMAGE
IMAGE calculates a one or two dimensional aerial image.
Syntax
IMAGE
[INFILE=<c>] [DEMAG=<n>] [GAP=<n>]
[OPAQUE|CLEAR] [DEFOCUS=<n>][CENTER]
[WIN.X.LOW=<n>] [WIN.X.HIGH=<n>] [WIN.Z.LOW=<n>] [WIN.Z.HIGH=<n>]
[DX=<n>] [DZ=<n>] [X.POINTS=<n>] [Z.POINTS=<n>] [N.PUPIL=<n>]
[MULT.IMAGE] [X.CROSS|Z.CROSS] [ONE.DIM]
Description
This statement calculates a 2D aerial image and sets parameters that control the accuracy,
input, and output of the imaging module. The IMAGE statement accepts layout information
created by MaskViews.
INFILE is the name of the mask data file from MaskViews. It contains coordinates of
rectangular mask features as well as the transmittance and phase of each feature. This file
name usually ends with the extension.sec.
Note: For more information on the alternative method of loading MaskViews layout information for image
calculations, see the LAYOUT statement.
DEMAG specifies demagnification factor. If specified, all elements of layout as well as all
parameter of image window and grid will decrease.
GAP specifies the mask-to-wafer gap for the case of contact printing. The units are microns.
OPAQUE and CLEAR specify the type of mask to be used. The background will be opaque
if you select OPAQUE, while the mask features will be clear. The background will be clear if
you select CLEAR and the mask features will be opaque.
DEFOCUS is a user specified defocus parameter. If < 0, above the resist. If > 0, below the
resist surface.
CENTER specfies that layout loaded using the INFILE parameter will be shifted so its center
is in the point (0,0)-- the origin of coordinates for the computational window.
WIN.X.LOW, WIN.X.HIGH, WIN.Z.LOW, and WIN.Z.HIGH set the minimum and
maximum x and z values that define the image window. If unspecified, default values from
the mask file will be used. The units are microns.
DX specifies the mesh resolution for the image window in x. If DX is not specified,
X.POINTS and Z.POINTS will be used. The units are microns.
DZ specifies the mesh resolution for the image window in z. The default is DZ=DX. The units
are microns.

IMAGE Statements
X.POINTS and Z.POINTS are the number of x and z coordinate points in the image
window, respectively. These parameters are used only if DX is not specified. Default value is
10 for both coordinates.
N.PUPIL defines or changes the number of mesh points in the projector’s exit pupil used in
imaging simulations. The value of N.PUPIL sets the number of mesh points along the exit
pupil’s radius. Larger values provide better accuracy. The default setting should be adequate
for accuracy. N.PUPIL also sets the size of the mask or image cell for imaging simulations.
Finally, N.PUPIL affects the discretization of the source. This means that if a very fine source
discretization is required, N.PUPIL should be set to a larger value. Note that computation time
grows linearly with the number of pupil mesh points and source points used in the simulation.
MULT.IMAGE specifies that the preceding and current images will be added. You can add
any number of images as long as the IMAGE command contains this boolean.
X.CROSS and Z.CROSS specify if the one dimensional image is parallel to the x-axis or z-
axis respectively.
ONE.DIM use a one dimensional image module that images a line drawn across a 2D layout.
This is the best method when the image will be used for subsequent EXPOSE statements.
Example
This statement loads a mask named MASK.SEC and specifies x resolution in the image
window of DX=0.1 micrometers. It then runs the imaging module.
IMAGE INFILE=MASK.SEC DX=0.1
For more information, see ILLUMINATION, PROJECTION, ILLUM.FILTER, PUPIL.FILTER,
ABERRATION, LAYOUT, and EXPOSE.

IMPLANT Statements
6.28 IMPLANT
IMPLANT specifies an ion implantation process step.
Syntax
IMPLANT
[GAUSS|PEARSON|FULL.LAT|MONTECARLO|BCA] [CRYSTAL|AMORPHOUS]
[PROFILE] [MEASURED.SIMS=<c> | EXTRACT.SIMS=<c>]
IMPURITY ENERGY=<n> DOSE=<n> [FULL.DOSE]
[TILT=<n>] [ROTATION=<n>] [FULLROTATION]
[PLUS.ONE] [DAM.FACTOR=<n>] [DAM.MOD=<c>] [PRINT.MOM]
[X.DISCR=<n>] [LAT.RATIO1] [LAT.RATIO2] [S.OXIDE=<n>]
[MATCH.DOSE|RP.SCALE|MAX.SCALE][SCALE.MOM][ANY.PEARSON]
[IMPL.TAB=<c>]
[N.ION=<n>] [MCSEED=<n>] [TEMPERATURE=<n>] [DIVERGENCE=<n>]
[IONBEAMWIDTH=<n>]
[IMPACT.POINT=<n>] [SMOOTH=<n>] [SAMPLING] [DAMAGE]
[CRIT.ENERGY=<n>]
[MISCUT.TH][MISCUT.PH]
[TRAJ.FILE=<n>] [N.TRAJ=<n>]
[Z1=<n>] [M1=<n>] [REFLECTIVE]
Description
This statement simulates ion implantation using different analytical and Monte Carlo models.
Model Selection Parameters
GAUSS, PEARSON, FULL.LAT, MONTECARLO, and BCA specify the implant model
that is being used. GAUSS selects a Gaussian distribution. PEARSON selects the Pearson-IV
distribution or where available dual Pearson-IV distributions. FULL.LAT is the same as
PEARSON with lateral component of the 2D distribution calculated using all available
moments instead of just a lateral standard deviation. MONTECARLO (synonym is BCA) activates
the Monte Carlo Implant Module, which based on the Binary Collision Approximation.
CRYSTAL and AMORPHOUS specify whether or not the silicon lattice structure is to be
taken into account during implant steps. The statements are mutually exclusive and CRYSTAL
is true by default. For implants through thick screen materials, you often need to specify
AMORPHOUS to avoid incorrect channeling profiles.
• For analytical implant models, these parameters select which set of tables are used for
silicon implant ranges. The CRYSTAL model uses the SVDP tables where available and is
the default.
• For MONTECARLO or BCA models these parameters control whether the crystalline lattice
structure is considered or not.
PROFILE specifies that a profile data to be used in silicon instead of analytical model
(GAUSS or PEARSON). The data might come either from measured SIMS profile or from
"SIMS" profile extracted by the EXTRACT command of DeckBuild.

IMPLANT Statements
MEASUREAD.SIMS specifies name of the file that contains measured SIMS profile used in
profile model indicated by the PROFILE statement. The file consist of two columns: depth in
microns and concentration in ion/cm3.
EXTRACT.SIMS specifies the name of the file that contains extracted SIMS profile.
Parameters Applicable for All Implant Models
IMPURITY specifies the impurity to be implanted (see Section 6.2.10 “Standard Impurities”
for the list of impurities). BF2 is also available.
ENERGY specifies the implant energy in keV.
DOSE specifies the dose of the implant. Dose is calculated in a plane normal to the implant
direction. The units are in cm-2.
FULL.DOSE specifies that the implanted dose is adjusted to compensate for the tilt angle.
This type of dose specification is often used for high tilt implants.
Adjusted Dose = DOSE/cos(TILT)
TILT specifies the tilt with respect to the vertical of the implantation ion beam. The units are
degrees. The default is 7°.
ROTATION specifies the angle of rotation of the implant relative to the plane of the
simulation. The units are degrees.The default is 30°.
FULLROTATION specifies that the implant be performed at all rotation angles.
PLUS.ONE (synonyms are UNIT.DAMAGE and D.PLUS), and DAM.FACTOR
(synonym is D.SCALE) specify the implant damage calculation. UNIT.DAMAGE specifies that
the interstitial profile should be a scaled version of the doping profile from the implant.
DAM.FACTOR specifies the scaling factor to be used for the UNIT.DAMAGE model. At a depth,
the interstitial concentration from the UNIT.DAMAGE model will equal to the implanted ion
concentration multiplied by DAM.FACTOR.
PRINT.MOM prints out moments for all ion/material combinations used in the analytical
model. In the case of Monte Carlo simulation, it prints out moments calculated from the
coordinates of ion in the standard structure file and can extracte them by the EXTRACT
function.
DAM.MOD specifies the name of the C-Interpreter file, which can be used to modify defect
concentration models.
Parameters Applicable Only for Analytical Implant Models
X.DISCR specifies the width of slices along the direction of the ion beam used to calculate
the implanted profile. The value used is scaled relative to the lateral straggling of the current
implant. By default, a slice width of between 0.1 and 0.2 of the average lateral straggle will
be used. This parameter allows you to override the internal selection of discretization along
the implant front. If the value of X.DISCR decreases, simulation accuracy and simulation time
will increase.
LAT.RATIO1 specifies a factor by which all lateral standard deviations for the first Pearson
distribution would be multiplied. Default is 1.0.
LAT.RATIO2 specifies a factor by which all lateral standard deviations for the second
Pearson distribution would be multiplied. Default is 0.2.

IMPLANT Statements
Note: The LAT.RATIO* parameters provide simple scaling of the default lateral standard deviation. Use the
MOMENTS statement for more complete lateral standard deviation modifications.
S.OXIDE specifies screen oxide parameter for the SVDP implant model. Default is 0.001
microns. The screen oxide thickness is not determined from the structure and must be user-
specified. See Section 3.5 “Ion Implantation Models” for more details and the on-line
examples on how to set this parameter automatically.
MATCH.DOSE, RP.SCALE (synonym is RP.EFF), and MAX.SCALE specify the method
for implant calculations in multi-material structures (see Section 3.5.2 “Multi-Layer
Implants”). Default is MATCH.DOSE.
SCALE.MOM specifies that moment scaling to be used with selected multilayer implant
model.
ANY.PEARSON specifies no restrictions on the combinations of allowed skewness and
kurtosis. This is true by default as required for the SVDP models. See Section 3.5 “Ion
Implantation Models” for details on potential problems with this setting. Athena versions
earlier than 4.0 had this parameter set to false by default.
IMPL.TAB specifies the table from the T-SUPREM4 default implant tables file.
Parameters Applicable Only for Monte Carlo/BCA Implant Models
N.ION specifies the number of ion trajectories to be calculated for the Monte Carlo method.
When the SAMPLING is not specified, the default is N.ION is 1,000 for 1D structures and
10,000 for 2D structures.
MCSEED specifies a seed for the random number generator used for the Monte Carlo
calculation.
TEMPERATURE specifies the temperature of the substrate during implantation.

DIVERGENCE specifies the implant beamwidth in degrees. When the BEAMWIDTH angle is
specified the TILT angle is varied between TILT +/- DIVERGENCE/2.0. Each ion will have
an angle somewhere in this range decided by a random number generator. Distribution of the
ions is uniform across the defined angular distribution. Correct specification of DIVERGENCE
is generally required for accurate zero degree implant ranges. Default is 1 degree.
IMPACT.POINT specifies (only in the Monte Carlo method) that the ion beam enters the
surface in the point with lateral coordinate x=(left+IMPACT.POINT*L), where left is the
x-coordinate of the left boundary of the structure and L is the length of the structure. This
parameter would be used for calculation of the point-source 2D distribution and spacial
moments using Monte Carlo method.
IONBEAMWIDTH specifies the ion beam width in nanometers. It can only be used with the
IMPACT.POINT parameter.
SMOOTH specifies that a special Gaussian convolution smoothing to be applied to the
Monte Carlo results. SMOOTH multiplied by estimated standard deviation of the whole profile
serves as the standard deviation for the Gaussian formula.
SAMPLING specifies that statistical sampling to be used in the Monte Carlo method.

IMPLANT Statements
Note: When you specify SAMPLING parameter, the multi-threading will be switched off and all ion trajectories will
be simulated on one CPU. In many cases, the statistical sampling method can be more accurate than
running larger amount of trajectories using multiple CPUs.
DAMAGE specifies that damage (in displacements/cm3) is calculated and saved into the
structure file during Monte Carlo implant simulation. This damage distribution is served only
for estimation of depth and lateral spread of amorphization area.
CRIT.ENERGY specifies critical deposition energy for amorphization. If the value of this
parameter is less than 1000, then it is used as a scaling factor for calibrated default value.
Otherwise, it will be user-specified value in units of eV/cm3.
MISCUT.TH and MISCUT.PH specify the wafer's miscut. See “Implantation Geometry in
SiC” on 214 for more information.
TRAJ.FILE specifies the name of the file in which ion trajectories calculated with the Monte
Carlo (BCA) method are to be saved.
Note: This parameter switches off statistical sampling if it’s specified by the SAMPLING parameter.
Note: When TRAJ.FILE parameter is specified, the Monte Carlo implant simulation will be performed on a
single processor, even if the multiple processors are available and parallel capability is specified by the -P
parameter in the GO ATHENA statement.
N.TRAJ specifies the number of ion trajectories to be saved in the TRAJ.FILE. The default is
minimum of N.ION and 2000.
Z1 specifies the atomic number of an "inert" ion used only for damage or amorphization of
substrate. No new impurity will be introduced into the structure except "damage", which will
affect subsequent Monte Carlo implants.
M1 specifies the atomic weight of the "inert" ion with atomic weight of Z1. If M2 is not
specified, the atomic weight of the main isotope will be used.
REFLECTIVE specifies that reflective boundary conditions will be used at the left and right
sides of the structure. The reflective boundary conditions are applicable for either TILT=0 or
for tilted implant with FULLROTATION. The reflective boundary conditions are also applicable
with non-zero ion beam DIVERGENCE.
Analytical Implant Example
This example specifies that a 100keV implant of phosphorus to be done with a dose of 1.0 e14
and with a tilt angle of 15° to the surface normal. The Pearson model is to be used to calculate
the doping profile.
IMPLANT PHOSPH DOSE=1E14 ENERGY=100 TILT=15
SVDP Boron Implant Example
This example shows the syntax for a zero tilt and 50keV boron implant through 5nm of screen
oxide. The oxide is defined by S.OXIDE and this definition is independent of any actual oxide
in the structure itself.
IMPLANT BORON DOSE=1E13 ENERGY=50 TILT=0 S.OXIDE=0.005

IMPLANT Statements
Monte Carlo Implant Example

This example specifies a 300keV boron implant at zero degrees tilt and rotation. Accurate
modeling of such implants is only possible in the BCA model. Since ion channeling is highly
dependent on the tilt angle.
IMPLANT BORON DOSE=1E13 ENERGY=300 BCA TILT=0 ROTATION=0
Implant Damage Example
This example implants phosphorus and invokes the unit damage model. The UNIT.DAMAGE
model creates an interstitial profile scaled to the implant doping profile. DAM.FACTOR is used
here to specify that the interstitial concentration will be ten times less than the doping
throughout the depth of the implant profile.
IMPLANT PHOSPHORUS DOSE=1E14 ENERGY=50 UNIT.DAMAGE DAM.FACTOR=0.1
SiC Implant Examples
Example 1
This example demonstrates equivalent implant directions with wafer’s major flat, <1 1 00>.
implant aluminum n.ion=40000 dose=0.63e13 tilt=0 rot=0 energy=60
bca
The equivalent implant direction with miscut wafer 4° <11 2 0> toward will be
implant aluminum n.ion=40000 dose=0.63e13 tilt=4 rot=90 miscut.th=4
miscut.ph=90 energy=60 bca
The two profiles will be statistically equivalent with slight shift due to cos(4°) projection of
the second profile.
Example 2
This example demonstrates equivalent implant directions with wafer’s major flat, <11 2 0>.
MISCUT.TH/PH are measured from the internal crystallographic coordinate system, not the
laboratory one defined by ROT.SUB.
bca
The equivalent implant direction with miscut wafer 4° toward ?1120? will be
The two profiles will be statistically equivalent with slight shift due to cos(4°) projection of
the second profile.
Example 3
The following example demonstrates equivalent implant conditions for 3° tilted off [0001]
direction toward <11 2 0> implant for ideal, and for miscut 4° toward <11 2 0> 4H-SiC wafer.

IMPLANT Statements

bca
The equivalent implant direction with miscut wafer 4° toward <11 2 0> will be
The two profiles will be statistically equivalent with slight shift due to both profiles
projection on the vertical, Y, coordinate.

IMPURITY Statements
6.29 IMPURITY
IMPURITY specifies impurity parameters.
Note: This statement supersedes the older syntax using separate statements for each impurity type. The
ARSENIC, ANTIMONY, BORON, INDIUM and PHOSPHORUS statements should no longer be used.
Syntax
IMPURITY
I.IMPURITY [DONOR|ACCEPTOR|NEUTRAL] MATERIAL
[AT.NUMBER=<n>][AT.MASS=<n>]
[DIX.0=<n>] [DIX.E=<n>] [DIP.0=<n>] [DIP.E=<n>]
[DIPP.0=<n>] [DIPP.E=<n>] [DIM.0=<n>] [DIM.E=<n>]
[DIMM.0=<n>] [DIMM.E=<n>] [DVX.0=<n>] [DVX.E=<n>]
[DVM.0=<n>] [DVM.E=<n>] [DVMM.0=<n>] [DVMM.E=<n>]
[DVP.0=<n>] [DVP.E=<n>] [DVPP.0=<n>] [DVPP.E=<n>]
[SOL.SOLUB|CLUSTER.ACT] [PROP.ACTIV=<n>]
[CTN.0=<n>] [CTN.E=<n>] [CTP.0=<n>][CTP.E=<n>]
[SS.CLEAR][SS.TEMP=<n>][SS.CONC=<n>]
[ACT.FACTOR=<n>] [TRACT.0=<n>] [TRACT.E=<n>] [TRACT.MIN=<n>]
[/MATERIAL] [SEG.0=<n>][SEG.E=<n>][TRN.0=<n>][TRN.E=<n>]
[TRNDL.0=<n>][TRNDL.E=<n>]
[PD.DIX.0=<<n>>] [PD.DIX.E=<n>] [PD.EFACT=<n>] [PD.SEG.E=<n>]
[PD.TAU=<n>][PD.SEGSITES=<n>] [PD.GROWTH.0=<n>] [PD.GROWTH.E=<n>]
[PD.CRATIO=<n>][PD.SEG.GBSI=<n>]
[AXX.STRAIN] [BXX.STRAIN] [CXX.STRAIN] [AYY.STRAIN] [BYY.STRAIN]

[CYY.STRAIN] [E.STRAIN][A.YOUNG]
Description
This statement allows to specify coefficients of impurity diffusion, transport, segregation, and
so on.
Generic Parameters
I.IMPURITY is the name of impurity which parameters to be specified (see Section 6.2.10
“Standard Impurities” for the list of impurities).
DONOR, ACCEPTOR and NEUTRAL specify the type of the impurity in the given
material. Default is NEUTRAL.
MATERIAL specify the material in which the impurity parameters apply as well as
MATERIAL1 for the segregation and transport parameters on the boundary between two
materials (see Section 6.2.9 “Standard and User-Defined Materials” for the list of materials).

IMPURITY Statements
AT.NUMBER and AT.MASS specify the atomic number and atomic mass of the impurity
respectively. This parameters are used in Monte Carlo/BCA implant calculations.
Diffusion Parameters
The units for all pre-exponential diffusion constants are cm2/sec, while the units for activation
energies are eV.
DIX.0 and DIX.E specify the diffusion coefficient for the impurity diffusing with neutral
interstitials. DIX.0 is the pre-exponential constant and DIX.E is the activation energy.
DIP.0 and DIP.E specify the diffusion coefficient for the impurity diffusing with single
positive interstitials. DIP.0 is the pre-exponential constant. DIP.E is the activation energy.
DIPP.0 and DIPP.E specify the diffusion coefficients for the impurity diffusing with double
positive interstitials.
DIM.0 and DIM.E specify the diffusion coefficient for the impurity diffusing with single
negative interstitials. DIM.0 is the pre-exponential constant. DIM.E is the activation.
DIMM.0 and DIMM.E specify the impurity diffusing with doubly negative interstitials.
DIMM.0 is the pre-exponential constant. DIMM.E is the activation energy.
DVX.0 and DVX.E specify the impurity diffusing with neutral vacancies. DVX.0 is the pre-
exponential constant. DVX.E is the activation energy.
DVM.0 and DVM.E specify the impurity diffusing with single-negative vacancies. DVM.0 is
the pre-exponential constant. DVM.E is the activation energy.
DVMM.0 and DVMM.E specify the impurity diffusing with double-negative vacancies.
DVMM.0 is the pre-exponential constant. DVMM.E is the activation energy.
DVP.0 and DVP.E specify the impurity diffusing with single-positive vacancies. DVP.0 is the
pre-exponential constant. DVP.E is the activation energy.
DVPP.0 and DVPP.E specify the impurity diffusing with double-positive vacancies. DVPP.0
is the pre-exponential constant. DVPP.E is the activation energy.
FI.0 and FI.E are the fractional interstitialcy parameters that determine whether the impurity
diffuses through interaction with interstitials or vacancies. Once the expression for total FI is
evaluated from these coefficients, the value of total FI can vary between 0 and 1. FI equal to
1 corresponds to a pure interstitial-based diffusion, while value of 0 corresponds to a pure
vacancy mechanism.
Activation Model Parameters
SOL.SOLUB specifies that solid solubility model and solid solubility tables will be used for
calculation of active concentration of the specified impurity in the specified material. This is
default for all cases except Arsenic in Si and Poly.
CLUSTER.ACT specifies that cluster activation will be used. It is default only for AS in Si
and Poly.
SS.CLEAR, SS.TEMP, and SS.CONC are the parameters for solid solubility data.
SS.CLEAR clears the currently stored solid solubility data for the specified impurity in the
specified material. SS.TEMP and SS.CONC add a single temperature and an associated solid
solubility concentration point to those already stored for the impurity. Units for SS.TEMP are
oC. Units for SS.CONC are cm-3.

IMPURITY Statements
CTN.0 and CTN.E specify the vacancy clustering coefficients for the impurity. CTN.0 is the
pre-exponential coefficient and CTN.E is the activation energy. By default, these parameters
are only used for Arsenic.
CTP.0 and CTP.E specify the vacancy clustering coefficients.
ACT.FACTOR specifies parameter for concentration dependent solid solubility activation
model. The value of ACT.FACTOR must be between 0.8 and 1.0.
PROP.ACTIVE specifies coefficient of the proportional activation model. If this parameter
is specified, then in every point of specified material, the active concentration of the specified
impurity will be equal to the chemical concentration multiplied by the value of this parameter.
This model is mainly used for silicon carbides. The value must be more or equal to 0.0 and
less or equal to 1.0.
TRACT.0, TRACT.E, and TRACT.MIN specify parameters of the transient activation
model. Units for TRACT.0 and TRAT.MIN are seconds. Units for TRACT.E are eV.
Interface Transport Parameters
/MATERIAL specify MATERIAL2 for the segregation and transport parameters on the
boundary between two materials (see Section 6.2.9 “Standard and User-Defined Materials”
for the list of materials).
SEG.0 and SEG.E allow the computation of the equilibrium segregation concentrations.
SEG.0 is the unitless pre-exponential constant. SEG.E is the activation energy in eV.
TRN.0 and TRN.E allow the specification of the transport velocity across the interface given.
TRN.0 is the pre-exponential constant (units are cm/sec). TRN.E is the activation energy
(units are eV).
TRNDL.0 and TRNDL.E specify parameters of the Interface Trap Model, describing dose
loss at silicon/oxide interface. Units for TRNDL.0 are cm/sec; units for TRNDL.E are eV.
Polysilicon Diffusion Model Parameters
PD.DIX.0 and PD.DIX.E specify impurity diffusivity along grain boundaries. PD.DIX.0 is
the pre-exponential factor of grain boundary diffusivity (units are cm2/sec). PD.DIX.E is the
activation energy for grain boundary diffusivity (units are eV).
PD.EFACT specifies entropy factor of grain boundary segregation coefficient.
PD.SEG.E specifies the activation energy of grain boundary segregation coefficient. Units
are eV.
PD.TAU specifies the grain boundary time constant. Units are seconds.
PD.SEGSITES specifies density of segregation sites at grain boundary. Units are sites/cm2.
PD.GROWTH.0 specifies the grain growth rate pre-exponential coefficient. Units are
eV*cm2/sec.
PD.GROWTH.E specifies the grain growth rate activation energy. Units are eV.
PD.CRATIO specifies initial ratio between impurity concentration in grain boundaries and
total concentration.
PD.SEG.GBSI specifies the factor which controls segregation between polysilicon grain
boundaries and Silicon.

IMPURITY Statements
Parameters Related to Stress Models

AXX.STRAIN specifies the constant term of the doping dependent XX component of
intrinsic strain.
BXX.STRAIN specifies the linear term of the doping dependent XX component of intrinsic
strain.
CXX.STRAIN specifies the quadratic term of the doping dependent XX component of
intrinsic strain.
AYY.STRAIN specifies the constant term of the doping dependent YY component of
intrinsic strain.
BYY.STRAIN specifies the linear term of the doping dependent YY component of intrinsic
strain.
CYY.STRAIN specifies the quadratic term of the doping dependent YY component of
intrinsic strain.
E.STRAIN specifies activation energy per volume change ratio for the stress-dependent
diffusion model.
A.YOUNG specifies the linear coefficient of doping dependent Young’s modulus. The units
are dyne/cm2.
Examples
The following statement changes the neutral interstitial diffusivity component of phosphorous
in silicon.
IMPURITY I.PHOSPHORUS SILICON DIX.0=3.85 DIX.E=3.85
The following statement changes the segregation parameters at the silicon-silicon dioxide
interface. The concentration of phosphorous in silicon will be 30 times the concentration of
phosphorous in oxide at equilibrium.
IMPURITY I.PHOSPHORUS SILICON /OXIDE SEG.0=30.0 SEG.E=0.0
The following syntax sets the temperature dependent impurity activation of Indium in Silicon.
IMPURITY I.INDIUM SILICON SS.TEMP=800 SS.CONC=<VAL1> SS.CLEAR
IMPURITY I.INDIUM SILICON SS.TEMP=900 SS.CONC=<VAL2>
IMPURITY I.INDIUM SILICON SS.TEMP=950 SS.CONC=<VAL3>
Note: The transport and segregation coefficients, TRN.0, TRN.E, SEG.0, and SEG.E, are known to be
inaccurate for some values of concentration, material combinations, and temperature ranges. If the
simulation is inaccurate, consider these coefficients for calibration.
For more examples, see DIFFUSE, METHOD, INTERSITITIAL, and VACANCY.

INITIALIZE Statements
6.30 INITIALIZE
INITIALIZE specifies the initial starting material and background doping levels. LOADFILE
is a synonym for this statement.
Syntax
INITIALIZE
[MATERIAL] [ORIENTATION=<n>] [ROT.SUB=<n>] [X.COMP=<n>]
[Y.COMP=<n>]
[C.FRACTION=<n>][C.IMPURITIES=<n>|RESISTIVITY=<n>][C.INTERST=<n>]
[C.VACANCY=<n>] [BORON|PHOSPHORUS|ARSENIC|ANTIMONY] [NO.IMPURITY]
[ONE.D|TWO.D|AUTO] [X.LOCAT=<n>] [CYLINDRICAL]
[INFILE=<c>] [STRUCTURE|INTENSITY]
[SPACE.MULT=<n>] [INTERVAL.R=<n>] [LINE.DATA] [SCALE=<n>] [FLIP.Y]
[DEPTH.STR=<n>] [WIDTH.STR=<n>]
Description
This command sets up the mesh from either a rectangular specification or from a previous
structure file. The statement also initializes the background doping concentration in all
regions.
Material Related Parameters
MATERIAL specifies the material to be initialized (see Section 6.2.9 “Standard and User-
ORIENTATION specifies the substrate orientation. Only 100, 110 and 111 are recognized.
The default is 100.
ROT.SUB specifies the major flat of the silicon substrate. It is measured in degrees from the
external x-axis of the crystallographic coordinate system. By default, ROT.SUB=-45, i.e., it
represents the [104] plane. This parameter is used only in BCA implantation module.
X.COMP specifies the composition fraction of the cation listed first in ternary or quaternary
compound semiconductor to be deposited. For example, Al is usually first cation of AlGaAs.
The full list of compound semiconductors that can be specified in Athena and the rules of
their naming can be found in Appendix B of Atlas User’s Manual. For backward
compatibility, C.FRACTION is the alias for X.COMP.
Y.COMP specifies the composition fraction of the anion listed first in quaternary compound
semiconductor to be deposited. For example, As is first anion in InGaAsP. The full list of
compound semiconductors that can be specified in Athena and the rules of their naming can
be found in Appendix B of Atlas User’s Manual.

Dopant Related Parameters

C.IMPURITIES specify the uniform impurity concentration in substrate material (see
Section 6.2.10 “Standard Impurities” for the list of impurities). Multiple parameters can be
used to define compensated doping in the substrate material.
RESISTIVITY specifies the resistivity of the initial substrate material. If RESISITVITY is
specified, the impurity concentration specified by C.*** parameter will be ignored and
calculated from the resistivity vs. concentration tables, which are available only for boron,
phosphorus, arsenic, and antimony. The units are -cm.
BORON, PHOSPHORUS, ARSENIC, ANTIMONY specify the type of impurity when
initial doping is defined by RESISTIVITY parameter.
Note: There are several additional methods to specify impurity concentration in substrate material. These methods
are introduced to provide TSUPREM4 compatibility and described in Section C.7 “Changes in the
INITIALIZE statement”.
C.INTERST specifies the uniform interstitial concentration in substrate material. Units are
cm-3.
C.VACANCY specifies the uniform vacancy concentration in substrate material. Units are
cm-3.
NO.IMPURITY specifies that the calculation be performed without impurities. No
impurities will be introduced during the simulation. This speeds calculation and allows quick
analysis of oxidation, deposit, and etch results.
Parameters Related to Dimensionality of Simulation
ONE.D, TWO.D, AUTO set whether the run will be in 1D, 2D, or the dimensionality
automatically determined from the process flow. ONE.D specifies that calculation will be
performed in 1D at a location indicated by the X.LOCAT parameter. TWO.D specifies that all
process steps will be done in a full two dimensional calculation. If name of these parameters
are specified or AUTO is used, Athena will perform 1D calculation until a two-dimensional
calculation is required. This typically happens at the first ETCH statement, which doesn’t
remove material across the whole width of the structure.
X.LOCAT specifies the position within the defined 2D mesh for performing 1D simulation.
CYLINDRICAL specifies the boundary conditions for cylindrically symmetrical structure.
In this case, the axis of rotation is X=0.0 and no negative x coordinates are allowed.
Parameters Related to Initialization from a File
INFILE specifies a file name for reading. This file must contain a previously saved structure
or intensity distribution (see Section 6.63 “STRUCTURE”). IN.FILE is a synonym for this
parameter.
STRUCTURE and INTENSITY specify which type of file is to be initialized. STRUCTURE is
the default.
Grid and Structure Related Parameters
SPACE.MULT specifies a global spacing multiplier to be applied to the spacings defined on
the previously specified LINE statements.

INTERVAL.R is the maximum ratio between the distances of adjoining mesh lines. The
default is 1.5. RATIO is a synonym for this parameter.
LINE.DATA specifies that locations of mesh lines be printed during execution.
SCALE allows an incoming mesh to be scaled. The default is 1.0.
FLIP.Y is a Boolean parameter that dictates the mesh should be mirrored about the x axis.
DEPTH.STR and WIDTH.STR specify the depth and width of the initial substrate structure
dimension for use with the Process Adaptive Meshing algorithm. Units are microns.
Example Starting from a file
The following statement reads in a previously saved structure from the TEST.STR file.
INITIALIZE INFILE=TEST.STR
Example Using an GaAs Substrate
The following statement creates GaAs substrate doped with Selenium concentration of
11015cm-3.
INITIALIZE GAAS C.SELENIUM=1E15
For more examples, see BOUNDARY, LINE, REGION, STRUCTURE, and BASE.MESH.

INTERSTITIAL and VACANCY Statements
6.31 INTERSTITIAL and VACANCY

INTERSTITIAL specifies coefficients of interstitial diffusion, recombination, and generation.
VACANCY specifies coefficients of vacancy diffusion, recombination, and generation.
Note: These two statements are almost equivalent. Most parameters that exist in the INTERSTITIAL
statement are also on the VACANCY statement.
INTERSTITIAL|VACANCY
[MATERIAL] [D.0=<n>][D.E=<n>] [CSTAR.0=<n>][CSTAR.E=<n>]
[NEU.0=<n>][NEU.E=<n>] [NEG.0=<n>][NEG.E=<n>
[DNEG.0=<n>][DNEG.E=<n>] [POS.0=<n>][POS.E=<n>
[DPOS.0=<n>][DPOS.E=<n>]
[KR.0=<n>][KR.E=<n>] [IVFACTOR=<n>] [IIFACTOR=<n>]
[KTRAP.0=<n>][KTRAP.E=<n>] [DAMALPHA=<n>]
[/MATERIAL=] [TIME.INJ] [GROWTH.INJ] [RECOMB]
[KSURF.0=<n>][KSURF.E=<n>] [KRAT.0=<n>][KRAT.E=<n>]
[KPOW.0=<n>][KPOW.E=<n>]
[VMOLE=<n>] [THETA.O=<n>][THETA.E=<n>] [GPOW.0=<n>[GPOW.E=<n>]
[WET02|DRY02] [REC.STR=<n>][INJ.STR=<n>]
[A.0=<n>][A.E=<n>] [T0.0=<n>][T0.E=<n>] [TPOW.0=<n>][TPOW.E=<n>]
[DCARBON.E=<n>][KCARBON.0=<n>][KCARBON.E=<n>]
Description
These two equivalent commands specify transport and generation/recombination coefficients
for interstitials and vacancies. The statements allow you to specify coefficients for any
material, though it is only practical for semiconductors. Athena has measured (or calibrated)
default values only for silicon and some interfaces with silicon.
MATERIAL specify the material for which the interstitial (or vacancy) parameters apply as
well as MATERIAL1 for the segregation and transport parameters on the boundary between
two materials (see Section 6.2.9 “Standard and User-Defined Materials” for the list of
materials). Default is Silicon.
Defect Diffusion Parameters
D.0 and D.E specify the interstitial (or vacancy) diffusion coefficient. D.0 is the pre-
exponential constant (the units are cm2 /sec), and D.E is the activation energy (the units are
eV).
CSTAR.0 and CSTAR.E specify of the total equilibrium concentration of interstitials (or
vacancies) in intrinsically-doped conditions. CSTAR.0 is the pre-exponential constant (the
units are cm-3), and CSTAR.E is the activation energy (the units are eV).

NEU.0, NEU.E, NEG.0, NEG.E, DNEG.0, DNEG.E, POS.0, POS.E, DPOS.0 and DPOS.E
specify the relative concentration of interstitials (or vacancies) in the various charge states
(neutral, negative, double negative, positive, double positive) under intrinsic doping
conditions. All *.0 parameters are unitless. All *.E parameters are in eV.
Bulk Defect Recombination Parameters
KR.0 and KR.E specify the interstitial (or vacancy) bulk recombination rate. KR.0 is the pre-
exponential constant (the units are cm-3sec-1), and KR.E is the activation energy in eV.
IVFACTOR and IIFACTOR specify I/V Bimolecular recombination ratios in HIGH.CONC
model. These parameters are valid only for the INTERSTITIAL statement.
KTRAP.0 and KTRAP.E specify the interstitial trap reaction rate. KTRAP.0 is the pre-
exponential constant (the units are cm-3sec-1), and KTRAP.E is the activation energy in eV.
Note: At present, it is very difficult to extract exact values for these parameters. The default values assume the trap
reaction is limited by the interstitial concentration.
DAMALPHA specifies the interstitial recombination rate in the dislocation loop region. The
units are sec-1.
Interface Defect Generation and Recombination Parameters
/MATERIAL specify MATERIAL2 for setting generation and recombination parameters on
the boundary between two materials (see Section 6.2.9 “Standard and User-Defined
Materials” for the list of materials).
TIME.INJ, GROWTH.INJ and RECOMB specify the type of reactions occurring at the
specified interface. The TIME.INJ parameter means that a time dependent injection model
should be chosen. The GROWTH.INJ parameter ties the injection to the interface growth
velocity. The RECOMB parameter indicates a finite surface recombination velocity
KSURF.0, KSURF.E, KRAT.0, KRAT.E, KPOW.0 and KPOW.E specify the interstitial (or
vacancy) surface recombination rate. KSURF.0 is the pre-exponential constant for surface
recombination rate under inert conditions (the units are cm/sec). KSURF is corresponding
activation energy in eV. KRAT.0 is the pre-exponential constant for the growth rate dependent
component of the surface recombination rate (unitless). KRAT.E is the corresponding
activation energy in eV. KPOW.0 is the pre-exponential constant of the power parameter in the
surface recombination rate formula (unitless), and KPOW.E is the corresponding activation
energy in eV.
VMOLE, THETA.0, THETA.E, GPOW.0 and GPOW.E specify interstitial (or vacancy)
generation parameters of the growth dependent generation model. VMOLE is the lattice density
of the consumed material (the units are cm-3). THETA.0 specifies the pre-exponential constant
for the fraction of consumed atoms injected as interstitial (or vacancy). THETA.E specifies the
activation energy for the fraction (the units are eV). GPOW.0 and GPOW.E specify pre-
exponential constant and activation energy of the power parameter of the growth injection
formula.
WETO2, DRYO2 specify whether the parameters THETA.0 and THETA.E are for wet
oxidation or dry oxidation. The default is DRYO2.

REC.STR and INJ.STR allow you to specify experimental models for interstitial (or
vacancy) recombination or injection at interfaces. Three macros are defined for use: T is the
time in seconds, and X and Y is the coordinates. If these are specified, they are used in place
of any default model.
A.0, A.E, T0.0, T0.E, TPOW.0 and TPOW.E specify parameters for time-dependent
injection model TIME.INJ. A.0 is the pre-exponential constant for the injection rate and A.E
is the corresponding activation energy. T.0 and T.E are the pre-exponential constant and
activation energy for the time constant in the time-dependent injection formula. TPOW.0 and
TPOW.E are the pre-exponential constant and activation energy for the power constant in the
time-dependent injection formula.
Parameters for Carbon Effects in SiGeC
KCARBON.0 and KCARBON.E specify interstitial recombination rate in carbon sink.
KCARBON.0 is the pre-exponential constant for the rate in sec-1, and KCARBON.E is the
corresponding activation energy in eV.
DCARBON.E specifies the coefficient of interstitial diffusion retardation in SiGe in presence
of carbon impurity. The units are eV.
Basic Example
The following statement specifies the silicon diffusion and equilibrium values for interstitials.
INTERST SILICON DI.0=5.0E-7 D.E=0.0 CSTAR.0=1.0E13 CSTAR.E=0.0
Defect Injection during Oxidation Example
The following statement specifies the oxide-silicon interface injection for DRYO2 ambient is
to be computed using the oxide growth velocity and with 1% of consumed silicon injected as
interstitials.
INTERST SILICON /OXIDE GROWTH VMOLE=5.0E22 \
THETA.0=0.01 THETA.E=0.0
Surface Recombination Example
The following statement specifies the surface recombination velocity at the nitride silicon
interface is 3.5  10-3 cm/sec.
INTERST SILICON /NITRIDE KSURF.0=3.5E-3 KSURF.E=0.0 KRAT.0=0.0
Experimental Injection Model Example
The following statement describes an injection at the silicon oxide interface that
exponentially decays in time.
INTERST SILICON /OXIDE INJ.STR = (10.0E4*EXP( T / 10.0 ))
General Comments
The equivalence of INTERSTITIAL and VACANCY statement syntax is done regardless of
physical meaning. For example, you can define vacancy injection during oxidation although
default parameters are zero.
The models used here are involved in ongoing research. Many of the parameters have
unknown dependencies, such as stress, temperature, starting silicon material, and stacking
fault density.
For more examples, see DIFFUSE and TRAP.

LAYOUT Statements
6.32 LAYOUT
LAYOUT describes manual input of mask features for Optolith.
Syntax
LAYOUT
[LAY.CLEAR=<n>]
[X.LOW=<n>][Z.LOW=<n>][X.HIGH=<n>][Z.HIGH=<n>][X.TRI=<n>]
[Z.TRI=<n>]
[HEIGHT=<n>][WIDTH=<n>][ROT.ANGLE=<n>][X.CIRCLE=<n>][Z.CIRCLE=<n>]
[RADIUS=<n>][RINGWIDTH=<n>][MULTIRING][PHASE=<n>][TRANSMIT=<n>]
[INFILE=<c>][MASK=<c>][SHIFT.MASK=<c>]
Description
This command is used to enter mask information for Optolith. Several LAYOUT statements can
be used in sequence to define complete mask patterns. All coordinate and size parameters are
in microns.
LAY.CLEAR specifies that the currently defined layout should be deleted prior to the
execution of the new layout definition.
X.LOW specifies the minimum x coordinate of the rectangular feature.
Z.LOW specifies the minimum z coordinate of the rectangular feature.
X.HIGH specifies the maximum x coordinate of the rectangular feature.
Z.HIGH specifies the maximum z coordinate of the rectangular feature.
X.TRI specifies the x coordinate of the right angle corner of the triangular feature.
Z.TRI specifies the z coordinate of the right angle corner of the triangular feature.
HEIGHT specifies the height of the right angle triangle feature.
WIDTH specifies the base width of the right angle triangle feature.
ROT.ANGLE specifies the angle of rotation of the feature (-180ROT.ANGLE180) with
respect to the x-axis. The default value is 0. The center of rotation is at the center of the
rectangle and at the right angle corner of the triangle, respectively.
Note: You can only use the X.TRI, Z.TRI, HEIGHT, WIDTH, and ROT.ANGLE parameters for projection
printing model.
X.CIRCLE specifies the x coordinate of the center of the circular or ring feature.
Z.CIRCLE specifies the z coordinate of the center of the circular or ring feature.
RADIUS specifies the radius of the circular feature or the outermost radius of the ring or
multiring feature.
RINGWIDTH specifies the width of ring(s) of the mask feature shaped as a single ring or a
series of concentrated rings.

LAYOUT Statements
MULTIRING specifies that mask feature consists from series of concentered rings with ring
widths and distances between rings specified by the RINGWIDTH parameter. The number N of
the rings in the feature should satisfy the following relation 2N * RINGWIDTH<RADIUS.
Note: You can only use the RINGWIDTH and MULTIRING parameters for proximity printing model specified by
the GAP parameter in the IMAGE statement.
PHASE specifies the phase shift produced by the feature (-180PHASE180). The default
value is 0.
TRANSMIT specifies the intensity transmittance of the feature (0TRANSMIT1). The
default value is unity.
INFILE specifies the name of the MaskViews layout file. This file contains the mask
information. The MaskViews mask information is a set of polygons for each mask layer with
attributes (transmittance and phase shift) for each polygon.
MASK specifies the name of mask to be used for image calculation.
SHIFT.MASK specifies the name of the additional mask layer - usually phase shifting layer.
Examples
The following statement describes a mask feature that is 2 microns in the x dimension and 0.4
microns in the z direction and rotated by 45 with respect to the x-axis.
LAYOUT X.LO=-1 X.HI=1 Z.LO=-0.2 Z.HI=0.2 ROT.ANGLE=45 \
TRANSMIT=1
For more examples, see IMAGE, ILLUMINATION, PROJECTION, ILLUM.FILTER,
PUPIL.FILTER, ABERRATION, and the MaskViews User’s Manual.

LINE Statements
6.33 LINE
LINE specifies a mesh line during grid definition.
Syntax
LINE
X|Y LOCATION=<n> [SPACING=<n>] [TAG=<c>] [TRI.LEFT|TRI.RIGHT]
Description
This statement defines the position and spacing of mesh lines. All LINE statements should
come before the REGION and BOUNDARY statements, which should then be followed by an
INITIALIZE statement.
X and Y specify whether a mesh line is horizontal or vertical.
LOCATION specifies the location along the chosen axis (in microns) at which the line
should be positioned. The x coordinate increases from left to right; the y coordinate increases
progressing from top to bottom going into the substrate. This is the opposite of normal
Cartesian y-axis progression which increases going upward.
SPACING specifies the local grid spacing (in microns). Athena adds mesh lines to the ones
given according to the following recipe. Each user line has a spacing whether user-specified
or inferred from the nearest neighbor. These spacings are then smoothed out so no adjacent
intervals have a ratio greater than the value given by INTERVAL.R on the INITIALIZE
statement (default is 1.5). New grid lines are then introduced so that the line spacing varies
geometrically from one end of the interval to the other. Refer to the example below.
TAG labels lines for later reference by BOUNDARY and REGION statements. The tag label can
be any word.
TRI.LEFT and TRI.RIGHT can be specified in the LINE X statement to control triangle
orientation in the initial grid. Initial simulation area is divided into non-uniform rectangular
elements defined by the LINE X and LINE Y statements. By default each box element is then
divided into two rectangular triangles by a diagonal going from the bottom-left to upper-right
corner of the box. In some applications it is preferable to have a symmetrical grid triangle
orientation locally.
One of the examples is etching of a non-vertical trench. If TRI.RIGHT (default) is specified
all boxes between this line and the line with TRI.LEFT (or the last line) will be divided by a
"bottom-left to upper-right" diagonal. If TRI.LEFT is specified the boxes will be divided by a
diagonal going from upper-left to bottom-right corner of the box.
Example
In the following specifications, there are 3 user-specified x lines and 2 user-specified y lines.
Spacing of the x lines is finer in the center than at the edges. After processing, Athena
produces a mesh with x lines at 0.0, 0.42, 0.69, 0.88, 1.0,1.12, 1.31, 1.58, 2.0. Around the
center, the spacing is 0.12, approximately what was requested. At the edge, the spacing is
0.42 because that was as coarse as the line spacing could get without having an interval ratio
greater than 1.5. If the interval ratio is set to 9, then we would have one interval of 0.9 and one
interval of 0.1 on each side. In this example, specifying a spacing of 1 would produce an x
line at 0.0 and 1.0.
LINE X LOC=0 SPA=1 TAG=LEFT

LINE Statements
LINE X LOC=1 SPA=0.1

LINE X LOC=2 SPA=1 TAG=RIGHT
LINE Y LOC=0 SPA=0.02 TAG=SURF
LINE Y LOC=3 SPA=0.5 TAG=BACK
Note: It is difficult to predict how many lines are going to be generated in each interval. The initial mesh
specification is quite important to the success of the simulation. Use the geometric mode by specifying the
NO.IMP parameter on the INITIALIZE statement to perform a fast simulation without impurities to
determine if the grid spacings are appropriate.
For more examples, see INITIALIZE, REGION, BASE.MESH, and BASE.PAR.

MASK Statements
6.34 MASK
MASK deposits and patterns photoresist or artificial masking material barrier through the
MaskViews interface.
Syntax
MASK
NAME=”<c>” [REVERSE] [DELTA=<n>]
Description
MASK is used in DeckBuild to provide interface to Silvaco’s general purpose layout editor,
MaskViews. When you specify a mask statement, Athena will deposit photoresist and pattern
it by etching. The etched pattern is determined by selected cut line in MaskViews. See the
MaskViews User’s Manual for a complete description of this feature.
NAME specifies the name of the layer that defines the photoresist patterning. Mask names
must appear inside of double quotes. This name must correspond to a mask name contained in
the layout file invoked using DeckBuild. The mask names are case sensitive and cannot be
abbreviated.
REVERSE specifies that the mask polarity should be reversed or that negative type
photoresist should be modeled.
DELTA specifies an offset in mask size. The offset corresponds to a change in CD (critical
dimension) of the mask. Each edge of the mask is moved by a distance DELTA to enlarge or
contract the mask feature.
Examples
The following statement deposits photoresist on the top of the simulation structure and etches
it with the pattern prescribed by the MaskViews layout. The layout file must be specified
using the MaskViews interface as described in the MaskViews User’s Manual or in Chapter 2
“Tutorial”.
MASK NAME="CONT"
For more examples, see STRIP.

MATERIAL Statements
6.35 MATERIAL
MATERIAL sets the coefficients for materials.
Syntax
MATERIAL
[MATERIAL] [NI.0=<n>] [NI.E=<n>] [NI.POW=<n>] [EPS=<n>][E.FIELD]
[VISC.0=<n>][VISC.E=<n>][VISC.X=<n>] [WETO2|DRY02]
[S11.ELAST=<n>] [S12.ELAST=<N>]
[S44.ELAST=<n>][LCTE=<c>][INTRIN.SIG=<n>]
[OXIDIZABLE] [IMMOBILE] [DENSITY=<n>] [COMPONENTS]
[AT.NUM.1=<n>][AT.NUM.2=<n>][AT.NUM.3=<n>][AT.NUM.4=<n>]
[AT.MASS.1=<n>][AT.MASS.2=<n>][AT.MASS.3=<n>][AT.MASS.4=<n>]
[ABUND.1=<n>][ABUND.2=<n>][ABUND.3=<n>][ABUND.4=<n>]]
[REFLOW][GAMMA.REFLO=<n>] [NO.FLIP]
[NIFACT.SIGE=<n>][EAFACT.SIGE=<n>][NIFACT.SIC][EAFACT.SIC]
Description
MATERIAL specify the material for which all parameters apply (see Section 6.2.9 “Standard
and User-Defined Materials” for the list of materials).
Parameters Related to Impurity Diffusion
NI.0, NI.E, and NI.POW specify parameters of the intrinsic electron concentration as a
function of temperature. NI.0 is the preexponential constant in the intrinsic electron
concentration formula. NI.E is the corresponding activation energy. NI.POW is the unitless
power constant. These parameters are used only in diffusion calculation and not in EXTRACT
electrical calculations
EPS specifies the relative dielectric permittivity of the material. This value is used to
calculate electric field in semiconductors during diffusion simulation. This value isn’t used in
EXTRACT electrical calculations.
E.FIELD specifies that electric field term will be included in the impurity diffusion equations
for this material. The default is true.
Parameters Related to Material Stress and Viscosity
VISC.0, VISC.E, and VISC.X specify the material viscosity parameters. VISC.0 is the pre-
exponential coefficient, in g/(cm·s). VISC.E is the activation energy, in eV. VISC.X is the
incompressibility factor.
WETO2 and DRYO2 specify whether the viscosity parameters are for wet or dry oxidation.
These parameters are valid only if the specified material is oxide.
LCTE specifies the linear coefficient of thermal expansion as a function of temperature (T).
It is expressed as a fraction rather than a percentage.
INTRIN.SIG specifies the initial uniform stress state of a material, such as a thin film of
nitride deposited on the substrate. The units are dyne cm-2.

MATERIAL Statements
S11.ELAST, S12.ELAST, and S44.ELAST are elastic compliances used to calculate

Young’s modulus and Poisson’s ratio for stress calculations and compless oxidation.
Parameters Related to Oxidation
OXIDIZABLE specifies that the material could be oxidized. If you set OXIDIZABLE to TRUE,
then all oxidation related parameters for this material will be set equal to those specified
earlier for Silicon. You can specify different values in the OXIDE statement.
IMMOBILE specifies that this material will not move during oxidation. The default is
FALSE. This is useful when a non-oxidizable material(s) is a continuation of a substrate, so it
is supposed to be immobile and not distorted or lifted as in a case when it serves as a
"floating" mask.
Parameters Related to Material Characteristics used by Monte Carlo Implant
DENSITY specifies the density of the material in g/cm3.
COMPONENTS specifies number of atomic components in the material.
AT.NUM.1, AT.NUM.2, AT.NUM.3, and AT.NUM.4 specify the atomic numbers of the
constituent atoms of the material. The number of AT.NUM parameters specified must
correspond to COMPONENTS.
AT.MASS.1, AT.MASS.2, AT.MASS.3, and AT.MASS.4 specify the atomic masses of the
constituent atoms of the material in atomic mass units. The number of AT.MASS parameters
specified must correspond to COMPONENTS.
ABUND.1, ABUND.2, ABUND.3, and ABUND.4 specify the relative fraction of the
constituent atoms of the material. The number of ABUND parameters specified must
correspond to COMPONENTS.
Note: At least one parameter from each of the four lines above are required to define materials for Monte Carlo
implants.
Parameters Related to REFLOW Calculations

REFLOW specifies that the material will flow when a DIFFUSE step including REFLOW is
defined.
GAMMA.REFLO specifies the surface tension parameter used in the reflow calculation.
Units are dyne/cm. The material viscosity (VISC.* parameters) will also affect the reflow
rate.
Parameters Related to the Grid Control
NO.FLIP specifies that triangle flipping procedure should not be applied to the specified
material.
Parameters Related to the Boron Diffusion Model in SiGe/SiGeC
NIFACT.SIGE specifies the linear coefficient for Ge dependency formula of intrinsic carrier
concentration for Boron diffusion model in SiGe/SiGeC.
EAFACT.SIGE specifies the linear coefficient for Ge dependency formula of intrinsic carrier
concentration for Boron diffusion model in SiGe/SiGeC.
NIFACT.SIC specifies linear coefficient for Ge content dependency formula of intrinsic
carrier concentration for Boron diffusion model in SiGe/SiGeC.

MATERIAL Statements
EAFACT.SIC specifies linear coefficient for Ge content dependency formula of intrinsic

carrier concentration for Boron diffusion model in SiGe/SiGeC.
Example
The following statement defines some properties of a material called BPSG. The material is
composed of silicon, oxygen, boron, and phosphorus with fraction composition 0.3, 0.6, 0.05,
and 0.05 respectively. Monte Carlo Implants could be performed into this material based on
this definition.
MATERIAL MATERIAL=BPSG COMPONENTS=4 DENSITY=2.362 \
AT.NUM.1=14 AT.NUM.2=8 AT.NUM.3=5 AT.NUM.4=15 \
AT.MASS.1=28.0 AT.MASS.2=16.0 AT.MASS.3=10.8 \
AT.MASS.4=31.0 \
ABUND.1=0.3 ABUND.2=0.6 ABUND.3=0.05 ABUND.4=0.05
For more information, see OXIDE, STRESS, and DIFFUSE.

METHOD Statements
6.36 METHOD
METHOD selects numerical methods and models for diffusion and oxidation.
Syntax
METHOD
[FERMI|TWO.DIM|STEADY|FULL.CPL]
[PLS] [IC] [VC] [DDC] [SS]
[CLUSTER.DAM] [HIGH.CONC] [I.LOOP.SINK] [POLY.DIFF]
[CLUST.TRANS] [DOSE.LOSS] [BDOSE.LOSS] [PDOSE.LOSS] [ASDOSE.LOSS]
[SiGECDF.MOD=<c>] [SiGECNI.MOD=<c>] [MODEL.SIGEC] [MIN.TEMP=<n>]
[IMPURITY|INTERST|VACANCY|OXIDANT|VELOCITY|TRAPS|PSI|PAC]
[REL.ERROR=<n>][ABS.ERROR=<n>]
[FE.RELERR=<n>][FE.ABSERR=<n>] [TD.RELERR=<n>][TD.ABSERR=<n>]
[ST.RELERR=<n>][ST.ABSERR=<n>] [FU.RELERR=<n>][FU.ABSERR=<n>]
[MIN.FILL] [MIN.FREQ=<n>] [GAUSS|CG] [BACK=<n>] [BLK.ITLIM=<n>]
[TIME.STR|ERROR|NEWTON] [DIAG|KNOT|FULL.FAC] [TRUNC.DEF=<n>]
[INIT.TIME=<n>][PDINIT.TIME] [T.DEFECT=<n>]
[OXIDE.GDT=<n>] [REDO.OXIDE=<n>] [TRBDF|FORMULA]
[ERFC|ERFG|ERF1|ERF2|COMPRESS|VISCOUS]
[LIFT.POLY][LIFT.OXIDE][LIFT.NITRID] [OX.THRESH=<n>] [SKIP.SIL]
[GRID.OXIDE=<n>][GRIDINIT.OX=<n>]
[GRID.SILICI=<n>][GRIDINIT.SI=<n>
[GLOOP.IMAX=<n>][GLOOP.EMIN=<n>][GLOOP.EMAX=<n>]
[OXIDE.EARLY=<n>][OXIDE.LATE=<n>][OXIDE.REL=<n>]
[OX.OBFIX=<n>][FILL] [PERIMETER=<n>] [MINTAREA.OX=<n>]
[ADAPT][DEPO.SMOOTH]ETCH.SMOOTH][DIFF.SMOOTH][STEP.SMOOTH]
[STRESS.HIST][DOPANT.STRESS][STRAIN.DIFF]
Description
This statement is used to set flags to select the various mathematical algorithms that will be
used to produce the simulation and to select the desired diffusion and oxidation model
complexity. Appropriate defaults for the numerical parameters are included in the athenamod
file so that you only need to specify the desired diffusion and oxidation model. The numerical
methods used in Athena for the solution of the diffusion equations are described in [7].

METHOD Statements
Parameters Related to DIFFUSION models

FERMI, TWO.DIM, STEADY, and FULL.CPL specify the type of diffusion equations to
be solved with particular regard to the point defect models (see Section 3.1 “Diffusion
Models”). The FERMI parameter specifies the defects are assumed to be a function of the
Fermi level only. The TWO.DIM parameter specifies that a full time dependent transient
simulation should be performed. The STEADY parameter specifies that the defects are assumed
to be in a steady state. The FULL.CPL parameter specifies that full coupling between defects
and dopants should be included. The default is FERMI. PD.FERMI is an alias for FERMI.
PD.TRANS is an alias for TWO.DIM. PD.FULL is an alias for FULL.CPL. PD.FERMI is an alias
for FERMI. PD.TRANS is an alias for TWO.DIM. PD.FULL is an alias for FULL.CPL.
Note: Chapter 2 “Tutorial” shows a complete description of the use of these diffusion models for typical
applications.
PLS, IC, VC, DDC, and SS specify advanced diffusion models (see Section 3.2 “Advanced
Diffusion Models”). If only the PLS parameter is specified, the classical dopant diffusion
model will be used. The parameters IC and VC will invoke additional interstitial and vacancy
clustering models. The DDC parameter switches on the dopant-defect clustering model. The
SS parameter will include the solid solubility model.
Note: These advanced diffusion models can be used only for boron, phosphorus and arsenic in silicon
technologies Also, these models cannot be used when oxidation or silicidation or both occur during the
simulated diffusion step. There are also some limitation on complexity of the 2D structures, which can be
handled by the solver. In most cases, when the solver cannot handle the structure, materials, impurities or
other conditions, it returns control to standard diffusion models.
CLUSTER.DAM specifies that the Stanford {311} cluster model is enabled, allowing a
scaled profile of {311} clusters during a subsequent implant. Only use this model when
FULL.CPL is also specified. It further causes a transient dissolution of the {311} clusters
leading to bulk interstitial injection. The CLUSTER statement is used to set parameters for this
model.
Note: For correct operation, set METHOD CLUSTER.DAM FULL.CPL before the IMPLANT statement that
generates the {311} clusters.
HIGH.CONC specifies that extra dopant concentration dependent point defect

recombination model is enabled. The IIFACTOR and IVFACTOR parameters on the
INTERSTITIAL command are used when METHOD HIGH.CONC is enabled.
I.LOOP.SINK specifies that a dislocation loop band can be specified during a subsequent
implant and that the loops may behave as an interstitial sink during diffusion. The
DISLOC.LOOP command is used to set parameters for this model.
POLY.DIFF specifies that the two-stream polysilicon diffusion model should be used. To
operate accurately, set this model before the deposition of the polysilicon material. See
Section 6.29 “IMPURITY” and Section 3.1.7 “Grain-based Polysilicon Diffusion Model” for
more information.
CLUST.TRANS enables the Transient Activation Model.

METHOD Statements
DOSE.LOSS specifies that Interface Trap Model for dose loss at silicon/oxide interface is
enabled for all three impurities B, P, and AS.
BDOSE.LOSS specifies that Interface Trap Model for dose loss at silicon/oxide interface is
enabled for boron.
PDOSE.LOSS specifies that Interface Trap Model for dose loss at silicon/oxide interface is
enabled for phosphorus.
ASDOSE.LOSS specifies that Interface Trap Model for dose loss at silicon/oxide interface is
enabled for arsenic.
MODEL.SIGEC enables special B diffusion model in SiGeC/SiGeC.
SIGECDF.MOD specifies the name of the C-Interpreter file for boron diffusivity model in
SiGe.
SIGECNI.MOD specifies the name of the C-Interpreter file instrinic carrier concentration
model used in boron diffusion mode in SiGe/SiGeC.
MIN.TEMP specifies the minimum temperature for which impurity diffusion is considered.
At temperatures below MIN.TEMP, the impurities are considered immobile. The default is
700C With caution, you can set this parameter to a lower value for certain diffusion steps.
Parameters Related to Numerics of Diffusion/Oxidation
IMPURITY specifies impurity for which one or several bound (tolerance) parameters will be
applied during diffusion/oxidation simulation (see Section 6.2.10 “Standard Impurities” for
the list of impurities).
INTERST, VACANCY, OXIDANT, VELOCITY, TRAPS, PSI and PAC specifies type of
solution for which one or several bound (tolerance) parameters will be applied during
diffusion/oxidation simulation.
REL.ERR indicates the precision with which the impurity solution must be solved. In
general, the actual error will be less than half of the indicated error. The defaults are 0.01 for
all impurities except the potential, which is solved to 0.001. If this parameter is used an
impurity should also be specified.
ABS.ERR specifies the error tolerance absolute value. For dopants, the absolute error
defaults to 1.0109. For defects, the absolute error defaults to 1.0105. For the potential, the
error defaults to 1.010-6. If this parameter is used, an impurity should also be specified.
FE.RELERR and FE.ABSERR specifies the relative error and absolute errors for the
FERMI model.
TD.RELERR and TD.ABSERR specifies the relative and absolute errors for the TWO.DIM
model.
ST.RELERR and ST.ABSERR specifies the relative and absolute errors for the STEADY
model.
FU.RELERR and FU.ABSERR specifies the relative and absolute errors for the FULL.CPL
model.

METHOD Statements
MIN.FILL and MIN.FREQ specify a minimum fill. It defaults to true. This is a highly
recommended option since it can reduce the matrix sizes by a factor of two or more and
operation speed is a function of the size of the matrix. MIN.FREQ is a parameter that controls
the frequency of the minimum fill reorderings. It is only partially implemented and has no
effect on the calculation.
GAUSS and CG specify the type of iteration performed on the linear system as a whole. CG
specifies that a conjugate residual should be used.
BACK specifies the number of back vectors that can be used in the CG outer iteration. The
default is three and the maximum possible value is six.
Note: A higher value of BACK will give faster convergence at the cost of more memory usage.
BLK.ITLIM is the maximum number of block iterations that can be taken. The block
iteration will finish at this point independent of convergence.
TIME.STE, ERROR, and NEWTON specify the frequency with which the matrix should
be factored. The default is TIME. The TIME parameter specifies that the matrix should be
factored twice per time step. This option takes advantage of the similarity in the matrix across
a time integration. The ERROR parameter indicates that the matrix should be factored
whenever the error in that block is decreasing. The NEWTON parameter forces factorization at
every NEWTON step.
DIAG, KNOT, and FULL.FAC specifies the amount of fill to be included in the factorization
of the matrix. FULL.FAC indicates that the entire amount of fill is to be computed. The DIAG
parameter indicates that only the diagonal blocks should be factored in the matrix. The KNOT
parameter is inactive. DIAG is the default parameter. Although under certain conditions (one-
dimensional stripes), FULL.FAC will perform better.
TRUNC.DEF specifies that defect concentrations that become negative due to numerical
difficulties be forced to a positive value.
Parameters Related to Timestep Control
INIT.TIME specifies the initial timestep value. The default is 0.1 seconds.
PDINIT.TIME specifies the initial time step for point defect diffusion. Point defects are held
fixed for the first timestep. The default is 10-5 seconds.
T.DEFECT specifies time in seconds for which point defect injection will be neglected
during an oxidation. The default is 5 seconds.
OXIDE.GDT limits the timestep during oxidation to a fraction (OXIDE.GDT) of the time
required to oxidize the thickness of one grid layer (GRID.OXIDE). The timestep may be
limited by oxidation and by diffusion, and the value of OXIDE.GDT will limit the timestep if it
is more stringent than the limits imposed by diffusion. OXIDE.GDT<<1 is recommended to
improve resolution of oxidizing diffusions. The default is 0.25.
REDO.OXIDE saves time by not computing the oxide flow field every time the diffusion
equation for impurities is solved. The REDO.OXIDE parameter specifies the percentage of the
time required to oxidize the thickness of one grid layer, which should elapse before resolving
the flow field. Usually REDO.OXIDE is much less than OXIDE.GDT, which is an upper bound
on how long the solution should wait. It is mainly intended to exclude solving oxidation at
each and every one of the first few millisecond time steps when defects are being tracked.

METHOD Statements
TRBDF and FORMULA specify the time integration method to be used. The TRBDF
parameter indicates that a combination trapezoidal rule/backward difference should be used.
The error is estimated using Milne’s device. The FORMULA method allows you to specify the
time step directly as a function of time (t), previous time step (dt), and grid time (gdt).
This option is primarily for testing. The TRBDF method is the default.
Parameters related to OXIDATION models
ERFC, ERFG, ERF1, ERF2, COMPRESS, and VISCOUS are oxidation models (see
Section 3.3 “Oxidation Models”). The ERFC parameter indicates that a simple error function
approximation to a bird’s beak shape should be used. The ERF1 and ERF2 models are analytic
approximations to the bird’s beak from the literature. The ERFG model chooses whichever of
ERF1 or ERF2 is most appropriate. All erf models are applicable only to the simplest case of
oxidation to the right of the mask edge. All relevant parameters in the OXIDE statement must
be explicitly specified when using any of the ERF* models. The COMPRESS model regards the
oxide as a compressible liquid. The VISCOUS model treats the oxide as an incompressible
viscous liquid. Oxide is actually believed to be incompressible, but the compressible model
runs faster. The default is the COMPRESS model.
Note: For Hints on the use of the different oxidation models, see Chapter 2 “Tutorial”.
Note: Use of the VERTICAL model is not recommended in Athena
LIFT.POLY, LIFT.OXIDE, and LIFT.NITRID specifies that the polysilicon, oxide, and
nitride materials can be lifted by oxidation or silicidation processes. These are true by
default, but you set them to false to eliminate the lifting portion of the calculation for
geometries where lifting is not expected to occur.
OX.THRESH specifies that the oxidation threshold model is enabled. This doesn’t allow
oxidation when the concentration of oxidant drops below a critical threshold value set by
MIN.OXIDANT on the OXIDE statement.
SKIP.SIL is a Boolean parameter which controls the computation of stress in silicon.
SKIP.SIL defaults to true. stress can only be computed when the VISCOUS oxide model
is used. The silicon substrate is treated as an elastic solid subject to the tensions generated by
the oxide flow. Indiscriminate use is not recommended. The silicon grid is usually much
larger than the oxide grid and stress computation is correspondingly more lengthy.
Parameters related to Grid Control during Oxidation
Many grid related problems during oxidation are related to the initial oxide deposition. See
Section 6.40 “OXIDE” for more about initial oxides.
GRID.OXIDE specifies the desired thickness, in microns, of grid layers to be added to the
growing oxide. It has an effect on time steps (refer to OXIDE.GDT). The default is 0.1 microns.
GRIDINIT.OX specifies the initial oxide grid spacing (in microns). The default is 0.1
microns.
GRID.SILICI specifies the maximum silicide grid spacing (in microns). The default is 0.1
microns.
GRIDINIT.SI specifies the initial silicide grid spacing (in microns). The default is 0.1
microns.

METHOD Statements
GLOOP.EMIN, GLOOP.EMAX, and GLOOP.IMAX controls loop detection during grid

manipulation. The default value is GLOOP.IMAX=170°. Loop detection checks for intrusions
and extrusions in the boundary. The intrusion-fixing algorithm is triggered by angles greater
than GLOOP.IMAX. A larger value means that more extreme intrusions can develop and
increases the possibility of a tangled grid. A smaller value leads to earlier intrusion-fixing; too
small a value will lead to inaccuracy due to premature intervention. Similar concerns apply to
the other parameters. The values are a compromise between safety and accuracy. The
extrusion-fixing algorithm is always triggered by angles greater than GLOOP.EMAX. It may be
triggered by lesser extrusions, anything greater than GLOOP.EMIN, if the extrusion is a single-
triangle error in the boundary. The default value is GLOOP.EMIN=130°. Neither of these
parameters should be less than 90 because the rectangular edges of the simulation space
would be smoothed.
OXIDE.EARLY, OXIDE.LATE, and OXIDE.REL should not normally be modified. They
relate to internal numerical mechanisms and are described here only for the sake of
completeness. A node whose spacing decreases proportionally by more than OXIDE.LATE is
marked for removal. Also, if any nodes are removed, then all nodes greater than
OXIDE.EARLY are also removed. For earlier node removal (fewer obtuse triangles), try
OXIDE.LATE=0.3 and OXIDE.EARLY=0.1. Though not logical, it is harmless for
OXIDE.EARLY to be greater than OXIDE.LATE. The OXIDE.REL parameter is the percentage
error in velocities for the non-linear viscous model. The default is 1.010-2 (that is, a 1.0
percent error). OXIDE.REL can be increased for a faster solution.
OX.OBFIX specifies the cosine squared of the worst angle allowed during oxidation.
FILL and PERIMETER specify which action to apply to voids that may form during
oxidation. FILL specifies that you must fill the voids with oxide materials. Default is false.
PERIMETER specifies the maximum perimeter of the voids to fill. Default is 0.2 microns.
MINTAREA.OX specifies the minimum area of triangles generated during oxidation,
unitless. The default is zero. If MINTAREA.OX is greater than zero, then the oxide triangles
with area smaller than MINTAREA.OX is multiplied by the total area will be marked for
elimination. The typical safe value for MINTAREA.OX is around 1e-06. The use of
MINTAREA.OX is recommended mostly when small (< 0.01 microns) values for GRID.OX is
used for non-planar complex oxidations.
Parameters related to Grid Control during ETCH
ETCH.EPS sets a tolerance on the grid movement during ETCH statements. This parameter is
defined in relative units. The default is 10-6 that corresponds to about 10 Angstroms.
Reducing this number will allow sub-10A etches to be exact. But the possibility of small
triangles being created during etches is high if the parameter is set too low. This parameter
should not be set to zero.
Parameters used in the Adaptive Meshing Module
ADAPT specifies that the adaptive meshing should be performed on the IMPLANT, DIFFUSE,
or EPITAXY statements (the default is false).
DEPO.SMOOTH specifies that the mesh smoothing should be performed after each
DEPOSIT statement.
ETCH.SMOOTH specifies that the mesh smoothing should be performed after each ETCH
statement.

METHOD Statements
DIFF.SMOOTH specifies that the mesh smoothing should be performed after each DIFFUSE
statement.
STEP.SMOOTH specifies that the mesh smoothing should be performed after each time step
on each DIFFUSE statement.
Parameters related to Stress Simulation
STRESS.HIST specifies that stresses to be calculated during etching, deposition, diffusion,
and epitaxy process steps. Default is False.
DOPANT.STRESS specifies that dopant-induced stress simulation model should be used
either with the STRESS statement or when STRESS.HIST is specified. Default is False.
STRAIN.DIFF specifies that stress induced diffusion model should be used when stresses
have been calculated previously (i.e., STRESS.HIST is specified).
Example setting tolerances
The following statement specifies that the arsenic equation should be solved with a relative
error of 1% and concentrations below 1 x 109 can be ignored.
METHOD ARSEN REL.ERR=0.01 ABS.ERR=1.0E9
Example setting numerical techniques
The following statement specifies that minimum fill reordering should be done and that the
entire system should be solved using a conjugate residual technique with three back vectors.
The initial time step should be 0.1 seconds and time should be integrated using the TRBDF
parameter. The FERMI model should be used for diffusion and the COMPRESS model for the
oxide growth.
METHOD MIN.FILL CG BACK=3 INIT.TI=0.1 TRBDF FERMI COMPRESS
Example setting diffusion model for power devices
The following step specifies that a simple diffusion model should be used appropriate for
power electronic devices.
METHOD POWER
DIFFUSION TEMP=1000 TIME=300 NITROGEN
Example setting diffusion models for RTA
The following statement invokes all {311} cluster models for RTA simulation. It must be set
before the IMPLANT statement that generates the cluster damage.
METHOD NEWTON FULL.CPL CLUSTER.DAM I.LOOP.SINK HIGH.CONC BACK=6
IMPLANT ....
DIFFUSE ....

MOMENTS Statements
6.37 MOMENTS
MOMENTS specifies tables and spacial moments used in analytical implant models.
Syntax
MOMENTS
[SVDP_TABLES | STD_TABLES | USER_STDT] [USER_TABLE=<c>]
[MATERIAL] [I.IMPUTITY] [DOSE=<n>] [TMA_TABLES] [TBL.PATH=<c>]
[ENERGY=<n>]
[RANGE=<n>] [STD.DEV=<n>] [GAMMA=<n>] [KURTOSIS=<n>
[LSTD.DEV][LGAMMA][LKURTOSIS] [SKEWXY]
[SRANGE=<n>][SSTD.DEV=<n>][SGAMMA=<n>][SKURTOSIS=<n>]
[LSSTD.DEV][SSKEWXY][SKURTXY][SKURTT][DRATIO=<n>]
[IGNORE_MOM]
Description
Parameters Used to Select Moment Tables
SVDP_TABLES specifies that the SIMS Verified Dual Pearson (SVDP) moments tables will
be used with dual Pearson implant model. Default is true (See Section 3.5.1 “Analytic
Implant Models”).
STD_TABLES specifies that SVDP_TABLES are ignored and standard tables are used with the
subsequent implant statements.
USER_STDT specifies the user-defined moments file (see the USER_TABLE parameter) will
be used with standard format. You can find a template for the user-defined moments file in
<install.area>/lib/athena/<version>/common/userimp.
USER_TABLE=<c> specifies the file that contains user-defined look-up implant parameter
tables.
TMA_TABLES specifies that moments from T-SUPREM4 format implant tables will be
used.
TBL_PATH specifies the full path to a user-defined T-SUPREM4 implant tables file. By
default, the T-SUPREM4 tables file is to be located in the [install]/lib/athena/
<version>/common/implant_tables folders.
Implant Definition Parameters
MATERIAL specifies the material for which the implant moments are set (see Section 6.2.9
I.IMPURITY specifies the implanted impurity for which the moments are set (see Section
6.2.10 “Standard Impurities” for the list of impurities). I.BF2 can be also specified.
DOSE is an incident ion dose (/cm2).
ENERGY sets the incident ion energy (keV).

MOMENTS Statements
Parameters Used for Specification of Spacial Moments

RANGE (RP) specifies the projected range. Units are microns.
STD.DEV (DRP) specifies the standard deviation Units are microns.
GAMMA (SKEWNESS) specifies the third moment. Default is 0.0.
KURTOSIS specifies the fourth moment. Default is 3.01.
LSTD.DEV(LDRP) specifies the lateral standard deviation Units are microns.
SKEWXY specifies the mixed third moment.
KURTXY specifies the lateral mixed fourth moment.
KURTT specifies the lateral fourth moment. Default is 3.0.
SRANGE (SRP) specifies the projected range for second Pearson Units are microns.
SSTD.DEV(SDRP) specifies the standard deviation for second Pearson Units are microns.
SGAMMA (SSKEW) specifies the third moment for second Pearson function. Default is 0.0.
SKURTOSIS specifies the fourth moment for second Pearson function. Default is 3.01.
LSSTD.DEV(LSDRP) specifies the lateral standard deviation for second Pearson Units are
microns.
SSKEWXY specifies the mixed third moment for second Pearson. Default is 0.0.
SKURTXY specifies the mixed fourth moment for second Pearson. Default is 0.0
SKURTT specifies the lateral fourth moment for second Pearson. Default is 3.0.
DRATIO specifies the dose ratio R in the double Pearson function. Default is 0.9
Reset Parameter
IGNORE_MOM specifies that all previous MOMENTS statements will be ignored.
Examples
The MOMENTS statement is used to define user moments through a convenient command
language. The following example sets the moments for boron implantation into the user-
defined material SAPPHIRE.
MOMENTS MATERIAL=SAPPHIRE I.BORON DOSE=1.6e12 ENERGY=25 \
RANGE=0.098 STD.DEV=0.045 GAMMA=-0.04 KURTOSIS=3.5
For more examples, see IMPLANT.

OPTICAL Statements
6.38 OPTICAL
OPTICAL sets the optical parameters of materials for Optolith.
Syntax
OPTICAL
[MATERIAL][NAME.RESIST=<c>]
[LAMBDA=<n>][I.LINE|G.LINE|H.LINE|DUV.LINE]
[REFRAC.REAL=<n>][REFRAC.IMAG=<n>]
[DELTA.REAL=<n>][DELTA.IMAG=<n>]
Description
This command sets the optical parameters, reflective index, and extinction coefficient or
REFRAC.REAL and REFRAC.IMAG for each material at a particular wavelength. If photoresist
is used, NAME.RESIST must also be specified.
MATERIAL specifies the material for which the optical parameters to be set (see Section
6.2.9 “Standard and User-Defined Materials” for the list of materials)
NAME.RESISTspecifies the name of the photoresist.
I.LINE, G.LINE, H.LINE, DUV.LINE, and LAMBDA specifies the line or the wavelength.
Units for LAMBDA are microns. WAVELENGTH is an alias for LAMBDA.
REFRAC.REAL specifies the real component of the refractive index.
REFRAC.IMAG specifies the imaginary component of the refractive index.
DELTA.REAL specifies the difference between the real components of the refractive index
for completely exposed and unexposed resist. This value is used when dose effect on the
refractive index is simulated.
DELTA.IMAG specifies the difference between the imaginary components of the refractive
index for completely exposed and unexposed resist. This value is used when dose effect on
the refractive index is simulated.
Examples
The OPTICAL statement is used to load refractive index values into Athena for each
wavelength. The following shows a typical statement.
OPTICAL SILICON WAVELENGTH=.365 REFRAC.REAL=4.5 REFRAC.IMAG=5.2
You can enter user-defined materials in the following format.
OPTICAL MATERIAL=XXX WAVELENGTH=.365 REFRAC.REAL=1.4 \
REFRAC.IMAG=.3
For more examples, see EXPOSE and IMAGE.

OPTION Statements
6.39 OPTION
OPTION specifies the level of run-time output.
Syntax
OPTION
[QUIET|NORMAL|VERBOSE|DEBUG|WARNING]
Description
This statement specifies the level of information sent to the TTY Terminal Window of
DeckBuild.
QUIET, NORMAL, VERBOSE, DEBUG, and WARNING determines the amount of
information that is output about errors, CPU times, and behavior of the algorithms. The
default is QUIET. The VERBOSE and DEBUG modes are intended mainly for debugging by
developers.
Examples
The following statement sets the routine output to include more information.
OPTION NORMAL

OXIDE Statements
6.40 OXIDE
OXIDE specifies coefficients for use during oxidation steps. AMBIENT is a synonym for OXIDE.
Syntax
OXIDE
DRY02|WET02 [ORIENT=<n>]
[LIN.L.0=<n>][LIN.L.E=<n>] [LIN.H.0=<n>][LIN.H.E=<n>
[L.BREAK=<n>][L.PDEP=<n>]
[PAR.L.0=<n>][PAR.L.E=<n>] [PAR.H.0=<n>][PAR.H.E=<n>]
[P.BREAK=<n>][P.PDEP=<n>]
[ORI.DEP][ORI.FAC=<n>]
[HCL.PC=<n>][HCLT=<n>][HCLP=<n>][HCL.PAR=<n>][HCL.LIN=<n>]
[THINOX.0=<n>][THINOX.E=<n>][THINOX.L=<n>][THINOX.P=<n>]
[BAF.DEP][BAF.EBK=<n>][BAF.PE=<n>][BAF.PPE=<n>]
[BAF.NE=<n>][BAF.NNE=<n>][BAF.K0=<n>][BAF.KE=<n>]
[STRESS.DEP][VC=<n>][VR=<n>][VD=<n>][VT=<n>][DLIM=<n>]
[MATERIAL][/MATERIAL]
[DIFF.0=<n>][DIFF.E=<n>][SEG.0=<n>][SEG.E=<n>]
[TRN.0=<n>][TRN.E=<n>]
[HENRY.COEFF=<n>] [THETA=<n>] [ALPHA=<n>] [MIN.OXIDANT=<n>]
[INITIAL=<n>] [SPLIT.ANGLE=<n>]
[SPREAD=<n>][MASK.EDGE=<n>][NIT.THICK=<n>][ERF.Q=<n>]
[ERF.DELTA=<n>][ERF.LBB=<n>][ERF.H=<n>]
Description
All parameters relating to oxidation are specified in this statement. Oxidation models are
specified in the METHOD statement. All oxidation models are described in Section 3.3
“Oxidation Models”. To properly set values for most coefficients, you need to know whether
to use wet or dry oxidation and to know the substrate orientation.
Note: If a required parameter is omitted (e.g., orientation when a linear rate coefficient is being specified), then the
statement is ignored without warning.
Oxide Growth Rate Parameters

DRYO2, WETO2 specifies the type of oxidation to which specified coefficients apply.
Required for everything except for one-dimensional coefficients and the volume ratio.
ORIENT is the substrate orientation the coefficients specified apply to the required for
orientation factor (see ORI.FAC) and thin oxide coefficients. Only 100, 110, and 111 are
recognized. The default is 100.

OXIDE Statements
LIN.L.0, LIN.L.E, LIN.H.0, LIN.H.E, L.BREAK, and L.PDEP specifies the linear rate
coefficients (B/A). A doubly activated Arrhenius model is assumed. L.BREAK is the
temperature breakpoint between the lower and higher ranges in degrees Celsius. LIN.L.0 is
the pre-exponential factor in microns/min. LIN.L.E is the activation energy in eV for the low
temperature range. LIN.H.0 and LIN.H.E are the corresponding high temperature numbers.
L.PDEP is the exponent of the pressure dependence. The value given is taken to apply to
<111> orientation and later adjusted by ORI.FAC according to the substrate orientation
present.
PAR.L.0, PAR.L.E, PAR.H.0, PAR.H.E, P.BREAK and P.PDEP specifies the parabolic
rate coefficients (B).
ORI.FAC is the ratio of B/A on the specified orientation to the orientation.
ORI.DEP specifies whether the local orientation at each point on the surface should be used
to calculate B/A. The default is true. If it is false, the substrate orientation is used at all points.
THINOX.0, THINOX.E, and THINOX.L specifies coefficients for the thin oxide model
proposed by Massoud [14]. THINOX.0 is the pre-exponential factor in microns/min,
THINOX.E is the activation energy in eV, and THINOX.L is the characteristic length in
microns.
THINOX.P is the thin oxide model pressure dependence.
HCL.PC, HCLT, HCLP, HCL.PAR, and HCL.LIN is where the numerical parameter,
HCL.PC, is the percentage of HCl in the gas stream. It defaults to 0. The HCl dependence of
the linear and parabolic coefficients is obtained from a look-up table specified in the model
file. The table rows are indexed by HCl percentage. Specify the row entries with the
parameter HCLP, which is an array of numerical values surrounded by double quotes and
separated by spaces or commas. The columns are indexed by temperature. Specify the
column entries with the parameter HCLT, which is an array of numerical values surrounded by
double quotes and separated by spaces or commas. Specify the dependence of B/A with the
parameter HCL.LIN, which is an array of numerical values surrounded by double quotes and
separated by spaces or commas. The number of entries in HCL.LIN must be the product of the
number of entries in HCLP and HCLT. Specify the dependence of B with the parameter
HCL.PAR, which is an array of numerical values surrounded by double quotes and separated
by spaces or commas. The number of entries in HCL.PAR must be the product of the number
of entries in HCLP and HCLT.
BAF.DEP, BAF.EBK, BAF.PE, BAF.PPE, BAF.NE, BAF.NNE, BAF.K0, and BAF.KE
relates to the doping dependence of the oxidation rate. The doping dependence is activated
when BAF.DEP is true (default). MATERIAL1 must be specified with these parameters (only
SILICON and POLYSILICON make sense here).
STRESS.DEP, VC, VR, VD, VT, and DLIM controls the stress dependence of oxidation,
which is only calculated under the VISCOUS model. STRESS.DEP turns on the dependence. VC
is the activation volume of viscosity. VR is the activation volume of the reaction rate with
respect to normal stress. VT is the activation volume of the reaction rate with respect to
tangential stress. VD is the activation volume of oxidant diffusion with respect to pressure.
DLIM is the maximum increase of diffusion permitted under tensile stress.
MATERIAL specifies MATERIAL1 for which parameters to be set (see Section 6.2.9
/MATERIAL specifies MATERIAL2 for which parameters to be set (see Section 6.2.9

OXIDE Statements
DIFF.0, DIFF.E, SEG.0, SEG.E, TRN.0, and TRN.E specifies the diffusion coefficients of
oxidant in MATERIAL1 and the boundary coefficients (“transport” and “segregation”) from
MATERIAL1 to MATERIAL2 as defined above. DIFF.0 is the diffusivity pre-exponential factor
in cm2/sec. DIFF.E is the energy in eV. The transport coefficient represents the gas-phase
mass transfer coefficient in terms of concentrations in the solid at the oxide-gas interface, the
chemical surface-reaction rate constant at the oxide-silicon surface, and a regular diffusive
transport coefficient at other interfaces. The segregation coefficient is 1 at the oxide-gas
interface, it is infinity at the oxide-silicon interface, and is a regular segregation coefficient at
other interfaces.
Note: Oxidant in materials other than oxide is allowed to diffuse and segregate but its concentration is then
ignored (for instance, no oxynitridation).
HENRY.COEFF (Henry’s coefficient) is the solubility of oxidant in MATERIAL1 measured in

cubic centimeters (cm3) at one atmosphere. THETA is the number of oxygen atoms
incorporated in a cubic centimeter of oxide. In the case of dry oxidation, it is equal to THETA
and in the case of wet oxidation, it is equal to 2 THETA. Usually, the Deal-Grove coefficients
should be changed instead of HENRY.COEFF.
THETA specifies the concentration of O2 atoms incorporated in the material. Units are cm-3.
ALPHA specifies the volume expansion ratio between MATERIAL1 and MATERIAL2. Only
SILICON, POLYSILICON and OXIDE make sense here.
MIN.OXIDANT specifies the minimum oxidant concentration for oxidation to occur. Units
are cm-3. This parameter is active only if METHOD OX.THRESH is used.
Parameters Related to Grid Control
INITIAL specifies the thickness of the native (initial) oxide at the start of oxidation step. If
any oxidizable surface of the structure is bare, an oxide layer of this thickness is deposited
before oxidation begins. Units are microns. Default is 0.002.
Note: The oxidation algorithm requires selective deposition of a native oxide onto all exposed silicon or polysilicon
areas prior to oxidation. Grid problems can result in complex structures. To resolve these problems, adjust
INITIAL or use the DEPOSIT statement to create the native oxide.
SPLIT.ANGLE governs the minimum angle at which the oxide will split open one more grid
spacing when oxidizing at a triple point (i.e., where silicon, oxide, and nitride coincide
together at a point). The default for the split angle is 22.5°. The SPLIT.ANGLE parameter for
triple point oxidation is material dependent. Specify the oxidizing MATERIAL1 without a “/”
and MATERIAL2 with a “/” using the following format: OXIDE SPLIT.ANGLE=35 SILICON
/NITRIDE. There are only three possible combinations and they are SILICON/NITRIDE,
SILICON/POLYSILICO, and POLYSILION/NITRIDE. You can use this to control lateral
encroachment during oxidation.

OXIDE Statements
Parameters of the Analytical Oxidation Models (ERF*)

SPREAD and MASK.EDGE are used only in the error-function approximation to a bird’s
beak shape. SPREAD is the relative lateral to vertical extension, which defaults to 1. The fitting
parameter makes the “erfc” bird’s beak look realistic. MASK.EDGE is the position of the mask
edge in microns and defaults to negative infinity. Oxide grows to the right of the mask edge.
ERF.Q and ERF.DELTA are the DELTA and Q parameters for the “erfg” model. Normally,
you don’t need to change them but they are available if necessary.
ERF.LBB is the length of the bird’s beak and applies to the “erfg” model only. It can be
specified as an expression in Eox (the field oxide thickness (m)), eox (the pad oxide
thickness (m)), Tox (the oxidation temperature (Kelvin)), and en (the nitride thickness, m).
The published expression can be found in the models file. Specifying ERF.LBB=EOX, for
instance, would give a lateral spread equal to the field thickness, similar to the Hee-Gook Lee
model with a spread of one.
ERF.H is the ratio of the nitride lifting to the field oxide thickness. It corresponds to the
Guillemot “H” parameter except that it is normalized to the field oxide thickness. It is
specified as an expression of Eox, eox, Tox, en.
NIT.THICK specifies the nitride thickness to substitute for the parameter EN.
Note: The ERFG model uses both oxide and nitride thickness. These values are not inferred from the structure.
Instead, the nitride thickness is user-specified in the OXIDE statement and the oxide thickness is computed
by adding the total oxide grown and the initial user-specified oxide thickness. If the structure has more than
20 angstroms (the default) of native oxide on it when diffusion begins, then thickness must be specified.
Beware of this when continuing a diffusion by any means (e.g., after reading in a previous structure).
Examples
The following modifies the parabolic oxidation rates for {100} silicon in a dry oxygen
ambient.
OXIDE DRY ORI=100 PAR.L.0=283.333 PAR.L.E=1.17
Note: If a required parameter is omitted (e.g., orientation when a linear rate coefficient is being specified), then the
statement is ignored without warning.
The following set the native oxide thickness at 10 Angstroms.

OXIDE INITIAL=0.001
The following defines that stress-dependent oxidation rates will be used with the viscous
oxidation model.
METHOD VISCOUS
OXIDE STRESS.DEP=t
For more examples, see DIFFUSE and METHOD.

POLISH Statements
6.41 POLISH
POLISH runs the chemical mechanical polishing (CMP) module.
Syntax
POLISH
MACHINE=<c>[TIME=<n>][HOURS|MINUTES|SECONDS]
[DX.MULT=<n>][DT.FACT=<n>][DT.MAX=<n>]
Description
This statement executes the chemical mechanical polishing module of Elite. The POLISH
statement must be preceded by a RATE.POLISH statement to define the polishing machine.
MACHINE specifies the name of the polish machine.
TIME specifies the time the machine is to be run.
HOURS, MINUTES, and SECONDS specifies the units of the TIME parameter.
DX.MULT is the accuracy multiplier for Elite polishes. The discretization size used for the
polish calculation will be multiplied by DX.MULT. For improved accuracy, decrease the value
of DX.MULT. For improved speed, increase the value of DX.MULT.
DT.FACT controls the timestep size. By default, the movement of a string node is limited to
less than or equal to one quarter of the median segment length. This is a good compromise
between simulation speed and the danger of loop formation. The optimization factor DT.FACT
must not exceed 0.5 but can decrease if necessary for more accuracy.
DT.MAX is used with Elite type polish calculations. By default, the upper limit for the micro
timestep DT.MAX is one tenth of the total etch time specified. This is a good compromise
between calculation accuracy and calculation time. But, sometimes it is useful to adapt this
value to the specific simulation problem. Allowing the time steps to become greater gives a
higher simulation speed but the accuracy may suffer. For smaller time steps, the simulation
speed will decrease but the accuracy may be greater.
Examples
The following statements illustrate running the chemical mechanical polish module. A
RATE.POLISH statement sets the values for the polish model and must precede the POLISH
statement.
RATE.POLISH OXIDE MACHINE=cmp u.s MAX.HARD=0.15 MIN.HARD=0.03 \
ISOTROPIC=0.001
POLISH MACHINE=cmp TIME=5 MIN
For more examples, see RATE.POLISH and ETCH.

PRINT.1D Statements
6.42 PRINT.1D
PRINT.1D prints values along a one–dimensional cross section or an material interface.
Note: Use of this statement is not recommended. All functions are available using the EXTRACT command within
DeckBuild.
Syntax
PRINT.1D
[X.VALUE=<n>|Y.VALUE=<n>]
[MATERIAL][/MATERIAL] [ARCLENGTH][LAYERS]
[X.MIN=<n>][X.MAX=<n>][FORMAT=<c>]
Description
This command prints the values along cross sections through the device. You can also use to
integrate along a specified line. The value printed is the value that has been selected (see
Section 6.55 “SELECT”).
X.VALUE specifies the x coordinate of a vertical cross-section along which the selected
values are to be printed. Units are microns.
Y.VALUE specifies the y coordinate of a vertical cross-section along which the selected
values are to be printed. Units are microns.
MATERIAL specifies the selected values in the named material at the interface with another
material named by /MATERIAL are to be printed (see Section 6.2.9 “Standard and User-
ARCLENGTH is only relevant when printing along an interface. If ARCLENGTH is chosen,
the printed ordinate is the arclength, measured in microns, along the boundary from the left
most point of the curve. If ARCLENGTH is not chosen, the x value of the interface location is
printed. The coordinate of the left most point is equal to its x coordinate in the mesh layers.
LAYERS instructs the selected print variable to integrate in each material it crosses. The
integrated value and material width is reported. Zero crossings of the variable are treated the
same as material interfaces.
X.MIN and X.MAX specify the minimum and maximum positions along the cross-section to
be printed.
FORMAT changes the print format for the variable, using standard format expressions of the
C-language. Default is “%-16e”.
Examples
The following statement prints the selected value at x equal to one micron between the top of
the mesh and the 3.0 micron point.
PRINT.1D X.VAL=1.0 X.MAX=3.0
The following prints the selected variable along the silicon side of the silicon/oxide interface.
PRINT.1D SILICON /OXIDE
For more examples, see SELECT and PRINTF.

PRINTF Statements
6.43 PRINTF
PRINTF is a string printer and desk calculator.
Note: Functions of this statement have been replaced by the EXTRACT statement
Description
The ECHO statement merely prints the string given to it. This is useful for placing comments in
an output file. The statement attempts to parse the string to a legal real number if possible. It
has a regular expression parser built-in. This allows ECHO to be used as a desk calculator.
Examples
The following command will send the string “Athena Is My Favorite Process
Simulator” to standard output.
ECHO Athena Is My Favorite Process Simulator
The following command will print 4096.
ECHO (2^3^4)
The following command will print 8.373, which is the solution to the arithmetic expression.
ECHO ( 15.0 - 12.0 * EXP( 4.0 - 2.0 / 6.0 ) )

PROFILE Statements
6.44 PROFILE
PROFILE reads a 1D doping profile into Athena.
Syntax
PROFILE
[INFILE=<c>] [MASTER | XY.SIMS | EXTRACT.SIMS]
[IMPURITY|INTERST|VACANCY|CLUSTER.DAM|DIS.LOOP]
[LAYER1.DIV=<n>] [LAYER2.DIV=<n>]....[LAYER20.DIV=<n>)
Description
This statement can be used to load a 1D stream of doping data into an Athena structure. The
data might come from a Secondary Ion Mass Spectroscopy (SIMS) profile, from the “SIMS”
profile extracted by the EXTRACT command of DeckBuild, or from a 1D simulation in
SSuprem3. Data is applied in 1D across the width of the mesh for subsequent 2D simulation.
INFILE specifies the name of the profile data file or Standard Structure File to be loaded.
MASTER or SSF indicates that the file to be loaded is an Silvaco Standard Format file. Files
generated by SSuprem3 are in this format.
XY.SIMS specifies that the profile file is in a SIMS format described in example below. This
is the default.
EXTRACT.SIMS specifies that the profile file is obtained by EXTRACT command of
DeckBuild (see DeckBuild User’s Manual for more information).
IMPURITY specifies the impurity type for profile data file. Corresponding active impurity
will be also added. See Section 6.2.10 “Standard Impurities” for the list of impurities.
INTERST, VACANCY, CLUSTER.DAM and DIS.LOOP specify that profile data file
includes a profile of interstitials, vacancies, {311} clusters or dislocation loops,
correspondingly.
LAYER1.DIV, LAYER2.DIV,..., LAYER20.DIV specifies the number of subdivisions for
each layer when loading SSuprem3 Structure files.
Examples
An example of a PROFILE statement is given below.
PROFILE INF=BORON.SIMS BORON
In this case, the PROFILE statement specifies that only boron information will be added to
the current working silicon structure. The data file BORON.SIMS should be in the following
format.
#THIS IS SIMS DATA
0.01 1E15
0.02 1.1E15
0.04 1.3E15
0.06 1.5E15
0.1 1.7E15
0.2 1.9E15

PROFILE Statements
0.4 2.6E15
. . .
In the following example, the PROFILE statement will read in a 1D Silvaco’s standard format
(SSF) file. All doping and layer information will be preserved. This allows you to start a
simulation in, for example, SSuprem3 and finish it in Athena. The Athena grid must be set up
in the conventional manner first. The PROFILE statement will then include any overlying
layers that may have been deposited or grown in creating the SSuprem3 structure. The value
LAYER<n>.DIV controls the number of grid points in the overlying layers. The default grid
spacing generated for overlying layers is 0.05 m.
PROFILE MASTER INF=SSUPREM3.STR LAYER1.DIV=3 LAYER2.DIV=6
The first layer above the substrate will have 3 vertical grid spacings and the second layer
above the substrate will have 6 vertical grid spacings. The file SSUPREM3.STR must be a SSF
file.
The following is a list of special cases and their solutions.
• If a SSuprem3 structure is deeper than the Athena structure, the PROFILE statement will extend the
value of the bottom grid point.
• If a SSuprem3 structure is shallower than the Athena structure, the PROFILE statement will clip the
Athena profile.
• Loading a SSF file works only with a bare silicon wafer as a starting point. If you try to use some
other material for a substrate, the results are unreliable and unpredictable.
• Any concentrations of dopant initialized in Athena will be overwritten if a PROFILE statement is
used to load a SSF file.

PROJECTION Statements
6.45 PROJECTION
PROJECTION specifies the basic optical projection parameters for Optolith.
Syntax
PROJECTION
[NA=<n>][FLARE=<n>]
Description
This statement specifies the numerical aperture NA, the defocus distance, and the possible
flare in the optical or resist systems.
NA is the numerical aperture of the optical projection system.
FLARE is the amount of flare for the particular imaging problem. FLARE must be expressed
in percentages.
Examples
The following statement sets the numerical aperture and flare value for the projection system.
PROJECTION NA=.5 FLARE=2
For more examples, see IMAGE, ILLUMINATION, ILLUM.FILTER, PUPIL.FILTER, LAYOUT,
and ABERRATION.

PUPIL.FILTER Statements
6.46 PUPIL.FILTER
PUPIL.FILTER specifies the projection pupil type and filtering for Optolith.
Syntax
PUPIL.FILTER
CIRCLE|SQUARE|GAUSSIAN|ANTIGAUSS
[GAMMA=<n>][IN.RADIUS=<n>][OUT.RADIUS=<n>][PHASE=<n>]
[TRANSMIT=<n>][CLEAR.FIL]
Description
This command allows you to specify four different pupil types and allows spatial filtering in
the Fourier transform plane.
CIRCLE, SQUARE, GAUSSIAN, and ANTIGAUSS defines or changes the shape of the
exit pupil of the projection system. The shape of the pupil must be entered as a character
string.
GAMMA defines or changes the GAMMA value for GAUSSIAN and ANTIGAUSS pupil
transmittance. GAMMA is a parameter that defines the truncation of the GAUSSIAN by the pupil.
In the limit of GAMMA0, the pupil transmittance will be uniform.
IN.RADIUS and OUT.RADIUS defines or changes the intensity transmittance and phase
transmittance of an annular zone inside the exit pupil or either the illumination or the
projection system. This qualifier is used to simulate spatial filtering techniques. IN.RADIUS
and OUT.RADIUS are used to define an annular zone in the exit pupil having the pupil
transmittance equal to TRANSMIT and producing the phase angle equal to PHASE. Radius
values are specified in fractions of unity and phase is specified in degrees. Note that the
annular zones should not overlap. The outer radius of an inner zone must be smaller than the
inner radius of an outer zone. The shape of the annular zone is specified by the shape
parameter above. The maximum radius is one.
PHASE specifies the phase shift in degrees produced by the pupil filter-180  PHASE
180).
TRANSMIT specifies the pupil transmittance caused by the pupil filter.
CLEAR.FIL resets the projection filter list.
Examples
This set of commands defines a square aperture in the projection pupil that is opaque over a
square annular region.
PUPIL.FILTER SQUARE
PUPIL.FILTER IN.RADIUS=.1 OUT.RADIUS=.2 PHASE=0 TRANSMIT=0
For more examples, see IMAGE, ILLUMINATION, PROJECTION, ILLUM.FILTER, LAYOUT, and
ABERRATION.

QUIT Statements
6.47 QUIT
QUIT terminates execution of Athena. The EXIT, STOP, and BYE statements are synonyms of
the QUIT statement.
Syntax
QUIT
Description
All statements after a QUIT statement will not be checked or executed.

RATE.DEPO Statements
6.48 RATE.DEPO
RATE.DEPO specifies the deposit rates of a machine, which is used in a subsequent DEPOSIT
statement.
Syntax
RATE.DEPO
MACHINE=<c> MATERIAL | NAME.RESIST=<c>]
CONICAL|CVD|PLANETAR|UNIDIRECT|DUALDIRECT|
HEMISPHERIC|MONTE1|MONTE2|CUSTOM1|CUSTOM2
DEP.RATE=<n> [INFILE=<c>]
[A.H|A.M|A.S|U.S|U.M|U.H|N.M]
[STEP.COV=<n>] [ANGLE1=<n>][ANGLE2=<n>][ANGLE3=<n>]
[C.AXIS=<n>] [P.AXIS=<n>] [DIST.PL=<n>]
[SIGMA.DEP=<n>] [SIGMA.0][SIGMA.E]
[SMOOTH.WIN=<n>][SMOOTH.STEP=<n>][MCSEED=<n>][STICK.COEF=<n>
Description
This statement is used to define deposition parameters and the machine name for one of ten
deposition models available in Elite.
MACHINE specifies the machine name for the RATE.DEPO statement.
MATERIAL specifies material to be deposited by the deposit machine (see Section 6.2.9
NAME.RESIST specifies the name of photoresist to be deposited.
CONICAL, CVD, PLANETARY, UNIDIRECT, DUALDIRECT, HEMISPHERIC,
MONTE1, MONTE2, CUSTOM1 and CUSTOM2 specify a particular model for the
machine definition
DEP.RATE specifies the deposition rate used by the models CONICAL, CVD, UNIDIREC,
DUALDIREC, HEMISPHE, PLANETAR, MONTE1, and MONTE2. DEP.RATE is a rate multiplier for
the CUSTOM1 and CUSTOM2 models.
INFILE specifies the name of a file containing angle and deposition rate information for the
CUSTOM model.
A.H, A.M, A.S, U.H, U.M, U.S, and N.M specify that the deposition rate DEP.RATE is in
Angstroms per hour, Angstroms per minute, Angstroms per second, microns per hour,
microns per minute, microns per second, and nanometers per minute, respectively. Default is
A.S.
STEP.COV specifies the step coverage used by the model CVD.
ANGLE1 specifies the angle parameter used by the models HEMISPHE, CONICAL, UNIDIREC,
DUALDIREC, and PLANETAR.

RATE.DEPO Statements
ANGLE2 specifies the angle parameter used by the models DUALDIREC, PLANETAR, and
HEMISPHE.
ANGLE3 specifies the angle parameter used by the model PLANETAR.
C.AXIS specifies the central axis length used by the models CONICAL and PLANETAR.
P.AXIS specifies the planetary axis length used by the models PLANETAR and CONICAL.
DIST.PL specifies the distance from wafer to planetary axis used by the model PLANETAR.
SIGMA.DEP specifies the surface diffusion parameter used by the models UNIDIRECT,
DUALDIRECT, HEMISPHERIC, PLANETARY, CONICAL, MONTE1, and MONTE2.
SIGMA.0 and SIGMA.E specify pre-exponential coefficient and activation energy of
temperature dependent surface diffusion. The temperature is specified on the DEPOSIT
command.
SMOOTH.WIN and SMOOTH.STEP specifies a window size in microns and a number of
smoothing passes for the simple geometric deposit smoothing algorithm.
MCSEED specifies a seed to be used for random number generation in the Monte Carlo
deposit models: MONTE1 and MONTE2.
STICK.COEF specifies the sticking coefficient for the MONTE1 model. Unitless, which must
be between 0.0 and 1.0.
Examples
The following statement defines a machine named TEST that deposits silicon nitride with a
rate of 1500 A/minute using the CVD model with step coverage of 75%.
RATE.DEPO MACHINE=TEST NITRIDE DEP.RATE=1500 A.M CVD STEP.COV=.75
For more examples, see DEPOSIT.

RATE.DEVELOP Statements
6.49 RATE.DEVELOP
RATE.DEVELOP sets development rate and exposure bleaching parameters for each type of
photoresist in Optolith.
Syntax
RATE.DEVELOP
[NAME.RESIST=<c>][G.LINE|H.LINE|I.LINE|DUV.LINE|LAMBDA=<n>
[A.DILL=<n>][B.DILL=<n>][C.DILL=<n>]
[E1.DILL=<n>][E2.DILL=<n>][E3.DILL=<n>
[RMAX.MACK=<n>][RMIN.MACK=<n>][MTH.MACK=<n>][N.MACK=<n>]
[RO.TREFONAS=<n>][Q.TREFONAS=<n>
[RO.HIRAI=<n>][RC.HIRAI=<n>][ALPHA.HIRAI=<n>
[R1.KIM=<n>][R2.KIM=<n>][R3.KIM=<n>][R4.KIM=<n>][R5.KIM=<n>]
[R6.KIM=<n>][R7.KIM=<n>][R8.KIM=<n>][R9.KIM=<n>][R10.KIM=<n>]
[C0.EIB=<n>][C1.EIB=<n>][C2.EIB=<n>][C3.EIB=<n>]
[DIX.0=<n>][DIX.E=<n>]
Description
This command sets the development rate parameters and exposure parameters for each type
of photoresist. These statements can be entered into the athenamod file, so that the parameters
are loaded each time Athena is started.
NAME.RESIST is the photoresist name for this set of parameters.
G.LINE, H.LINE, I.LINE, DUV.LINE, and LAMDBA specify the wavelength for each set
of photoresist parameters.
A.DILL, B.DILL, and C.DILL are the A, B, and C constants for the Dill exposure model.
E1.DILL, E2.DILL, and E3.DILL defines the E1, E2, or E3 parameter for Dill’s
development rate function. These parameters are dimensionless.
RMAX.MACK, RMIN. MACK, MTH.MACK, and N.MACK are the constants for the
Mack development model. RMAX.MACK specifies the development rate of the fully exposed
resist. RMAX.MACK must be specified in microns/sec. RMIN.MACK specifies the development
rate of the unexposed resist. RMIN.MACK must be specified in microns/sec. MTH.MACK is the
threshold normalized PAC concentration. MTH.MACK is dimensionless. N.MACK specifies the
developer sensitivity. N.MACK is dimensionless.
R0.TREFONAS and Q.TREFONAS are constants for the Trefonas development model.
R0.TREFONAS specifies a development rate constant. R0.TREFONAS must be specified in
microns/sec. Q.TREFONAS specifies a development rate constant.

RATE.DEVELOP Statements
R0.HIRAI, RC.HIRAI, and ALPHA.HIRAI are constants for the Hirai development
model. R0.HIRAI specifies the development rate of the fully exposed resist material.
R0.HIRAI must be specified in microns/sec. RC.HIRA specifies a development rate for
unexposed resist. RC.HIRAI must be specified in microns/sec. ALPHA.HIRAI specifies a
dimensionless reaction constant.
R1.KIM, R2.KIM, R3.KIM, R4.KIM, R5.KIM, R6.KIM, R7.KIM, R8.KIM, R9.KIM,
and R10.KIM are constants for the Kim development model. R1.KIM corresponds to the
dissolution rate of the resist material if it has been fully exposed, that is if all the PAC has
been decomposed. R1.KIM must be expressed in microns/sec. R2.KIM corresponds to the
dissolution rate of the unexposed resist material. R2.KIM must be expressed in microns/sec.
R3.KIM corresponds to the dissolution sensitivity of the resist material. R3.KIM is
dimensionless. R4.KIM corresponds to a specific depth into the resist film for surface
retardation effects. R4.KIM must be specified in microns. R5.KIM describes extraordinary
retardation effects. R5.KIM is dimensionless, positive, and less than one. R6.KIM describes
extraordinary retardation effects. R6.KIM is dimensionless, positive, and less than one.
R7.KIM describes extraordinary retardation effects. R7.KIM is dimensionless, positive, and
less than one. R8.KIM describes extraordinary retardation effects. R8.KIM is dimensionless,
positive, and less than one. R9.KIM describes extraordinary retardation effects. R9.KIM is
dimensionless, positive, and less than one. R10.KIM describes extraordinary retardation
effects. R10.KIM is dimensionless, positive, and less than one.
CO.EIB, C1.EIB, C2.EIB, and C3.EIB are the parameters for the Eib development model.
DIX.0 and DIX.E specify pre-exponential constant in cm2/sec and activation energy in eV for
diffusion of photoactive compound that are used in the post exposure bake.
Examples
The following statement defines the Dill development parameters for a user-defined resist
called SECRETX.
RATE.DEVELOP NAME.RESIST=TEST E1.DILL=1 E2.DILL=0.5 E3.DILL=.003
For more examples, see EXPOSE, BAKE, and DEVELOP.

RATE.DOPE Statements
6.50 RATE.DOPE
RATE.DOPE specifies the enhancement parameters for dopant enhanced etching in Elite.
Syntax
RATE.DOPE
MACHINE = <c> MATERIAL I.IMPURITY
[ENH.MAX = <n>] [ENH.SCALE = <n>] [ENH.MINC = <n>]
Description
This statement is used to define dopant enhanced etching and can be applied to an etch
machine defined using the RATE.ETCH statement.
Note: Dopant enhanced etching is not applicable to MC.PLASMA etch model.
MACHINE specifies the machine name for which the dopant enhanced model to be applied.
MATERIAL specifies material in which the dopant enhanced model to be used (see Section
6.2.9 “Standard and User-Defined Materials” for the list of materials).
IMPURITY, INTERST, VACANCY, SXX, SYY, and SXY specify impurity (or other
solution) which concentration is used in the dopant enhanced etching model (see Section
6.2.10 “Standard Impurities” for the list of impurities).
ENH.MAX specifies the maximum enhancement.
ENH.MINC specifies the solution value below which enhancement decays.
ENH.SCALE specifies the spread of the enhancement over solution values (i.e., how quickly
the enhancement factor reach its maximum).
Note: For exponentially varying solutions, e.g., oxidation stress and dopant concentrations, both C and
ENH.MINC are taken to be log base 10 of their respective value.

RATE.ETCH Statements
6.51 RATE.ETCH
RATE.ETCH specifies the etch rate parameters for a machine, which is used in a subsequent
ETCH statement in Elite.
Syntax
RATE.ETCH
MACHINE=<c> MATERIAL | NAME.RESIST=<n>
WET.ETCH | RIE | PLASMA | MC.PLASMA
A.H | A.M | A.S | U.H | U.M | U.S | N.M
[DIRECTIONAL=<n>] [ISOTROPIC=<n>] [CHEMICAL=<n>] [DIVERGENCE=<n>]
[PRESSURE=<n>] [TGAS=<n>] [TION=<n>] [VPDC = <n>] [VPAC=<n>]
[LSHDC=<n>][LSHAC=<n>][FREQ=<n][MGAS=<n>][MION=<n>][QIO=<n>]
[QCHT= <n>]
[CHILD.LANGM|COLLISION|LINEAR|CONSTANT] [IONS.ONLY]
[NPARTICLES=<n>] [ENERGY.DIV = <n>][OUTF.TABLE = <<n>>]
[OUTF.ANGLE = <c>]
[ER.LINEAR|ER.INHIB|ER.COVERAGE|ER.THERMAL]
[K.I=<n>][K.F=<n>][K.D=<n>][SPARAM=<n>][THETA=<n>]
[IONFLUX.THR=<n>][MAX.IONFLUX=<n>][MAX.CHEMFL=<n>][MAX.DEPOFL=<n>]
[ION.TYPES = <n>] [MC.POLYMPT = <n>] [MC.RFLCTDIF = <n>]
[MC.ETCH1 = <n>] [MC.ETCH2 = <n>] [MC.ALB1 = <n>] [MC.ALB2 = <n>]
[MC.PLM.ALB = <n>] [MC.NORM.T1 = <n>] [MC.NORM.T2= <n>]
[MC.LAT.T1 = <n>] [MC.LAT.T2= <n>] [MC.ION.CU1= <n>]
[MC.ION.CU2= <n>]
[MC.PARTS1 = <n>] [MC.PARTS1 = <n>] [MC.ANGLE1=<n>] [MC.ANGLE2=<n>]
Description
This statement is used to define parameters and the machine name for one of four etch models
available in Elite.
MACHINE specifies the machine name for the RATE.ETCH statement.
MATERIAL specifies material for which parameters of the etch machine to be applied (see
Section 6.2.9 “Standard and User-Defined Materials” for the list of materials).
NAME.RESIST specifies the name of photoresist to be etched.
WET.ETCH,RIE, PLASMA, and MC.PLASMA specify a particular model for the
machine definition.

Parameters used for RIE and WET.ETCH models

A.H, A.M, A.S, U.H, U.M, U.S, and N.M specifies that the etch rates are in Angstroms per
hour, Angstroms per minute, Angstroms per second, microns per hour, microns per minute,
microns per second, and nanometers per minute respectively.
DIRECTIONAL specifies the directional component of the etching rate used by the RIE
model. The ionic etch rate is the contribution of the ions to the chemically oriented etching
mechanisms. The ions are assumed to have an anisotropic angular distribution specified by
divergence parameter.
ISOTROPIC specifies the isotropic etch rate used by the WET.ETCH and RIE models. The
isotropic etch rate is the contribution of thermal atoms, radicals, and molecules coming out of
the plasma. These are assumed to have an isotropic angular distribution. Therefore, the
isotropic etching may lead to an underetching of the mask.
CHEMICAL, DIVERGENCE: CHEMICAL is the etch rate in the RIE model normal to the
ion beam when the DIVERGENCE is specified as non-zero. DIVERGENCE specifies the beam
divergence used by the RIE model. The angular distribution of the ions coming down to the
wafer is Gaussian.
Parameters used for Plasma Etch Model
PRESSURE specifies the plasma etcher reactor pressure. Units are mTorr. Default 50 mTorr.
TGAS specifies the plasma etcher reactor gas temperatures. Units are oK. Default is 300 oK.
TION specifies the plasma etcher reactor ion temperatures. Units are oK. Default is 300 oK.
VPDC specifies the DC bias in the plasma sheath. Units are V. Default is 32.5 V.
VPAC specifies the AC voltage in the sheath-bulk interface. Units are V. Default is 32.5 V.
FREQ specifies frequency of the AC current Units are Mhz. Default is 13.6 Mhz.
LSHDC specifies the mean sheath thickness. Units are mm. Default is 0.005 mm.
LSHAC specifies the AC component of the sheath thickness. Units are mm. Default is 0.0.
MGAS specifies the atomic mass of the gas atoms. Default is 40.
MION specifies the atomic mass of the plasma ions. Default is 40.
QIO specifies the momentum transfer cross-section. Units are m2. Default is 1.7e-19
QCHT specifies the charge exchange cross-section. Units are m2. Default is 2.1e-19.
CHILD.LANG, COLLISION, LINEAR, and CONSTANT specify a model used in
calculation of the voltage drop in the plasma sheath. Default is LINEAR.
IONS.ONLY specifies that neutrals to be ignored in plasma simulation. Default is false
NPARTICLES specifies number of particles used for Monte Carlo calculation of the ion flux
coming from plasma. Default is 10,000.
ENERGY.DIV specifies number of energy divisions used for calculation of the plasma ion
flux. Default is 50.

OUTF.TABLE specifies the name of an output file in which complete table of simulated
plasma ions and neutral distributions is saved. The table cannot be loaded using TonyPlot.
The meanings of the columns in the table are:
• i - index of energy from 0 to Nrow=ENERGY.DIV -1, where ENERGY.DIV is the number of
energy divisions specified in the RATE.ETCH statement (default is 50).
• k - index of angle from 0 to Ncol=14. The interval [0, 90] degrees is divided into 15
intervals and each of them are divided into 4 sub-intervals each of 1.5 degrees wide.
• Cergy<n> - number of ions in each sub-intervals, n=0. .3.
• Cergyn<n> - number of neutrals in each sub-intervals, n=0. .3.
• Angadd - the sum of Cergy<n> and Cergyn<n> for each (i,k) pair. It corresponds to the
energy-angle distribution of the particles.
• Angtot - the sum of Angadd for each k, it corresponds to the angle distribution of
particles.
• Erel - the result of i/Nrow +1/(2*Nrow) corresponding to the medium energy in eV of
the 50 normalized intervals [0.0, 0.02], [0.02, 0.04] . . . [0.98, 1.0].
OUTF.ANGLE specifies the name of an output file in which energy-angular ion flux
distribution is saved. The distribution can be plotted using TonyPlot.
ER.LINEAR, ER.INHIB, ER.COVERAGE, and ER.THERMAL specify surface kinetics
model to be used: Simple linear, Adsorbed inhibiting layer, threshold coverage, and thermal
spike models, correspondingly. Default is ER.LINEAR.
K.I specifies the plasma etch rate linear coefficient related to the ion flux.
K.F specifies the plasma etch rate linear coefficient related to the chemical flux.
K.D specifies the plasma etch rate linear coefficient related to the deposition flux.
SPARAM specifies S-parameter of threshold coverage and thermal spike models.
THETA specifies theta parameter of threshold coverage and thermal spike models.
IONFLUX.THR specifies the flux threshold value below which the flux is not considered for
etching. Default is 0.0.
MAX.IONFLUX specifies a multiplier for ion flux generated by the plasma etching
machine. Default is 1.0.
MAX.CHEMFL specifies a multiplier for chemical flux generated by the plasma etching
MAX.DEPOFL specifies a multiplier for deposition flux generated by the plasma etching
Parameters used for Monte Carlo Plasma Etch Model
ION.TYPES specifies the number of different ions in etching plasma.
MC.POLYMPT specifies the number of MC simulated polymer particles normalized to the
volume of the ejected material.
MC.RFLCTDIF specifies the reflection diffusiveness. 1 corresponds to completely diffusive
reflection, 0 corresponds to ideal mirror reflection.
MC.ETCH1 specifies the etch rate parameter for the first type of ions, unitless.
MC.ETCH2 specifies the etch rate parameter for the second type of ions, unitless.

MC.ALB1 specifies the reflection parameter for the first type of ions, unitless. This
coefficient can vary from 0 (no reflection) to 1 (100% reflection).
MC.ALB2 specifies the reflection parameter for the second type of ions, which is unitless.
This coefficient can vary from 0 (no reflection) to 1 (100% reflection).
MC.PLM.ALB specifies the reflection parameter for polymer particles, which is unitless.
This coefficient can vary from 0 (no reflection) to 1 (100% reflection).
MC.NORM.T1 specifies the plasma normal temperature for the first type of ions, which is
unitless.
MC.NORM.T2 specifies the plasma normal temperature for the second type of ions, which
is unitless.
MC.LAT.T1 specifies the plasma lateral temperature for the first type of ions, which is
unitless.
MC.LAT.T2 specifies the plasma lateral temperature for the second type of ions, which is
unitless.
MC.ION.CU1 specifies the plasma ion current density for the first type of ions, ions/second/
cm2.
MC.ION.CU2 specifies the plasma ion current density for the second type of ions, ions/
second/cm2.
MC.PARTS1 specifies the number of MC simulated particles for the first type of ions.
MC.PARTS2 specifies the number of MC simulated particles for the second type of ions.
MC.ANGLE1 specifies the incident angle for the first type of ions. The default is 0 (normal
incidence).
MC.ANGLE2 specifies the incident angle for the second type of ions. The default is 0
(normal incidence).
Wet Etch Example
The following example defines an etch machine that attacks silicon with wet etch
characteristics and an etch rate of 0.1 micron/minute.
RATE.ETCH MACHINE=TEST SILICON WET.ETCH ISOTROPIC=.1 U.M
Monte Carlo Plasma Etch Example
The following statement defines parameters of Monte Carlo Plasma Etch machine as well as
etching characteristics of Silicon associated with this machine.
RATE.ETCH MACHINE=MCETCH SILICON MC.PLASMA ION.TYPES=1 \
MC.PARTS1=20000 MC.NORM.T1=14.0 MC.LAT.T1=2.0 \
MC.ION.CU1=15 MC.ETCH1=1e-05 MC.ALB1=0.2 MC.PLM.ALB=0.5 \
MC.POLYMPT=5000 MC.RFLCTDIF=0.5

RATE.POLISH Statements
6.52 RATE.POLISH
RATE.POLISH specifies the polishing parameters for a chemical mechanical polishing (CMP)
module.
Syntax
RATE.POLISH
MACHINE = <c> MATERIAL | NAME.RESIST=<n>
[A.H|A.M|A.S|U.S|U.M|U.H|N.M]
[SOFT.RATE][HEIGHT.FAC=<n>][LENGTH.FAC=<n>][KINETIC.FAC=<n>]
[MAX.HARD=<n>] [MIN.HARD=<n>]
[ISOTROPIC=<n>]
Description
This command sets the parameters for the CMP machine used in the POLISH statement. The
parameters must be set for each material to be polished. There are two polish models, hard
and soft, that can be used together or separately. Define these models by specifying their
parameters.
MACHINE specifies the machine name.
MATERIAL specifies material for which parameters of the CMP machine to be applied (see
Section 6.2.9 “Standard and User-Defined Materials” for the list of materials).
NAME.RESIST is the user-defined photoresist to be polished.
A.H, A.M, A.S, U.H, U.M, U.S, and N.M specifies that the rates are in Angstroms per hour,
Angstroms per minute, Angstroms per second, microns per hour, microns per minute,
microns per second, and nanometers per minute respectively.
SOFT.RATE is the rate for the soft polish model.
HEIGHT.FAC is the vertical deformation scale for the soft polish model. Units are microns.
LENGTH.FAC is the horizontal deformation scale for the soft polish model. Units are
microns.
KINETIC.FAC is the Kinetic factor (soft polish model). The vertical polish rate increases as
the surface becomes more vertical.
MAX.HARD is the maximum rate for the hard polish model. Corresponds to a pattern factor
of zero.
MIN.HARD is the minimum rate for the hard polish model. Corresponds to a pattern factor
of one.
ISOTROPIC specifies the isotropic etch rate used by the POLISH model.

RATE.POLISH Statements
Examples
The following statements describe a polishing machine named CMP for nitride and oxide.
RATE.POLISH MACHINE=cmp NITRIDE SOFT=4 N.M HEIGHT.FAC=0.02 \
LENGTH.FAC=80 KINETIC.FAC=10 \
RATE.POLISH MACHINE=cmp OXIDE SOFT=25 HEIGHT.FAC=0.02 \
LENGTH.FAC=30 KINETIC.FAC=10
For more examples, see POLISH and RATE.ETCH.

REGION Statements
6.53 REGION
REGION specifies a material to be assigned to a defined mesh region.
Note: Typically, the REGION statement is not required since initial substrate material is specified on the INIT
statement.
Syntax
REGION
MATERIAL [XLO=<c>][YLO=<c>][XHI=<c>][YHI=<c>]
Description
This command specifies the material in a rectangular mesh. REGION statements should follow
LINE statements. Material must be specified for every triangle in a mesh. Therefore for each
rectangular mesh, there must be at least one REGION statement specifying, which material is
included within the mesh. If you do not include REGION statement between the LINE
statement and the INITIALIZE statement, you can define the material on the INITIALIZE
statement.
MATERIAL specifies the material in a region (see Section 6.2.9 “Standard and User-
XLO, YLO, XHI, and YHI specifies the bounds of the region rectangle. The value
<string> should be one of the tags created in a preceding LINE statement.
Examples
The following REGION statement specifies silicon as the material for the entire mesh.
LINE X LOC=0 SPA=1 TAG=LEFT
LINE X LOC=1 SPA=0.1
LINE X LOC=2 SPA=1 TAG=RIGHT
LINE Y LOC=0 SPA=0.02 TAG=SURF
LINE Y LOC=1 SPA=0.1 TAG=BACK
REGION SILICON XLO=LEFT XHI=RIGHT YLO=SURF YHI=BACK
INIT
Note: If you do not use REGION statement and no material appears on the INIT statement, then Athena
assumes Silicon is the starting material. If you do not specify enough regions to describe the material at
every part of the grid, it may not be detected until the execution of a subsequent command.
For more examples, see INITIALIZE.

RELAX Statements
6.54 RELAX
RELAX loosens the grid in an Athena mesh.
Syntax
RELAX
[MATERIAL] [X.MIN=<n>][X.MAX=<n>][Y.MIN=<n>][Y.MAX=<n>]
[DIR.X|DIR.Y] [SURFACE][DX.SURF=<n>
Description
This statement allows you to increase grid spacing. You can place the RELAX statement
anywhere within the input file. RELAX commands, however, are ignored if Athena is in 1D
mode. The RELAX statement also includes an algorithm for relaxing grid on the surface of the
simulation structure.
MATERIAL specifies that RELAX will only apply to the regions of this MATERIAL (see
Section 6.2.9 “Standard and User-Defined Materials” for the list of materials). If MATERIAL is
not specified, RELAX will be applied to all materials in the box.
X.MIN, X.MAX, Y.MIN, and Y.MAX specifies the corner coordinates of the RELAX box.
Units are microns. Default is bounding box of the current simulation structure.
DIR.X or DIR.Y allow the direction of the grid relax to be controlled. DIR.X and DIR.Y are
true by default (i.e., when the RELAX statement is encountered, the grid is relaxed in both
directions by default). When DIR.X or DIR.Y is selected as false (i.e., DIR.X=F or DIR.Y=F),
then the grid is only relaxed in the direction that is left as true.
SURFACE specifies to relax the surface grid.
DX.SURF specifies a minimum size for surface segments.
Examples
RELAX SILICON X.MAX=1 Y.MIN=0
This statement changes a grid over a rectangular area in silicon from the left side of a
structure to 1 and from y=0 to the bottom of the silicon.
Note: RELAX will not make any changes to a grid if obtuse triangles would result from the mesh relaxation.
Consequently, RELAX will typically only work on meshes that were initially defined using LINE statements
in Athena. For other structures, you can use DevEdit. For more information about DevEdit, see the DevEdit
User’s Manual.

SELECT Statements
6.55 SELECT
SELECT selects the variable for printing using the PRINT.1D statement.
Note: This command has been superseded for use with PRINT.1D by the EXTRACT command. See the
DeckBuild User’s Manual for more information.
Syntax
SELECT
[Z=<c>][TEMPERATURE=<n>]
Description
SELECT specifies the variable that will be printed by the PRINT.1D statement. You can only
use one variable at a time. Each SELECT statement overrides any previous statements.
Z is set equal to the selected variable. The operators *, /, +, -, ^ all work as standard
algebraic operators would. Z can be set to any of the vector variables shown on the next page.
Table 6-4 Select Operator Variables.
Vector Variables Description
ANTIMONY antimony concentration

ARSENIC arsenic concentration
BORON boron concentration
CI.STAR equilibrium interstitial concentration
CV.STAR equilibrium vacancies concentration
DOPING net active concentration
ELECTRONS electron concentration
INTERSTITIAL interstitial concentration
NI intrinsic electron concentration
OXYGEN oxygen concentration
PHOSPHORUS phosphorus concentration
Sxx, Sxy, Syy components of stress in rectangular coordinates
TRAP unfilled interstitial trap concentration
VACANCY vacancy concentration
X x coordinates
Y y coordinates
X.V x velocity
Y.V y velocity

SELECT Statements
Table 6-5 Select Functions.
Function Description
abs absolute value

active active portion of the specified dopant
erf error function
erfc complimentary error function
exp exponential
gradx numerically differentiates the argument with respect to x location
grady numerically differentiates the argument with respect to y location
log logarithm
log10 logarithm base 10
<mat1>@<mat2> returns the y value of the interface between <mat1> and <mat2> along a
vertical slice at the given location
scale scales the value given by the maximum value
sqrt square root
TEMPERATURE specifies the temperature at which expressions are evaluated. It defaults

to the last diffusion temperature. This parameter has to be specified (by default or explicitly)
when printing a net active concentration or preparing a Atlas structure file.
Examples
The following example will choose the base 10 logarithm of the arsenic concentration as the
PRINT.1D variable.
SELECT Z=LOG10(ARSEN)
The following chooses the difference between the phosphorus and an analytic profile as the
PRINT.1D variable.
SELECT Z=(PHOS - 1.0E18 * EXP (Y * Y / 1.0E-8))
The following chooses the excess vacancy interstitial product as the PRINT.1D variable.
SELECT Z=(INTER * VACAN - CI.STAR * CV.STAR)
Note: When using log or log10 functions, make sure the argument is positive and non-zero. For example, always
use log10(abs(doping)+1).
For more examples, see the PRINT.1D

SET Statements
6.56 SET
SET specifies strings or numbers for variable substitution.
Note: This commands executed under DeckBuild and is documented fully in the DeckBuild User’s
Manual.
Syntax
SET
variable = <value>
Numerical Variable Example
The following statement defines a variable and performs an expression on it for use later
within the Athena processing syntax.
SET MYDOSE=1e13
SET HALFMYDOSE=$"MYDOSE"/2
IMPLANT BORON DOSE=$"HALFMYDOSE"
String Variable Example
The following uses SET to define a string variable. The saved file will be called
mosfet_fred.str.
SET MYNAME=fred
STRUCTURE OUTFILE=mosfet_$"myname".str
For more examples, see EXTRACT.

SETMODE Statements
6.57 SETMODE
SETMODE specifies execution mode parameters.
Syntax
SETMODE
[NOEXECUTE | ECHO]
Description
This command turns on the following execution mode parameters. The UNSET statement
allows the same parameters to be turned off.
NOEXECUTE puts all entered statements into a check only mode. If this flag is on, Athena
will only check the legality of the input syntax and not execute any statements.
ECHO instructs Athena to echo all input lines to the run-time output. Note that in DeckBuild,
this is not required as all lines are echoed to the bottom run-time window or run-time output
file by default.
Examples
The following statement causes Athena to echo each command it receives.
SETMODE ECHO
For more examples, see UNSETMODE.
Note: The parser does not recognize abbreviated forms of these commands. It requires that you enter
NOEXECUTE and ECHO verbatim.

SILICIDE Statements
6.58 SILICIDE
SILICIDE specifies the silicidation coefficients.
Syntax
SILICIDE
[SILICON|POLYSILICON|TUNGSTEN|TITANIUM|PLATINUM|COBALT|
WSIX|TISIX|PTSIX|COSIX|MATERIAL=<c>]
[/SILICON|/POLYSILICO|/TUNGSTEN|/TITANIUM|/PLATINUM|/COBALT
/WSIX|/TISIX|PTSIX|/COSIX|/MATERIAL=<c>]
[MTTYPE=<c>] [/MTTYPE=<c>] [ALPHA=<n>]
Description
SILICON, POLYSILICON, TUNGSTEN, GAAS, TITANIUM, PLATINUM, COBALT,
WSIX, TISIX, PTSIX, and MATERIAL specify the first material to which the parameters
apply.
/SILICON, /POLYSILICO, /TUNGSTEN, /TITANIUM, /PLATINUM, /COBALT, /
WSIX, /TISIX, /PTSIX, / COSIX, and /MATERIAL specify the second material to which
parameters apply.
MTTYPE specifies the type (metal or silicide) of the user-defined MATERIAL.
/MTTYPE specifies the type (metal or silicide) of the user-defined /MATERIAL.
ALPHA specifies the volume expansion ratio between MATERIAL and /MATERIAL.
Examples
The following example specifies the volume expansion between user-defined material TiSi2
and standard material titanium.
SILICIDE MATERIAL=TISI2 MTTYPE=SILICIDE /MATERIAL=TITANIUM
ALPHA=0.4

SOURCE Statements
6.59 SOURCE
SOURCE executes statements from the specified file.
Syntax
SOURCE
<filename>
Description
SOURCE reads statements from an input file. Statements are read from the file until an end-
of-file marker is found. SOURCE is especially useful for executing a large group of statements.
SOURCE places the named file in the current input stream. SOURCE statements can be nested
up to the limit of open file descriptors (system dependent).
Examples
The following statement causes the contents of a file named test.in to be included into the
input stream.
SOURCE TEST.IN
Note: To support the use of this function when running under the VWF Automation Tools, place the file to be
sourced into a directory directly visible to the simulation run, regardless where the simulator is executing.

STRESS Statements
6.60 STRESS
STRESS calculates elastic stresses.
Syntax
STRESS
[TEMP1=<n>][TEMP2=<n>][NEL=<n>]
Description
This command calculates stresses due to thin film intrinsic stress or thermal mismatch.
TEMP1 and TEMP2 are the initial and final temperatures in oC for calculating thermal
mismatch stresses.
NEL is the number of nodes per triangle element. Both linear approximation with 3 or 4
nodes and quadratic approximation with 6 or 7 nodes are allowed. Linear approximation with
3 nodes is the fastest and is recommended for very large structures. Default is quadratic
approximation with 6 nodes.
Examples
The following calculates the stresses in the substrate and film arising from a nitride layer,
which has an intrinsic stress of 1.4 1010 dynes cm-2 when deposited uniformly.
MATERIAL NITRIDE INTRIN.SIG=1.4E10
STRESS
The following calculates thermal mismatch stress in the whole structure as the result of a
temperature change from 1000 to 100 oC.
STRESS TEMP1=1000 TEMP2=100
For more examples, see MATERIAL.

STRETCH Statements
6.61 STRETCH
STRETCH stretches structures about a specified location.
Syntax
STRETCH
MATERIAL=<c> [LENGTH=<n>] [X.VAL=<n>] [Y.VAL=<n>][STRETCH.VAL=<n>]
[SPACING=<n>][DIVISION=<n>][SNAP]
Description
This statement specifies that the device is to be stretched about a specified location. If device
characterization as a function of length is of interest, the stretch function will save massive
amounts of CPU time in generating multiple gate length structures. The stretch capability is
also useful for power devices.
MATERIAL specifies material that defines the stretch region (see Section 6.2.9 “Standard
and User-Defined Materials” for the list of materials). Default is SILICON.
LENGTH specifies the final value in microns to which the specified material region is
stretched. Alternatively, you can specify X.VAL using STRETCH.VAL to specify the position
of a vertical cut line and the distance to be stretched respectively. The grid spacing within the
stretched region is defined either by spacing or by division.
X.VAL and Y.VAL specifies the horizontal or vertical position in microns at which stretch
occurs. LENGTH overrides the STRETCH.VAL, X.VAL, and Y.VAL parameters. If LENGTH is
specified, the cut line stretch location defaults to the center of the specified material. The
default material is polysilicon.
SPACING specifies the grid spacing within the stretched region. Units are microns.
DIVISION specifies the number of grid divisions within the stretched region.
SNAP indicates that X.VAL should “snap” (change value or locate) to the nearest grid point
before stretching. SNAP is recommended to minimize the potential for obtuse triangle
generation. SNAP is set to true by default.
Stretch Examples
The following statement will stretch a device about the center of its polysilicon region. This
device can have been a MOSFET with a polysilicon gate 1 micron long. The STRETCH
command creates a 1.8 micron-long MOSFET in this case.
STRETCH LENGTH=1.8
The following example will stretch an oxide isolation structure from the x position of 2.3
microns by a value of 1.3 microns. The stretched region contains 14 grid spaces. This case
can be useful for generating large isolation regions that take too long to simulate numerically.
STRETCH OXIDE X.VAL=2.3 DIVISIONS=14 STRETCH.VAL=1.3

STRETCH Statements
Note: The stretch function may not be valid or physically correct in the case of very short initial structures (e.g. with
RSCE effect in MOSFETs). The location selected for stretching should correspond exactly to a grid line for
best results. It will provide best grid quality if the stretch location does not touch areas in which the grid has
been relaxed. The STRETCH command often results in grid failure for complex structures and is not
recommended for complex topographies. DevEdit provides a superior stretch feature for these cases.

STRIP Statements
6.62 STRIP
STRIP removes all photoresist and barrier materials.
Syntax
STRIP
[MATERIAL]
Description
MATERIAL specifies the material to be stripped (see Section 6.2.9 “Standard and User-
Defined Materials” for the list of materials). If no material is specified, the STRIP command
removes both photoresist and barrier materials.
Examples
The following sequence of statements deposits photoresist patterned with the mask level
named CONT, etches oxide through the mask, and removes the photoresist with the STRIP
statement.
MASK NAME="CONT"
ETCH OXIDE DRY THICK=.2
STRIP
This example requires the use of MaskViews.
For more examples, see MASK and ETCH.

STRUCTURE Statements
6.63 STRUCTURE
STRUCTURE writes the mesh and solution information, aerial image information, or flips or
mirrors the structure. SAVEFILE is a synonym for this statement.
Syntax
STRUCTURE
[OUTFILE=<c>]
[FLIP.Y][MIRROR][LEFT|RIGHT|TOP|BOTTOM]
[INTENSITY][MASK][REMOVE.GAS] [SIGE.CONV][TWO.DIM]
Description
This statement writes the entire mesh and solution set to a file. The saved data is from the
current set of solution and impurity values.
OUTFILE specifies the name of the file to be written. Existing files with the same name are
overwritten by newly specified files. OUT.FILE is an alias for this parameter.
FLIP.Y indicates that the structure should be flipped around the x axis. This is used to invert
structures for backside processing.
MIRROR, LEFT, RIGHT, TOP, and BOTTOM mirrors the grid about its left or right, top
or bottom boundary respectively. This is useful for turning half of a MOSFET
simulationstructure into full structure for subsequent Atlas simulation. The default reflection
is about the right axis.
INTENSITY specifies the aerial image intensity distribution to be saved in the output file.
MASK specifies layout mask information to be saved in the output file.
REMOVE.GAS specifies that the gas region is to be removed from the output structure.
Currently, the overlaying gas region is automatically added to the structure for Monte Carlo
etch and BCA implant simulations.
SIGE.CONV converts the layer of silicon that is highly doped with Ge into a Si1-x Gex layer
so it can be used in Atlas.
TWO.DIM specifies that the structure to be transformed into 2D when simulation is
performed in AUTO mode (see Section 6.30 “INITIALIZE”) and the structure is still 1D.
Examples
The following statement writes the current structure to a file called test.str.
STRUCTURE OUTFILE=TEST.STR
The following statement saves an aerial image and masks calculated by Optolith to a file
called test.str.
STRUCTURE OUTFILE=TEST.STR INTENSITY MASK
The following statement mirrors the structure about its left boundary.
STRUCTURE MIRROR LEFT

STRUCTURE Statements
Note: The STRUCTURE command will only save all mesh and solution information. It will not save any defined
model or machine methods. If you exit a simulator the middle of an input file, you may need to manually
parse the preceding METHOD and IMPURITY commands to reinitialize specified parameters. This
function is handled automatically when running under the VWF Automation Tools.
For more examples, see INITIALIZE.

SYSTEM Statements
6.64 SYSTEM
SYSTEM allows execution of any UNIX C-shell command within an input file
Note: The SYSTEM statement is executed by DeckBuild and is fully documented in the DeckBuild User’s
Manual.
Note: The SYSTEM command must be enabled using an option on the DeckBuild Main Control menu under
CategoryOptions.
Examples
The following command will remove all files named test*.str before a DIFFUSE statement
where the DUMP parameter is used.
system \rm -rf test*.str
DIFFUSE .... DUMP=1 DUMP.PREF=test
The SYSTEM command and UNIX commands are case sensitive.
UNIX commands can be concatenated on a single line using the semicolon (;) operator. For
example, to run a third party program that reads and writes Silvaco format files with the fixed
names input.str and output.str.
STRUCTURE OUTF=mysave.str
system mv mysave.str input.str; source myprog.exe; mv output.str
myrestart.str
INIT INF=myrestart.str
The UNIX re-direct symbol, >, is not supported by the system command. The UNIX echo
and sed syntax can be used instead to output values or variables to a given filename. For
example, to save the extracted value of the variable, $myvariable, to the file called myfile.
system echo $”myvariable” | sed -n “w myfile”

TONYPLOT Statements
6.65 TONYPLOT
Tonyplot starts the graphical post-processor TonyPlot.
Note: The Tonyplot statement is executed by DeckBuild, which is fully documented in the deckBuild
User’s Manual.
Examples
All graphics in Athena is performed by saving a file and then loading the file into TonyPlot.
The command
tonyplot
causes Athena to automatically save a file and plot it in TonyPlot. The TonyPlot window will
appear displaying the material boundaries. Use the Plot:Display menu to see more graphics
options.
The following command will display the myfile.str file.
tonyplot -st myfile.str
The following command will overlay the results of myfile1.str and myfile2.str.
tonyplot -overlay myfile1.str myfile2.str
Note: For documentation of the extensive features of TonyPlot for graphical display and analysis, consult the
TonyPlot chapter of the VWF Interactive Tools User’s Manual, Vol. I.

TRAP Statements
6.66 TRAP
TRAP sets the coefficients of interstitial traps.
Syntax
TRAP
MATERIAL [ENABLE] [TOTAL=<n>] [FRAC.0=<n>][FRAC.E=<n>]
Description
This statement allows you to specify values for coefficients of the interstitial traps. The
statement allows coefficients to be specified for each of the materials. Athena has default
values only for silicon. Polysilicon parameters default to those for silicon.
MATERIAL specifies the material for which the parameters apply (see Section 6.2.9
ENABLE indicates that traps should be enabled in the material specified.
TOTAL specifies the total number of traps, in cm-3. The default for silicon is 5.01017 cm-3.
This value is appropriate for Czochralski silicon material.
FRAC.0 and FRAC.E allows the specification of the equilibrium empty trap ratio.
Examples
The following statement turns on interstitial traps and sets the total to 5.01017 and the
fraction to a half.
TRAP SILICON TOTAL=5.0E17 FRAC.0=0.5 FRAC.E=0.0 ENABLE
Note: The trap concentration depend upon the thermal history of the wafer, starting material, stress and
temperature. This history is not considered in the trap model in Athena.
For more examples, see INTERSTITIAL and VACANCY.

UNSETMODE Statements
6.67 UNSETMODE
UNSETMOD unsets execution mode parameters defined in the SETMODE statement.
Note: When SET variable=value is used in DeckBuild, it is impossible to UNSETMODE the variable
Syntax
UNSETMODE
[NOEXECUTE | ECHO]
Description
This command turns off the following execution mode parameters. The SETMODE statement
allows you to turn on the same parameters.
NOEXECUTE puts all entered statements into a check only mode. If this flag is on, Athena
will only check the syntax of the input commands and not actually run them.
ECHO instructs Athena to echo all input lines.
Examples
The following turns off statement echoing.
UNSETMODE ECHO
Note: UNSET is a synonym for this command.
Note: The parser does not recognize abbreviated forms of these commands. It requires that you enter
NOEXECUTE and ECHO verbatim.
For more examples, see SETMODE.

Appendix A
C-Interpreter
C-Interpreter Overview
A.1 C-Interpreter Overview

Athena has a C language interpreter (C-Interpreter) that allows you to modify the models
contained in Athena. In order to use this capability, write a C language analytical function
describing the model. If you’re not familiar with the C language, then we suggest that you
read any of the popular C language books such as [122].
The function arguments of the C-Interpreter functions are fixed in Athena. Thus, you need to
make sure that the arguments and return values match those expected by Athena. To help you,
a set of templates for functions available for the current release of Athena can be obtained by
typing:
athena -T filename
The filename is the name of the file where you want the template to be copied. You can also
obtain the C-Interpreter templates by selecting CommandsTemplates... in Deckbuild. The
following example shows how to use the C-Interpreter function get_damage_values to
modify the default Plus One Model (See Chapter 3 “SSuprem4 Models”) for interstitials
generated during ion implantation.
A.2 Example
/*
Template for the C-Interpreter function for defect formation
during ion implantation
*/
void get_damage_values
(
/* input parameters */
int imp, /* impurity index: As - 2; P - 3; Sb - 4; B - 5; etc.
*/
int mater, /* material index: Si - 3 */
double x, /* x-coordinate in micron*/
double y, /* y-coordinate in microns */
double implanted_conc,/* implanted concentration in 1/cm**3 */
double implanted_dam, /*accumulated damage in eV/cm**3),
do not use without Monte Carlo BCA */
/* return parameters */
double *I_val, /* Interstitial concentration */
double *V_val, /* Vacancy concentration */
double *CL_val, /* 311-Cluster concentration */
double *DL_val /* Dislocation loops concentration */
)
{

Example
if ( mater == 3 ) /* only in Silicon */

{
/* The function modifies +1 interstitial generation model.
The interstitials are generated only in unamorphized layer
where damage is less than 0.1 of atomic density of Si */
if ( implanted_dam < 5e21)
*I_val = implanted_conc;
}
return;
}
The function receives the following input values:
• material index,
• x and y coordinates of the point in the structure where damage is calculated,
• implant concentration at this point,
• implant damage but only when the Monte Carlo BCA model is used for the current
implant calculations.
This allows you to return the values of interstitial and vacancy concentration, and
concentrations of {311} clusters and dislocation loops. If one or few of return values are not
modified in the function the corresponding concentrations will remain unchanged after the
implant.
The function then needs to stored as a file (i.e., damage.lib). The model stored in the
function can then be activated by specifying the DAM.MOD=DAMAGE.LIB parameter in the
IMPLANT statement.
Note: Prior to Athena version 5.4.0.R, DAMAGEMOD.FN (DAM.MOD) was a parameter in the MOMENTS
statement
When you execute the IMPLANT statement using analytical or Monte Carlo models, the
specified C-Interpreter function will be used in place of the build in function.

Appendix B
Default Coefficients
Oxidation Rate Coefficients Default Coefficients
This appendix contains the list of impurity and material default coefficients, default model
parameters, and other parameters used in Athena calculations. Most of these coefficients are
initialized in the athenamod file. The file athenamod will appear when you select
CommandsModels in DeckBuild while Athena is the current simulator. Almost all of
these coefficients can be modified to match measured results.
You should check the contents of athenamod for updates to default values that may be more
current than those shown in the following lists.
B.1 Oxidation Rate Coefficients

B.1.1 Dry Ambient For <111> Orientation
These parameters are from the bibliography reference [37].
Table B-1 Parabolic and Linear Rate Constants for Dry Ambient
Parameter Value
PAR.H.0(µm2/min) 12.8667
PAR.H.E (eV) 1.23
P.BREAK (ºC) 0.
LIN.H.0 (µm/min) 1.038105
LIN.H.E (eV) 2.0
L.BREAK (ºC) 0
B.1.2 Wet Ambient for <111> Orientation

These parameters are from the bibliography reference [39].
Table B-2 Parabolic and Linear Rate Constants for Wet Ambient
Parameter Value
PAR.L.0 (µm2/min) 283.333
PAR.L.E (eV) 1.17
PAR.H.0 (µm2/min) 7.0
PAR.H.E (eV) 0.78
P.BREAK (ºC) 950.
LIN.L.0 (µm/min) 3.45104
LIN.L.E (eV) 1.6
LIN.H.0 (µm/min) 2.95106
LIN.H.E (eV) 2.05
L.BREAK (ºC) 900.

B.1.3 Orientation Factors For Linear Coefficients (both Ambients)
Table B-3 Linear Coefficient Orientation Factors

Orientation Value
For <100> orientation (unitless) ORI.FAC = 0.595
B.1.4 Pressure Dependence

For Dry Oxidation, DRY: L.PDEP = 0.75 and P.PDEP = 1.0. For Wet Oxidation, WET:
L.PDEP = 1.0 and P.PDEP = 1.0.
Table B-4 Thin oxide coefficients (only for dry ambient)

Orientation THINOX.0 THINOX.E (eV) THINOX.L () THINOX.P [14]
(2/min)
<111> 5.87106 2.32 0.0078 1.0
<110> 5.37104 1.80 0.0060 1.0
<100> 6.57106 2.37 0.0069 1.0
B.1.5 Chlorine Dependence
Table B-5 Chlorine dependence of dry coefficients for three temperatures: 900, 1000, 1100ºC
HCL.PC HCL.LIN HCL.PAR
900 1000 1100 900 1000 1100
0.0 1.0 1.0 1.0 1.0 1.0 1.0
1.0 1.75 1.25 1.621 1.083 1.658 1.355
3.0 1.75 1.486 2.207 1.25 1.840 1.490
5.0 1.75 1.486 2.207 1.444 2.075 1.641
7.0 1.75 1.486 2.207 1.639 2.332 1.816
10.0 1.75 1.486 2.207 2.028 2.759 2.102

B.1.6 Doping Dependence Of Oxidation Rate
Table B-6 Doping Dependence of Oxidation Rate

Parameter Value
BAF.EBK 241.6
BAF.PE 0.46
BAF.PPE 1.0
BAF.NE 0.145
BAF.NNE 0.62
BAF.K0 2.6e3
BAF.KE 1.1
See the bibliography reference [40] for more details.
B.1.7 Coefficients for the Analytical Guillemot Model
Table B-7 Coefficents for the Analytical Guillemot Model

Parameter Value
Spread 1.0
INITIAL 0.002
MASK.EDGE -200
ERF.Q 0.05
ERF.DELTA 0.04
ERF.LBB (8.25e-3*(1580.3 - Tox)*
(Eox ^0.67)*(eox ^0.3)*exp
(-( en -0.08)^2)/0.06))
ERF.H (402*(0.445 - 1.75*en)*exp
(- Tox /200)
See the bibliography reference [43] for more details.
B.1.8 Numerical Oxidation Coefficients

For Dry Oxidation, HENRY.COEFF=5E16,THETA= 2.2E22, TRN.0= 1E+3.
For Wet Oxidation, HENRY.COEFF= 3e19,THETA= 2.2e22,TRN.0= 1e+6.
B.1.9 Stress-dependent Growth Model Coefficients

Vc = 78, Vr = 9.7, Vd = 1.9, Vt = 0.0, Dlim = 1.0

B.1.10 Linear Coefficients Of Thermal Expansion

These parameters can be accessed by specifying the LCTE parameter in the MATERIAL
statement.
Table B-8 Linear Coefficients of Thermal Expansion

Parameter Value
SILICON LCTE = 3.052e-6 + 2 * 6.206e-10 * ( T - 293)
OXIDE LCTE = 1.206e-7 + 2 * 2.543e-10 * ( T - 293)
ALUMINUM LCTE = 2.438e-5 + 2 * 6.660e- 9 * ( T - 293)
NITRIDE LCTE = 3.0e-6
POLY LCTE = 3.052e-6 + 2 * 6.206e-10 * ( T - 293)
B.1.11 Volume Expansion Ratio

The volume expansion ratio, ALPHA, can be set in the OXIDE statement.
Table B-9 Volume Expansion Ratio

Ratio Value
silicon/oxide (unitless) 0.44
poly/oxide (unitless) 0.44
Other combinations (unitless) 1.00

Impurity Diffusion Coefficients Default Coefficients
B.2 Impurity Diffusion Coefficients

Table B-10 Impurity Diffusion Coefficients
Parameter Antimony Arsenic Boron Phosphorus
Silicon [124]
DIX.0 (cm2/s) 0.214 8.0 0.037 3.85
DIX.E (eV) 3.65 4.05 3.46 3.66
DIP.0 (cm2/s) 0.0 0.0 0.72 0.0
DIP.E (eV) 0.0 0.0 3.46 0.0
DIM.0 (cm2/s) 15.0 12.8 0.0 4.44
DIM.E (eV) 4.08 4.05 0.0 4.00
2
DIMM.0 (cm /s) 0.0 0.0 0.0 44.2
DIMM.E (eV) 0.0 0.0 0.0 4.37
CTN.0 (cm2/s) 5.1910-24
CTN.E (eV) 0.60
FI (unitless) 0.05 0.20 0.94
Polysilicon
DIX.0 (cm2/s) 21.4 6.6 3.66 385.0
DIX.E (eV) 3.65 3.44 3.46 3.66
DIP.0 (cm2/s) 0.0 0.0 72.0 0.0
DIP.E (eV) 0.0 0.0 3.46 0.0
2 1500.0 1200.0 0.0 443.9
DIM.0 (cm /s)
DIM.E (eV) 4.08 4.05 0.0 4.05
2
DIMM.0 (cm /s) 0.0 0.0 0.0 4420.0
DIMM.E (eV) 0.0 0.0 0.0 4.37
CTN.0 (cm2/s) 5.19 X 10-24
CTN.3(eV) 0.60
Oxide
2 16
DIX.0 (cm /s) 1.3110 1.75 3.1610-4 7.610-3
DIX.E (eV) 8.75 4.89 3.53 3.5
Tungsten Silicide [125]
2
DIX.0 (cm /s) 2.6 2.6 1.010-4 4.2
DIX.E (eV) 2.11 2.11 1.17 2.14
Titanium Silicide [125]
2
DIX.0 (cm /s) 4.8 4.8 1.510-7 392.0
DIX.0 (eV) 2.13 21.3 2.0 2.64
Platinum Silicide [125]
DIX.0 (cm2/s) 2.6 2.6 1.010-3 4.2
DIX.0 (eV) 2.11 2.11 1.17 2.14
All other coefficients for refractory metals and their silicides are set to 0.0.

Impurity Segregation Coefficients Default Coefficients
B.3 Impurity Segregation Coefficients
Table B-11 Impurity Segregation Coefficients

Silicon/oxide
SEG.0  30.0 30.0 1126.0 30.0
(unitless)
SEG.E (eV) 0.0 0.0 0.91 0..0
Poly/oxide
SEG.0  30.0 30.0 1126.0 30.0
(unitless)
SEG.E (eV) 0.0 0.0 0.91 0..0
Other Impurities and Pairs of Materials
SEG.E (eV) 0.0
B.4 Interface Transport Coefficients
Table B-12 Interface Transport Coefficients

Silicon/gas [126]
TRN.0 (cm/s) -3 1.5 27.9 1.5
2.510
TRN.E (eV) 1.04 1.99 2.48 1.99
Poly/gas
TRN.0 (cm/s) -3 1.5 27.9 1.5
2.510
TRN.E (eV) 1.04 1.99 2.48 1.99
Other Impurities and Pairs of Materials
TRN.0 (cm/s) 1.5510-7
TRN.E (eV) 0.0
B.5 Solid Solubility In Silicon

Solubility can be modified for a particular temperature using the SS.TEMP and SS.CONC
parameters in each of the impurity statements.
Table B-13 Solid Solubility in Silicon [127], [128]

Temperature Boron Phosphorus Antimony
(ºC) [cm3] [cm3] [cm3]
800 3.44991019 2.3000e19
825 4.12911019
850 4.90271019 2.79431020

Point Defect Parameters Default Coefficients
Table B-13 Solid Solubility in Silicon [127], [128]

Temperature Boron Phosphorus Antimony
(ºC) [cm3] [cm3] [cm3]
875 5.77771019
900 6.76151019 3.15851020 3.00001019
925 7.86101019
950 9.08321019
975 1.04351020
1000 1.019221020 3.39811020 4.00001019
1025 1.035521020
1050 1.53311020
1075 1.72631020
1100 1.93561020 3.79431020 4.80001019
1125 2.016131020
1150 2.040411020
1175 2.66431020
1200 2.94231020
1225 3.23871020
1250 3.55361020 6.62001019
1275 3.88761020
B.6 Point Defect Parameters

These parameters are for silicon only.
Table B-14 Point Defect Parameters [129]

Bulk Parameters Interstitial Vacancy
D.0 600.0 13.0
D.E (eV) 2.44 2.92
CSTAR.0 (cm-3) 5.01022 2.311021
CSTAR.E (eV) 2.36 1.08
KR.0 8.1610-4 8.1610-4
KR.E (eV) 3.19 3.19

Charge State Information Interstitial Vacancy
NEU.0 1.0 1.0
NEU.E (eV) 0.0 0.0
NEG.0 5.68 5.68

Defect Interface Recombination Parameters Default Coefficients

Charge State Information Interstitial Vacancy
NEG.E (eV) 0.50 0.145
DNEG.0 0.0 32.47
DNEG.E (eV) 0.0 0.62
POS.0 5.68 5.68
POS.E (eV) 0.26 0.45
B.7 Defect Interface Recombination Parameters
Table B-16 Defect Interface Recombination Parameters [129]

Silicon/oxide Interstitial Vacancy
KSURF.0 4.65105 4.18102
KSURF.E (eV) 2.35 2.47
6
KRAT.0 4.4710 2.631012
KRAT.E (eV) 0.94 2.88
KPOW.0 0.84 1.7910-07
KPOW.E (eV) 0.05 -1.91
Silicon/nitride
KSURF.0 1.010-05 1.010-09
Silicon/oxynitride
KSURF.0 1.010-02 1.010-05
Silicon/gas
-04
KSURF.0 1.010 7.010-08
KSURF.E (eV) 0.0 4.08
KPOW.0 1.0 1.0
All parameters for other combinations are 0.0.
B.8 Defect Growth Injection Interface Parameters
Table B-17 Defect Growth Injection Interface Parameters [129]

Silicon/oxide Interstitial Vacancy
THETA.0 [129] 4.58108 0.0
THETA.E (eV) [129] 2.36 0.0
GPOW.0 0.0 1.0
GPOW.E (eV) 0.0 0.0
VMOLE 5.01022 5.01022

Material Parameters Default Coefficients
All parameters for other combinations are 0.0.
B.9 Material Parameters

B.9.1 Diffusion Related
Table B-18 Diffusion Related Material Parameters

Parameter Silicon Poly Oxide Oxynitride Nitride Photo Alumin
NI.0 [130] 3.91016 3.91016 1.0 1.0 1.0 1.0 1.0
NI.POW [130] 1.5 1.5 0.0 0.0 0.0 0.0 0.0
NI.E [130] 0.605 0.605 0.0 0.0 0.0 0.0 0.0
EPS 11.9 11.9 3.9 7.5 7.5 1.0 1.0
B.9.2 Stress and Oxidation Related
Table B-19 Viscosity Parameters

Parameter VISC.0 VISC.E VISC.X
OXIDE (wet) 1.9910-7 5,292 0.499
OXIDE (dry) 3.110-3 7.405 0.499
NITRIDE 15 0 0.499
1.810
SILICON 11030 0 0.499
POLY 51016 0 0.499
OXYNI 12 0 0.499
510
Table B-20 Material Elastic Constants (Compliancies) in units of 10-12 dyn/cm2 [136],[137]
Material S11.ELAST S12.ELAST S44.ELAST
Oxide 1.515 -0.303 3.636
Nitride 0.257 -0.077 0.668
Silicon 0.768 -0.214 1.26
Polysilicon 0.535 -0.150 1.37
Oxynitride 0.257 -0.077 0.668
Aluminum 1.43 -0.500 3.86
6H_SiC 0.134 -0.06 0.388
4H_SiC 0.134 -0.06 0.388
3C_SiC 1.06 -0.476 1.82
Germanium 0.964 -0.26 1.49
SiGe 0.858 -0.236 1.366

Appendix C
TSUPREM-4 and TSUPREM-3
Compatibility Features
General Syntax Capabilities TSUPREM-4 and TSUPREM-3 Compatibility Features
The following changes in Athena syntax and functionalities are implemented in order to
achieve better compatibility with TSUPREM-4and TSUPREM-3 simulators.
C.1 General Syntax Capabilities

• Added capability to specify that default values of some parameters correspond to those of
TSUPREM-4. The modified keyfile athenakey.tma with some modified default values
is introduced. To run Athena with default parameters specified in athenakey.tma file the
syntax is:
go athena simflags="-tma"
For example, TSUPREM-4 defaults for TILT and ROTATION parameters in the IMPLANT
statement are 0°, while Athena uses 7 and 30° respectively.
• A plus character '+' can be used as a line continuation sign instead of standard backslash
'\'.
• Boolean parameters can be set to false by preceding the parameter name with '^' or '!'
character.
• The '$' character can be used to specify the comment line. This should only be used at
the beginning of the line, because the '$' character can be used for substitution of
parameters defined by SET or DEFINE statements of DeckBuild.
• The maximum length of parameter names has extended from 12 to 16 characters. (some
TSUPREM-3 names are longer than 12 characters).
• The first character of a parameter name can be a numeral now.
C.2 Execution Control Capabilities Provided by DeckBuild

The detailed description on these new functionalities will be published in the DeckBuild
manual. Here, we highlight only key features related to the compatibility issues.
C.2.1 DEFINE Statement and Substitutions Capability

DEFINE statement specifies strings for substitution in the following input statements until the
UNDEFINE statement is encountered. The following DEFINE statements and corresponding
substitutions are allowed.
DEFINE dconditions temp=1000 time=10 dry

DIFFUSE dconditions
DEFINE t1 5.0
DEFINE t2 10
DIFFUSE temp=900 time = ( t1 + t2 )

or
DIFFUSE temp=900 time = $t1 + $t2

Execution Control Capabilities Provided by DeckBuild TSUPREM-4 and TSUPREM-3 Compatibility Features
or
DIFFUSE temp=900 time = @t1 + @t2

In case when you have to redefine a string, you should use the %DEFINE statement.
C.2.2 IF/ELSEIF/ELSE/IF.END Capability.

This allows you to perform segments of input deck depending on conditions set in the IF
COND=( condition) or IFELSE COND=(condition). For example, the following sequence
extracts the gate oxide thickness. If it is greater than required 100 A, then the extra oxide
thickness is etched. Otherwise, the lacking oxide thickness is deposited.
extract name="gateox" thickness material="SiO~2" mat.occno=1
x.val=0.5
extract name="gateoxdiff" 1.e-5*($gateox - 100.0)
IF cond=($gateoxdiff > 0.0)
etch oxide thick=$gateoxdiff
ELSE
deposit oxide thick=-($gateoxdiff)
IF.END
C.2.3 LOOP/L.END/ASSIGN/L.MODIFY Capability

LOOP and L.END statements defines the beginning and end of an input deck segment, which
will be processed several times. The number of passes is specified by the STEPS parameter of
the LOOP statement.
The ASSIGN statement allows you to assign a numerical or character value to a name. It is
similar to the DEFINE statement. The main difference is the capability to vary the assigned
value within the LOOP cycle using RATIO and DELTA parameters. Also, only '$' and '@'
characters can be used for substitution.
The L.MODIFY statement allows you to alter the processing of the current LOOP cycle. For
example, if you use BREAK parameter, the current LOOP cycle gets interrupted and control
comes to the first line after the L.END statement.
The following example shows LOOP/L.END/ASSIGN/L.MODIFY capability. It demonstrates a
simple way to estimate a diffusion time needed to grow the gate oxide with the thickness of
100 A.
First, the initial value 0.0 is assigned to the name 'tt'. Within the loop, variable value is
assigned to the name 't'. Its inital value is 1.0 and increases by a factor of 1.2 in each
subsequent loop iteration. The current value of 't' is used as diffusion time.
The next ASSIGN statement calculates the total diffusion time.
The EXTRACT statement finds current oxide thickness 'gateox'. If the required condition is
that extracted 'gateox' exceeds 100 A, the LOOP cycle is interrupted and the total diffusion
time is printed.

MESH Statement TSUPREM-4 and TSUPREM-3 Compatibility Features
ASSIGN name=tt n.val=0.0

LOOP steps=20
ASSIGN name=t n.val=1.0 ratio=1.2
diffuse time=$t temperature=950 dry
ASSIGN name=tt n.val=$tt+$t
extract name="gateox" thickness material="SiO~2" mat.occno=1
x.val=0.1
IF cond = ( $gateox > 100.0 )
L.MODIFY break
IF.END
L.END
echo $tt
C.3 MESH Statement

The new MESH statement provides an alternative to standard mesh generation using the LINE
statements. It also specifies some parameters used in automatic grid generation when layout
information is provided by the Mask Data File generated by Taurus Layout and loaded in the
MASK statement.
DX.MAX specifies the maximum grid spacing in the horizontal direction. It is used when the
grid in the x-direction is specified using the Mask Data File.
DX.MIN specifies the minimum grid spacing in the horizontal direction. It is used when the
grid in the x-direction is specified using the Mask Data File.
DX.RATIO specifies the maximum interval ratio between adjacent grid points in the
horizontal direction. You can also specify this parameter in the INITIALIZE statement. The
default is 1.5.
DY.ACTIV specifies the grid spacing in y-direction at the bottom of the active region.
DY.BOT specifies the grid spacing in y-direction at the bottom of the structure.
DY.RATIO specifies maximum interval ratio between adjacent grid points in vertical
direction. This parameter could be also specified in the INITIALIZE statement. The default
is 1.5.
DY.SURF specifies the grid spacing in the surface region (i.e., between y=0 and
y=LY.SURF).
FAST is equivalent to the AUTO (default) parameter in the INITIALIZE statement.
Note: The MESH statement with LY.SURF and other related parameters cannot be used together with LINE Y
statements.
GRID.FAC specifies a global spacing multiplier, which will be applied to all spacing
parameters when you generate a grid with the INITIALIZE statement. This parameter is
equivalent to the SPACE.MULT parameter of the INITIALIZE statement.
LY.SURF specifies the depth of of the surface region in the default vertical grid (y-direction).

Using MASK statement with the parameter IN.FILE and XLINES for Automatic grid generation in the horizontal direction TSU-
LY.ACTIV specifies the bottom of the active region in the default vertical grid (y-direction).
LY.BOT specifies the depth of the bottom of the structure in the default vertical grid.
C.4 Using MASK statement with the parameter IN.FILE and XLINES for
Automatic grid generation in the horizontal direction
This capability can be used only if Athena runs within DeckBuild. If DeckBuild encounters
the MASK statement with the parameters IN.FILE=<maskfile.tl1> and XLINES, it
recognizes that the file should be in Taurus Layout Mask Data format.
The following information from the maskfile.tl1 are used to build the grid:
• The scale factor (units per micron).
• The minimum and maximum coordinates of mask specification, which define the left and
right of simulation space.
• The number of mask in the file.
• The first line of each mask description includes the mask name (e.g., Poly) and number of
opaque segments. Each subsequent line gives the minimum and maximum coordinates of
each of these segments.
DeckBuild will generate the LINE X statements, which are used by Athena according the
following rules.
1. The lines will be generated at each mask edge.
2. The grid spacing at these lines will be equal to DX.MIN specified in the MESH statement.
3. If none of the mask edges coincides with left or right boundary of the simulation space,
the LINE statement corresponding to such boundaries will be without spacing.
4. Additional one or two LINE statements will be generated between the lines corresponding
to mask edges. The SPAC parameters at these additional lines will be minimum of DX.MAX
and DX.MIN*(DX.RATIO)**n, where DX.MIN, DX.MAX and DX.RATIO are parameters
specified in the MESH statements. This will guarantee that grid spacings in the horizontal
grid will be increased far from mask edges.
These rules are illustrated by the following example of a structure with two POLY gates. If
the Mask Data File has the following fragment
1.000000E+03
0 3000
1
POLY 2
800 1200
1800 2200
and the MESH statement
MESH dx.min=0.01 dx.max=0.1 dx.ratio=2 follows, then DeckBuild will generate the
next sequence of the LINE X statements.
line x loc=0.000000 tag=left
line x loc=0.800000 spac=0.010000

Using mask information with the EXPOSE MASK=<maskname> statement.TSUPREM-4 and TSUPREM-3 Compatibility Features
line x loc=0.950000 spac=0.080000

line x loc=1.050000 spac=0.080000
line x loc=1.200000 spac=0.010000
line x loc=1.350000 spac=0.080000
line x loc=1.650000 spac=0.080000
line x loc=1.800000 spac=0.010000
line x loc=1.950000 spac=0.080000
line x loc=2.050000 spac=0.080000
line x loc=2.200000 spac=0.010000
line x loc=3.000000 tag=right
Note: MASK IN.FILE=<maskfile.tl1> cannot be used together with LINE X statements.
Note: .sec files generated by SILVACO's MaskViews tools provide superior capabilities in generating grid and
mask processing.
C.5 Using mask information with the EXPOSE MASK=<maskname>

statement.
This capability can be used only if Athena runs within DeckBuild. If DeckBuild encounters
the statement EXPOSE with the parameter MASK=<maskname>, it provides Athena with a
sequence of ETCH statements, which will remove photoresist below all transparent regions of
the specified mask. The EXPOSE MASK=<maskname> statement should be preceded by a
DEPOSIT PHOTO statement and followed by a DEVELOP (without parameters) statement. For
example, if you load the same Mask Data File as used above in the MASK statement, the
following input deck fragment appears in the input deck.
DEPO PHOTO THICK=1
EXPOSE MASK=POLY
DEVELOP
DeckBuild will then generate the next sequence of the ETCH statements.
etch photo start x=-0.100000 y=-1000
etch cont x=-0.100000 y=1000
etch cont x=0.800000 y=1000
etch done x=0.800000 y=-1000
etch photo start x=1.200000 y=-1000

etch cont x=1.200000 y=1000
etch cont x=1.800000 y=1000
etch done x=1.800000 y=-1000

Using mask information with the EXPOSE MASK=<maskname> statement.TSUPREM-4 and TSUPREM-3 Compatibility Features

etch cont x=2.200000 y=1000
etch cont x=3.100000 y=1000
etch done x=3.100000 y=-1000
If you use negative photoresist, the photoresist will be removed underneath all opaque
regions.
DEPO PHOTO NEGATIVE THICK=1
EXPOSE MASK=POLY
DEVELOP
Note: Because this is a special DeckBuild capability the DEPOSIT statement in above sequence has some
lexical limitations. The following parameter names/abbreviations must be used: 

DEPOSIT/DEPO
NEGATIVE/NEG
POSITIVE/POS
THICKNES/THICK
PHOTORES/PHOTO
In this case, DeckBuild will generate an alternative sequence of the ETCH statements.
etch cont x=0.800000 y=1000
etch cont x=1.200000 y=1000
etch done x=1.200000 y=-1000

etch cont x=1.800000 y=1000
etch cont x=2.200000 y=1000
etch done x=2.200000 y=-1000
Note: .sec files generated by Silvaco's MaskViews tool provide superior capabilities in the simulation grid
generation and mask processing control.

Aliases and substitutions for some statements TSUPREM-4 and TSUPREM-3 Compatibility Features
C.6 Aliases and substitutions for some statements

AMBIENT is alias for OXIDE.
DIFFUSION is alias for DIFFUSE.
Note: PRINT.1D issues a warning message if there is no SELECT statement prior to it.
ELIMINATE is not used in Athena and therefore is ignored. A warning is then issued. The
RELAX statement should be used instead. The RELAX capability is similar but more flexible
since it can be used in anywhere in the input deck.
ELECTRICAL and MOBILITY are not used in Athena and therefore are ignored. A
warning is issued. The EXTRACT capabilities of DeckBuild should be used instead.
EXTRACT is not used within Athena and therefore is ignored. DeckBuild has superior
extract capabilities. The EQUATION, REACTION, and INTERMEDIATE statements are not part of
Athena. Therefore, they are ignored.
LOADFILE is alias for INITIALIZE.
PLOT (in TSUPREM-3 decks) is ignored and TonyPlot should be used instead.
PLOT.1D, PLOT.2D, PLOT.3D, CONTOUR, LABEL, and COLOR statements are
depreciated in Athena. Warnings are issued. TonyPlot should be used instead.
PRINT (in TSUPREM-3 decks) is alias for PRINT.1D.
SAVEFILE is alias for STRUCTURE.
C.7 Changes in the INITIALIZE statement

Boolean parameters <100>, <110>, and <111> that specify crystalline orientation of the
silicon substrate are aliases for ORIENTATION=100, ORIENTATION=110, and
ORIENTATION=111.
DX.RATIO (new parameter) specifies maximum ration between adjacent mesh lines in
x-direction. It is equivalent to the same parameter in the MESH statement.
DY.RATIO (new parameter) specifies maximum ration between adjacent mesh lines in
y-direction. It is equivalent to the same parameter in the MESH statement.
IN.FILE is alias for INFILE.
RATIO is an alias for INTERVAL.R.
There are several additional ways to specify initial substrate doping. You can specify the
impurity name by using IMPURITY=<impname> parameter, where <impname> could be
TS4.BORON=<n>, TS4.PHOSPHORUS=<n>, TS4.ARSENIC=<n>, and TS4.ANTIMONY=<n>.
You can specify the corresponding doping either by I.CONC=<conc> or
I.RESIST=<resistivity>. You can also specify initial concentration or corresponding
resistivity for an individual impurity by using TS4.BORON=<n>, TS4.PHOSPHORUS=<n>,
TS4.ANTIMONY=<n>, and TS4.ARSENIC=<n>, while boolean parameters TS4.CONCENTR and
TS4.RESISTIV will indicate that the value <n> will be impurity concentration or resistivity,
correspondingly.

Changes in the DEPOSIT statement TSUPREM-4 and TSUPREM-3 Compatibility Features
A one-dimensional grid structure can now be specified without using LINE statements (or
their equivalents MESH, MASK or loading of a .sec file from MaskViews). The syntax of the
INITIALIZE statement of SSuprem3 and its derivatives can be used:
• DX specifies the nominal grid spacing in the initial grid.
• MIN.DX specifies the minimal grid spacing.
• SPACES specifies the number of the grid spaces in the initial structure.
• THICKNESS specifies total thickness of the initial structure.
• TIF is ignored. Athena uses Silvaco Structure File (SSF) format.
• XDX specifies the distance from the top of the initial structure at which nominal grid
spacing is placed.
C.8 Changes in the DEPOSIT statement

There are several additional ways to specify doping in the deposited layer. You can specify
the impurity name by the IMPURITY=<impname> parameter, where <impname> could be
TS4.BORON=<n>, TS4.PHOSPHORUS=<n>, TS4.ARSENIC=<n>, and TS4.ANTIMONY=<n>.
You can specify the corresponding doping either by I.CONC=<conc> or
I.RESIST=<resistivity>. You can also specify concentration or corresponding resistivity
for an individual impurity in deposited material by using TS4.BORON=<n>,
TS4.PHOSPHORUS=<n>, TS4=<n>.ANTIMONY, and TS4.ARSENIC=<n>, while boolean
parameters TS4.CONCENTR and TS4.RESISTIV will indicate that the value <n> will be
impurity concentration or resistivity, correspondingly.
ARC.SPACE is an alias for MIN.SPACE.
To provide compatibility with SSuprem3, the following aliases have been introduced:
• DX for DY
• XDX for YDY
• MIN.DX for MIN.DY
C.9 Changes in the DIFFUSE statement

There are several additional ways to specify impurity concentration in the ambient gas. You
can specify the impurity name using the IMPURITY=<impname> parameter, where
<impname> could be TS4.BORON=<n>, TS4.PHOSPHORUS=<n>, TS4.ARSENIC=<n>, and
TS4.ANTIMONY=<n>. You can specify the corresponding concentration by I.CONC=<conc>.
Alternatively, you can specify concentration of an individual impurity by using
TS4.BORON=<conc>, TS4.PHOSPHOR=<conc>, and so on.
The new parameter SS.IMPURITY (where the generic name IMPURITY could be substituted
by any standard impurity name) specifies that concentration of the named impurity in the
ambient gas is set to its solid solubility in silicon at the current temperature.
C.10 Changes in the ETCH statement

TRAPEZOI is an alias for the DRY.
If THICKNESS is not specified, it assumed to be infinite.

Changes in the STRUCUTURE (SAVEFILE) statement TSUPREM-4 and TSUPREM-3 Compatibility Features
C.11 Changes in the STRUCUTURE (SAVEFILE) statement

OUT.FILE is an alias for OUTFILE.
TIF, DEVICE, and MEDICI are ignored because Athena and other Silvaco TCAD tools use
the universal SSF data format.
C.12 Changes in the IMPLANT statement.

D.PLUS is an alias for PLUS.ONE and UNIT.DAMAGE.
D.SCALE is an alias for DAM.FACTOR.
IMPL.TAB specifies the table from the T-SUPREM4 default implant tables file to be used.
C.13 Changes in the ELECTRODE statement

BOTTOM is an alias for SUBSTRATE.
C.14 Changes in the METHOD statement

OX.ADAPT, IMP.ADAPT, and DIF.ADAPT are aliases for ADAPT.
PD.FERMI is an alias for FERMI.
PD.TRANS is an alias for TWO.DIM.
PD.FULL is an alias for FULL.CPL.
C.15 Changes in the MATERIAL statement

E.FIELD specifies that the electric field terms are to be accounted for in the diffusion
calculations. This parameter is always set to TRUE in semiconductors.
POLYCRYS is an alias for the POLY.DIFF parameter in the METHOD statement.

Bibliography
1. J.D. Plummer, M.D. Deal, and P.B. Griffin, Silicon VLSI Technology. Fundamentals,
Practice and Modeling, Prentice Hall, 2000.
2. G.F. Carey, W.B. Richardson, C.S. Reed, and B. J. Mulvaney. Circuit, Device and
Process Simulation. Mathematical and Numerical Aspects, John Wiley and Sons, 1996.
3. C. C. Lin and M. E. Law, “2-D Mesh Adaption and Flux Discretizations for Dopant
Diffusion Modeling”, IEEE Trans. CAD, v. 15, p. 194, 1995.
4. C. C. Lin and M. E. Law, “Mesh Adaptation and Flux Discretizations for Dopant
Diffusion Modeling”, NUPAD-V, p. 151, 1994.
5. D. Mathiot, and J.C. Pfister. “Dopant Diffusion in Silicon: A Consistent View Involving
Nonequilibrium Defects”, J. of Appl. Phys, v. 55, p. 3518, 1984.
6. P.M. Fahey, Point Defects and Dopant Diffusion in Silicon, Ph.D Thesis, Integrated
Circuits Laboratory, Department of Electrical Engineering, Stanford University, June
1985.
7. M.E. Law, Two Dimensional Numerical Simulation of Dopant Diffusion in Silicon, Ph.D.
Thesis, Department of Electrical Engineering, Stanford University, 1988.
8. J. A. Van Vechten and C. D. Thurmond, “Entropy of Ionization and Temperature
Variation of Ionization Levels of Defects in Semiconductors,” Phys. Rev. B, v. 14, p.
3539, 1976.
9. R.B. Fair, (ed. F.F.Y. Wang). “Concentration Profiles of Diffused Dopants in Silicon”,
Impurity Doping Process in Silicon. North Holland, New York, 1981.
10. F.Lau et.al., “A Model for Phosphorus Segregation at the Silicon-Silicon Dioxide
Interface”, Appl. Phys. A, v. 49, p 671, 1989.
11. Y.S. Oh, D. Ward, “A Calibrated Model for Trapping of Implanted Dopants at Material
Interface During Annealing, IEDM Tech. Digest, p. 509, 1998
12. P.B. Griffin, and J.D. Plummer. “Process Physics Determining 2-D Impurity Profiles in
VLSI Devices”, IEDM Tech. Digest, p. 522, 1986
13. S.M. Hu, “On Interstitial and Vacancy Concentration in Presence of Injection”, J. Appl.
Phys., v. 57, p. 1069, 1985.
14. S. Crowder, Processing Physics in SOI Material, Ph.D Thesis, Department of Electrical
Engineering, Stanford University, 1995.
15. Chakarov, I. and M. Temkin, “Modeling of Ion Implantation in SiC Crystals”, Nucl.
Instr. And Meth. In Phys. Res. v. B-242 (2005):690.
16. http://www.qualitymaterial.net/products.html. Date Accessed: July 3, 2013.
17. http://www.semiconductorwafers.net/4H-Semi-insulating-SiC.html. Date Accessed: July
3, 2013.
18. B.J.Mulvaney, W.B.Richardson, and T.L.Crandle, “PEPPER - A Process Simulator for
VLSI”, IEEE Trans. on Computer-Aided Design, v. 8, p. 336, 1989.
19. L.Mei, M.River, Y.Kwart, and R.W.Dutton, “Grain Growth Mechanism in Polysilicon”,
Proc. 4th Intern. Symp. on Silicon Materials and Technology, v. 81, p. 1007, 1981.

Bibliography
20. L. Mei and R.W. Dutton, “A Process Simulation Model For Multilayer Structures
Involving Polycrystalline Silicon”, IEEE Trans. Electron Devices, v. ED-29, p. 1726,
1982.
21. F. Boucard, Dopant diusion modelling in silicon for shallow junctions processing.Ph. d.
thesis, Louis Pasteur University, September 2003.
22. C. Ortiz and D. Mathiot, "A new kinetic model for the nucleation and growth of
selfinterstitialclusters in silicon," Mat. Res. Soc. Symp. Proc. "Si Front-end
processingphysics and technology of dopant defect interactions lll", edited by E.C.Jones,
K.S. Jones, M.D. Giles, P.Stolk, J.Matsuo, vol. 669, no. -, pp. J5-6, 2001.
23. N. Cowern, G. Mannino, P. A. Stolk, F. Roozeboom, H. G. A. Huizing, J. G. M. van
Berkum, F. Cristiano, A. Claverie, and M. Jaraíz, "Energetics of self-interstitial clusters
in si," Phys. Rev. Lett., vol. 82, no. 22, p. 4460, 1999.
24. T. J. Lenosky, B. Sadigh, S. K. Theiss, M.-J. Caturla, and T. D. de la Rubia, "Ab initio
energetics of boron-interstitial clusters in crystalline si," Appl. Phys. Lett., vol. 77, no.
12, p. 1834, 2000.
25. P. H. Keys, Phosphorus-defect interactions during thermal annealing of ion implanted
silicon. Ph.d. thesis, University of Florida, 2001.
26. A. H. Gencer, Modelling and simulation of transient enhanced diusion based on
interaction of point and extended defects. Ph. d. thesis, Boston University, 1999.
27. L. Pelaz, M. Jaraiz, G. Gilmer, H. Gossmann, C. Raerty, D. Eaglesham, and J. M. Poate,
"B diusion and clustering in ion implanted si: The role of b cluster precursors," Appl.
Phys. Lett., vol. 70, no. 17, p. 2285, 1997.
28. W. Orr Arienzo, R. Glang, R. F. Lever, R. K. Lewis, and F. F. Morehead, "Boron diusion
in silicon at high concentrations," J. Appl. Phys, vol. 63, no. 1, p. 116, 1988.
29. M. Yoshida, E. Arai, H. Nakamura, and Y. Terunuma, "Excess vacancy generation
mechanism at phosphorus diusion into silicon," J.Appl.Phys., vol. 45, no. 4, p. 1498,
1974.
30. S. Solmi, F. Barualdi, and R. Canteri, "Diffusion of boron in silicon during
postimplantation annealing," J. Appl. Phys., vol. 69, no. 4, p. 2135, 1991.
31. B. Colombeau, Interactions entre les défauts étendus et anomalies de diusion des
dopants dans le silicium : modéle physique et simulation prédictives. Thèse de doctorat,
Universite Paul-Sabatier (Toulouse), Septembre 2001.
32. P. Fastenko, Modeling and simulation of arsenic activation and diusion in silicon. Ph. d.
thesis, University of Washington, 2002.
33. B.E. Deal, and A.S. Grove. “General Relationship for the Thermal Oxidation of Silicon”,
J. Appl. Phys. v. 36, p. 3770, 1965.
34. D. Chin, Two Dimensional Oxidation. Modeling and Applications., Ph.D Thesis,
Department of Electrical Engineering, Stanford University, 1983.
35. H.Z. Massoud, Thermal Oxidation of Silicon in Dry Oxygen-Growth Kinetics and
Charge Characterization in the Thin Regime, Technical Report, Stanford Electronic
Laboratories, Stanford University, 1983.
36. H. Eyring, “Viscosity, Plasticity, and Diffusion as Examples of Absolute Reaction Rate”,
J. Chem. Phys. v. 4, p. 283, 1936.

Bibliography
37. B.E. Deal, “Thermal Oxidation Kinetics of Silicon in Pyrogenic H2O and 5% HCl/H2O
Mixtures”, J. Electrochem. Soc. v. 125, p. 576, 1978.
38. D.W. Hess and B.E.Deal. “Kinetics of the Thermal Oxidation of Silicon in O2/HCl
Mixtures”, J. Electrochem. Soc. v. 124, p. 735,1977.
39. R.R. Razouk, L.N.Lie, and B.E.Deal. “Kinetics of High Pressure Oxidation of Silicon in
Pyrogenic Steam”, J. Electrochem. Soc. v. 128, p. 2214, 1981.
40. C.P. Ho, and J.D.Plummer, “Si/SiO2 Interface Oxidation Kinetics: A Physical Model for
the Influence of High Substrate Doping Levels”, J. Electrochem. Soc. v. 126, p. 1516,
1979.
41. B. E. Deal and M. Sklar, “Thermal Oxidation of Heavily Doped Silicon,” J.
ELectrochem. Soc., v. 112, p. 430, 1965.
42. L. N. Lie, R. R. Razouk, and B. E. Deal, “High Pressure Oxidation of Silicon and Dry
Oxygen,” J. Electrochem. Soc., v. 129, p. 2828, 1982.
43. N. Guillemot, “A New Analytical Model of the Bird’s Beak”, IEEE Trans. on Electron
Devices. v. ED-34, p. 1033, 1987.
44. C.L. L.S.Hung, J.Gyulai, J.W.Mayer, S.S.Lau, and M-A.Nicolet, "Kinetics of TiSi2
Formation by Thing Ti Films on Si", J. Appl. Phys., v. 54, p. 5076, 1983
45. C.A.Pico, and M.G.Lagally, "Kinetics of Titanium silicide Formation on Single-Crystal
Si: Experiment and Modeling", J. Appl. Phys., v. 64, p. 4957, 1988.
46. S.Sen-Hou Ko, Sh. Murarka, A.R. Sitaram, "Ellipsometric measurements of the CoSi2
formation from very thin cobalt films on silicon", J. Appl. Phys., v.71, p. 5892, 1992.
47. M.A. Nicolet and S.S. Lau, in VLSI Handbook, Ed. Norman G. Einspruch, Academic
Press, p.430, 1985.
48. J. Lindhard, M. Scharff, and H.E. Schiott. “Range Concepts and Heavy Ion Ranges”,
Kgl. Dan. Vid. Selsk. Mat.-fys. Medd., v. 33, 1963.
49. R. Smith (Ed.), Atomic and Ion Collisions in Solids and at Surfaces, Cambridge
University Press, 1997.
50. D.G. Ashworth, R. Oven and B. Mundin, “Representation of Ion Implantation Profiles be
Pearson Frequency Distribution Curves”, J. Phys. D, v. 23, p. 870, 1990.
51. A. F. Tasch, “An Improved Approach to Accurately Model Shallow B and BF2 Implants
in Silicon”, J. Electrochem. Soc., v. 136, p. 810, 1989.
52. K.B. Parab et.al., “Analysis of Ultra-Shallow Doping Profiles Obtained by Low Energy
Ion Implantation”, J. Vac. Sci. Technol., v. B14, p. 260, 1996.
53. G.A.J. Amaratunga, K. Sabine, and A.G.R. Evans. “The Modeling of Ion Implantation in
a Three-Layer Structure Using the Method of Dose Matching”, IEEE Trans. Electron.
Dev. v. ED-32, p. 1899, 1985.
54. A.F. Burenkov, F.F.Komarov, and M.M.Temkin. “Analytical Calculation of Ion
Implantation through Mask Windows” (in Russian), Microelektronika, v. 16, p. 15, 1987.
55. A.F. Burenkov, A.G. Kurganov, and G.G. Konoplyanik, “Two-Dimensional Local Ion
Implantation Distribution” (in Russian), Povekhnost (Surface Sciences), v.8, p.52, 1989.
56. J. Lorenz, W. Kruger, and A. Barthel “Simulation of the Lateral Spread of Implanted
Ions: Theory”, NASECODE-VI, Ed. J.J.H.Miller, Boole Press, p.513, 1989.

Bibliography
57. D.G. Ashworth, M.D.J. Bowyer, and R. Oven, “Representation of Ion Implantation
Distributions in Two and Three Dimensions”, J. Phys. D, v. 24, p. 1120, 1991.
58. G. Hobler, E. Langer, and S. Selberherr, “Two-Dimensional Modeling of Ion
Implantation with Spatial Moments”, Solid-State Electronics, v. 30, p. 445, 1987.
59. M. Temkin and I. Chakarov, “Computationally Effective Model for 2D Ion Implantation
Simulation”, Semiconductor Process and Device Performance Modeling, Eds, S.T.
Dunham, J.S. Nelson, MRS, p. 27, 1998.
60. R.Oven, D.G.Ashworth, and M.D.J. Bowyer, “Formulas for the Distribution of Ions
Under an Ideal Mask”, J. Phys. D,. v. 25, p.1235, 1992.
61. J.E. Gibbons, W. S. Johnson, and S. W. Mylroie. Projected Range Statistics (2nd
edition), Stroudsburg, Pennsylvania: Dowden, Hutchinson, & Ross, Inc., 1975.
62. A.F. Burenkov, F.F.Komarov, M.A.Kumakhov, and M.M.Temkin. Tables of Ion
Implantation Spatial Distributions. Gordon & Breach Science Publishers, 1986.
63. O. B. Firsov, Soviet Physics JETP, v. 33, p. 696, 1957
64. J. F. Ziegler, J. P. Biersack, U. Littmark, The stopping and range of ions in solids, v. 1,
Pergamon Press, 1985.
65. UT-Marlowe Version 5.0 User manual, University of Texas, Austin, USA, 1999.
66. O. B. Firsov, Soviet Physics JETP, v. 36, p. 1517, 1959.
67. W. Brandt and M. Kitagawa, “Effective Stopping-Power Charges of Swift Ions in
Condensed Matter.”, Phys. Rev. B, v. 25, p. 5631, 1982.
68. F. L. Vook, Defects in Semiconductors, p. 60, 1972
69. I. R. Chakarov and R. P. Webb, ”CRYSTAL -- Binary Collision Simulation of Atomic
Collision and Damage Buildup in Crystalline Silicon”, Radiation Effects, v. 130-131, p.
447, 1994.
70. H. Kahn and T. E Harris, “Estimation of Particle Transmission by Random Sampling”,
National Bureau of Standards Applied Mathematics Series, v. 12, p. 27, 1951.
71. B. A. Bayes, “Statistical Techniques for Simulation Models”, The Australian Computer
Journal v. 2, p. 190, 1975.
72. J. M. Hemmersley and D. C. Handscomb, Monte Carlo Methods, Methuen and Co., Ltd.,
London, 1964.
73. M. Villién-Altamirano and J. Villién-Altamirano,. “RESTART: A Method for
Accelerating Rare Event Simulations”, Proc. 13th Int. Teletraffic Congress, ITC 13
(Queueing, Performance and Control in ATM), P. C. Cohen J. W., Ed., North-Holland,
Copenhagen, Denmark, p. 71, 1991;
74. A. Phillips and P. J. Price, “Monte Carlo Calculations on Hot Electron Energy Tails”,
Appl. Phys. Lett., v. 30, 1977
75. S.H.Yang, D. Lim, S. Morris, and A. F. Tasch, “A More Efficient Approach for Monte
Carlo Simulation of Deeply-Channeled Implanted Profiles in Single-Crystal Silicon”,
Proc. NUPAD, p. 97, 1994.
76. K. M. Beardmore and N. Gronbech-Jensen, “Efficient Molecular Dynamics Scheme for
the Calculation of Dopant Profiles Due to Ion Implantation”, Phys. Rev. E, 57, 1998, p.
7278.

Bibliography
77. J. M Hernández-Mangas, J. Arias, M. Jaraiz, L. Bailón, and J. Barbolla, “Algorithm for

Statistical Noise Reduction on Three-Dimensional Ion Implant Simulations”. Nucl. Instr.
Meth. in Physics Research B, 174, 2001, p, 433-438.
78. W. Bohmayr, A. Burenkov., J. Lorenz, H. Ryssel, and S. Selberherr, “Trajectory Split
Method for Monte Carlo Simulation of Ion Implantation”, IEEE Transactions on
Semiconductor Manufacturing, v. 8, p. 402, 1995.
79. P Glasserman, P. Heidelberger, P. Shahabuddin,and T.Zajic, A look at Multilevel
Splitting, Technical report RC-20692, IBM Research Division, T. J. Watson Research
Center, Yorktown Heights, New York, 1997.
80. S.E. Hansen and M. Deal, “SUPREM-IV.GS: Two-Dimensional Process Simulation for
Silicon and Gallium Arsenide”, Integrated Circuits Laboratory, Stanford University,
1993.
81. O. Madelung (Ed.). Semiconductors - Basic Data. Springer Verlag, 1996.
82. R. Anholt et.al., ”Ion Implantation Into Gallium Arsenide”, J. Appl. Phys. v. 64, p.3429,
1988.
83. K. Rajendran, "Simulation of Boron Diffusion in Strained SiGe Epitaxial Layers", Proc.
SYSPAD, p. 206, 2000.
84. R.F. Lever, "Boron Diffusion across Silicon-SiGe Boundaries", J. of Appl. Phys., v. 83,
p. 1988, 1998.
85. J.L. Ngau, P.B. Griffin, and J.D. Plummer, “Modelling the Suppression of Boron
Transient Enhanced Diffusion in Silicon by Substitutional Carbon Incorporation”,
Japanese. of Appl. Phys., v. 90, p. 1768, 2001
86. R.Jewett, A String Model Etching Algorithm, M. S. Thesis, University of California,
Berkeley, 1979.
87. W.G.Oldham et.al., “A General Simulator For VLSI Lithography And Etching
Processes: Part II - Application To Deposition And Etching”, IEEE Trans. on Electron
Devices, v. ED-27, p. 1455,1980.
88. A.R.Neureuther, C.H.Ting, and C.Y.Lin, “Application of Line-Edge Profile Simulation
to Thin-film Deposition Process”, IEEE Trans. on Electron Devices, v. ED-27, p. 1449,
1980.
89. A.R.Neureuther, “Basic Models And Algorithms For Wafer Topography Simulation”,
Problems and New Solutions for Device and Process Modeling, Ed. J.J.H. Miller, Boole
Press, Dublin, p. 99, 1985.
90. A.R.Neureuther, “Algorithms For Wafer Topography Simulation”, NASECODE IV,
Dublin, Ireland, p.58, 1985.
91. SAMPLE User Guide, Department of Electrical Engineering and Computer Sciences, UC
Berkeley, 1991.
92. M. Sikkens, et.al., Opt. Eng., v. 25, p. 142, 1986.
93. R.N. Tait, T. Smy and M.J. Brett “A Ballistic Deposition Model for Films Evaporated
Over Topography”, Thin Solid Films, v. 187 p. 375, 1990.
94. R.N. Tait, S.K. Dew, T. Smy and M.J. Brett “Ballistic Simulation of Optical Coatings
Deposited Over Topography”, SPIE Proc. v. 1324, p. 112, 1990.

Bibliography
95. R.N. Tait, T. Smy and M.J. Brett “Simulation and Measurement of Density Variation in
Mo Films Sputter Deposited Over Oxide Steps”, J. Vac. Sci. Technol. v. A8, p. 1593,
1990.
96. S.F.Meier, Etching Simulation Of Nonplanar Layers, M.S. Thesis, UC Berkeley, 1987.
97. J.L.Reynolds, A.R.Neureuther, and W.G.Oldham, “Simulation of Dry Etched Line
Etched Profiles”, J. Vac. Sci. Technol., v. 16, p. 1772, 1979.
98. A.R.Neureuther, C.Y.Liu, and C.H.Ting, “Modeling Ion Milling”, J. Vac. Sci. Technol.,
p. 1167, 1979.
99. S. Takagi, K. Iyanagi, S. Onoue, T. Shinmura, and M. Fujino, "Topography Simulation
of Reactive Ion Etching Combined with Plasma Simulation, Sheath Model, and Surface
Reaction Model", Japanese J. Appl. Phys., v. 41, no. 6A, p. 3947, 2002.
100.P.Sutardja, Y.Shacham-Diamand, and W.G.Oldham, “Two Dimensional Simulation of
Glass Reflow And Silicon Oxidation”, VLSI Technology Technical Digest, p. 39, 1986.
101.P.A.Burke, “Semi-Empirical Modeling of SiO2 Chemical-Mechanical Polishing
Planarization”, Proc. VMIC Conf., p. 379, 1991.
102.J.Warnock, “A Two-Dimensional Process Model for Chemimechanical Polish
Planarization”, J. Electrochem. Soc. v. 138, p. 2398, 1991.
103.B.M. Watraslewicz, “Image Formation in Microscopy at High Numerical Aperture”,.
Optical Acta, v. 12, p. 167, 1965.
104.B. Richards, E.Wolf, “Electromagnetic Diffraction in Optical Systems II. Structure of
the Image Field In An Aplanatic System”, Proc. Phys Soc., v. A 253, p. 358, 1959.
105.H.H. Hopkins, “On the Diffraction Theory of Optical Images”, Proc. Roy. Soc., v. A
217, p. 408, 1953.
106.H.H. Hopkins, “Applications of Coherence Theory In Microscopy And Interferometry”,
J. Opt. Soc. Am., v. 47, p. 508, 1957.
107.J.Tsujiuchi, “Image Forming Performance of Projection Systems”, Japanese. J. of Appl.
Phys., v. 4, Suppl. I., Proc. Conf. On Photographic And Spectroscopic Optics, p. 251,
1965.
108.W.H.A. Fincham, M.H.Freeman, Optics, Butterworths, London, Ch. 15, 1980.
109.J. Van Roey, J.van der Donk, P.E. Lagasse. “Beam-Propagation Method: Analysis and
Assessment”. J. Opt. Soc. Am., v. 71, p. 803, 1981.
110.J. Z. Y. Guo, F. Cerrina. “Modeling X-ray Proximity Lithography”. IBM J. Res.
Develop., v. 37, p. 331, 1993.
111.A. Erdmann, C. L. Henderson, C. G. Wilson, W. Henke. “Influence of Optical
Nonlinearities of the Photoresist on the Photolithographic Process: Basics”. SPIE Proc.,
v. 3051, p. 529, 1997.
112.A. Erdmann, C. L. Henderson, C. G. Wilson, R. R. Dammel. “Some aspects of thick
film resist performance and modeling”. SPIE Proc., v. 3333, p. 1201, 1998.
113.D.A. Bernard, “Simulation Of Post Exposure Bake Effects On Photolithographic
Performance of a Resist Film”, Phillips Journal of Research, v. 42, p. 566, 1987.
114.F.H. Dill, W.P. Hornberger, P.S. Hauge, J.M.Shaw, “Characterization of Positive
Photoresist”, IEEE Trans. Electron Devices, v. ED-22, p. 445, 1975.

Bibliography
115.D.J. Kim, et. al., “Development of Positive Photoresist”, IEEE Trans. Electron Devices,
v. ED-31, p. 1730, 1984.
116.C.A. Mack, “PROLITH: A Comprehensive Optical Lithography Model”, SPIE Proc., v.
538, p.207, 1985.
117.P. Trefonas III, et al., “New Principle For Image Enhancement In Single Layer Positive
Photoresist”, SPIE Proc., v. 771, p. 194, 1987.
118.Y. Hirai et.al., “Process Modeling For Photoresist Development And Design Of Drl/sd
(double Resist Layer) by a Single Development Process”, IEEE Trans. on CAD, v. CAD-
6, p. 403, 1987.
119.J.Van Roey, J. vander Donk, P.E. Lagasse, “Beam-propagation method: analysis and
assessment”, J. Opt. Soc. Am., Vol. 71, No. 7, p.803, July 1981.
120.J.Z.Y.Guo, F.Cerrina, “Modeling X-ray proximity lithography”, IBM J. Res. Develop.,
Vol. 37, No. 3, p.331, May 1993.
121.J.W.Goodman, Introduction to Fourier Optics, McGraw-Hill, New York, 1969.
122.B. Kernighan and D.Ritchie, The C Programming Language, Prentice-Hall, 1988.
123.D.B. Kao, J.P.McVittie, W.D.Nix, and K.C.Saraswat. “Two Dimensional Thermal
Oxidation of Silicon I. Experiments”, IEEE Trans. Electron. Dev., ED-34, p.1008, 1987.
124.R.B. Fair, Impurity Doping Process in Silicon, Ed. F.F.Y. Wang, North Holland,
Amsterdam, 1981.
125.C.L. Chu, “Characterization of Lateral Diffusion in Silicides”, IEDM Tech. Digest, p.
245, 1990.
126.P.H. Langer and J.I. Goldstein. “Boron Autodoping During Silane Epitaxy”, J.
Electrochem. Soc., v. 124, p. 591, 1977.
127.G.L. Vick and .M. Whittle. “Solid Solubility and Diffusion Coefficients of Boron in
Silicon”, J. Electrochem. Soc. v. 116, p. 1142, 1969.
128.F.A. Trumbore “Solid Solubilities of Impurity Elements in Germanium and Silicon”,
Bell System Tech. J. v. 39, p. 205, 1960.
129.H. Park and M. Law, “Point Defect Based Modeling of Low Dose Silicon Implant
Damage and Oxidation Effects on Phosphorus and Boron Diffusion in Silicon”, J. of
Appl. Phys. v. 72, p. 3431, 1992
130.F.J. Morin and J.P. Maita. “Electrical Properties of Silicon Containing Arsenic and
Boron”, Phys. Rev. v. 96, p. 28, 1954.
131.M. Giles, “Transient Phosphorus Diffusion Below the Amorphization Threshold”, J.
Electrochem. Soc. v. 138, p. 1160, 1991.
132.P.Gas et.al., "Diffusion of Sb, Ga, Ge, and As in TiSi2," J.Appl.Phys., v. 63, p. 5335,
1988.
133.P.Gas et.al., "Boron, Phosporus, and Arsenic Diffusion in TiSi2," J.Appl.Phys., v. 60,
p.1634, 1986.
134.C.M.Osburn et.al., "The Effect of Titanium Silicide Formation on dopant
Redistribution," J.Electrochem. Soc., v. 135, p.1490, 1988.
135.Chi On Chui et.al., “Activation and Diffusion Studies of Ion-implanted p and n Dopants
in Germanium”, Appl. Phys. Lett., v.83, p. 3275, 2003

Bibliography
136.S. Uppal, “Diffusion of Ion-implanted Boron in germanium”, J. Appl. Phys., v.90, p.

4293, 2001.
137.N.R. Zandenberg, et.al., “Boron and Phosphorus Diffusion in Strained and Relaxed Si
an SiGe”, J. Appl. Phys., v.94, p.3883, 2003.
138.Yi-M. Sheu et.al., “Modeling Mechanical Stress Effect on Dopant Diffusion in Scaled
MOSFETs”, IEEE Trans. Electron Devices, v. 52, p.30, 2005.
139.V. Senez, P. Ferreira, and B. Baccus, “2-D Simulation of Local Oxidation of Silicon:
Calibrated Viscoelastic Flow Analysis,” IEEE Trans. Electron Devices, v. 43, p. 720,
1996.
140.W.A. Brantley, J. Appl. Phys., Vol. 44, p. 534, 1973.

Index
A B
Adaptive Meshing .............................................................. 116 Bank-Weiser Error Estimator ................................................119
Adaptive Meshing Control ................................................ 120 Basic Diffusion and Oxidation Models .......................................64
Base Mesh Formation ..................................................... 120 Beam Propagation Method (BPM) ..................................270, 279
Heat Cycle .................................................................... 117
Bimolecular Recombination
Interface Mesh Control .................................................... 120
CNET ..........................................................................159
Ion Implantation ............................................................. 116
Binary Collision Approximation ..............................................328
Advanced Diffusion Model Examples ...................................... 166
Cowern’s Experiment ...................................................... 168 Bipolar Process Flow ............................................................74
Implantation Diffusion Experiment ...................................... 170 Conclusions ....................................................................79
Pelaz Experiment ........................................................... 169 The Base Current Profile – Low Injection ...............................78
Predeposition ................................................................ 166 The Base Current Profile – Medium Injection ..........................77
Tuning Base and Collector Currents – All Regions ...................75
Advanced Diffusion Models .................................................. 151
Tuning the Base Current – All Regions ..................................75
Classical Model of Dopant Diffusion (CNET) ......................... 153
Tuning the Collector Current – All Regions .............................76
Electrical Deactivation and Clustering (DDC) ........................ 163
Interstitials Clusters (IC) ................................................... 160 Buzz Saw Model. See Hard Polish Model ................................256
Solid Solubility ............................................................... 160
Vacancy Cluster (VC) ...................................................... 162 C
Advanced Features .............................................................. 82 Changing the Method Statement During the Process Flow
Deposition and Wet/Dry Etching .......................................... 85 Switching guidelines .........................................................54
MaskViews ..................................................................... 92
Charge States
Structure Manipulation Tools ............................................... 82
CNET ..........................................................................153
Analytic Implant Models
Chemical and Active Concentration Values ..............................127
Dual Pearson ................................................................ 198
Gaussian ...................................................................... 195 Chemical Mechnical Polish (CMP)
Pearson ....................................................................... 195 Hard Polish Model ..........................................................256
Screen Oxide Thickness Parameter (S.OXIDE) ..................... 199 Soft Polish Model ...........................................................258
SIMS-Verified Dual Pearson (SVDP) Model .......................... 198 C-Interpreter .....................................................222, 254, 404
Arrhenius expression .......................................................... 129 CNET Charge States
Athena .............................................................................. 25 Dopant-defect pairs ........................................................154
Point Defects .................................................................153
Athena 1D .................................................................... 18, 25
CNET Flux Equations
Athena Features and Capabilities ............................................ 13
Dopant--Defects Pairs .....................................................156
Athena Input/Output ............................................................. 26 Point Defects .................................................................155
Input .............................................................................. 26
CNET Generation-Recombination Terms
Output ........................................................................... 27
Formation of Pairs ..................................................158, 159
Standard Structure File Format(SSF) .................................... 27
Frenckel Pair Recombination ............................................158
Athena/Optolith ................................................................. 103
Compound Semiconductor Simulation
CD Extraction, Smile Plots, And Looping Procedures ............. 116
Diffusion Models ............................................................226
Illumination System ......................................................... 107
Ion Implantation Models ...................................................227
Imaging Control ............................................................. 111
Mask ........................................................................... 104 Concentration Jump Condition ..............................................131
Material Properties ......................................................... 113 Continuity Equation, see also Diffusion Equation .......................128
Projection System .......................................................... 110 Correct Substrate Depth Modelling
Structure Exposure ......................................................... 113 Diffusion .........................................................................61
athenaimp .......................................................................... 26 Ion Implantation ...............................................................57
athenamod ........................................................... 22, 26, 286
athenamod.97 ..................................................................... 24 D
athenares ........................................................................... 26 DDC
Arsenic ........................................................................165
Boron ..........................................................................163
Phosphorus ..................................................................164

Deactivation Threshold ........................................................146 Transient Activation Model ................................................148
Deal-Grove Model ..............................................................188 Epitaxy Simulation ..............................................................224
DeckBuild ...........................................................18, 390, 400 ERFG model .....................................................................188
Interactive Mode ..............................................................23 Etch Models ..............................................................225, 247
Running Athena inside .......................................................24 Isotropic .......................................................................247
Deckbuild Linear ..........................................................................253
Batch Mode .....................................................................23 Monte Carlo ..................................................................250
No Windows Batch Mode ...................................................24 Plasma .........................................................................249
defect ..............................................................................127 RIE ..............................................................................247
Defect Diffusion Exposure Module ...............................................................270
Time Step Control ...........................................................143 Resist exposure with accounting dose effect .........................272
DEPOSIT .........................................................................302 Resist exposure without accounting dose effect .....................272
Deposition Models ......................................................224, 238
Conformal .....................................................................238 F
Conical ........................................................................244 Flux Equations
Custom ........................................................................246 CNET ..........................................................................155
CVD ............................................................................238 Flux Expression .................................................................130
Doped Layers ................................................................224
Flux Jump Condition ...........................................................131
Dual Directional .............................................................240
Epitaxy Simulation ..........................................................224 Free Point Defect Damage ...................................................220
Grid Control ..................................................................224 Fresnel diffraction .......................................................278, 279
Hemispheric ..................................................................241 Fully Coupled Equations
Monte Carlo ..................................................................245 CNET ..........................................................................155
Planetary ......................................................................242 Fully Coupled Model
Unidirectional ................................................................238 High Concentration Extension ...........................................144
Deposition Wet/Dry Etching Fully-Coupled Model
Defining Elite Deposition Machines .......................................87 RTA Diffusion Modelling ...................................................145
Defining Elite Etch Machines ...............................................90
Modifying Athena/Elite Default Machines ...............................87
Using A Specified Etch Machine ..........................................91
G
Development Models Generation-Recombination Terms
Dill ..............................................................................274 CNET ..........................................................................157
Hirai ............................................................................276 Grain-based Polysilicon Diffusion Models ................................149
Kim .............................................................................275 Two stream ...................................................................149
Mack ...........................................................................275 Grid Control ................................................................40, 356
Trefonas .......................................................................275 Gummel Plot .......................................................................74
Development Modules .........................................................274
Device Structure ..................................................................26 H
Diffusion Equation ..............................................................129
Helmholtz equation .............................................................270
Diffusion Models ................................................................127
Huygens diffraction approximation .........................................278
Electrical Deactivation and Clustering Models .......................146
Fermi ...........................................................................131
Fully Coupled Model .......................................................143 I
Impurity Segregation .......................................................132 Imaging Module .................................................................262
Steady State .................................................................142
impurity ............................................................................127
Two Dimensional ............................................................134
Impurity Segregation Model ..................................................132
Dislocation Loop Based Enhanced Bulk Recombination ..............142
Interface Trap ................................................................133
Dislocation Loops ...............................................................142
Initial Structure ....................................................................28
DLTS ..............................................................................128 Electrodes ......................................................................48
dopant .............................................................................127 Initial Substrate ................................................................33
Dose Loss Model, see Interface Trap ......................................133 Rectangular Grid ..............................................................28
Reducing Grid Points in Non-Essential Areas using the
E Relax Parameter ..............................................................44
Reflecting a Structure in the “Y” Plane using the
Electrical Deactivation and Clustering Models Mirror Parameter ..............................................................47
Electrical Activation Model ................................................146

Simple Film Depositions .................................................... 36 MOSFET Process Flow
Simple Geometrical Etches ................................................. 41 Conclusion ......................................................................73
Structure File for Plotting or Initializing an ATHENA Input file Input ..............................................................................68
for Further Processing ....................................................... 50 PMOS Tuning ..................................................................72
Ion Implantation Damage Predictive Powers of Tuned Process Parameters Checking .......73
C-Interpreter ................................................................. 222 Tuning Diffusion Parameters ...............................................72
Cluster Model ................................................................ 221 Tuning Implantation Parameters ..........................................71
Dislocation Loops Model .................................................. 221 Tuning Oxidation Parameters ..............................................69
Plus 1 Model ................................................................. 220 Using Atlas for MOS Process Tuning ....................................72
Ion Implantation Models ...................................................... 195 Multi-Layer Implants
Analytic ........................................................................ 195 DOSE.MATCH ...............................................................200
Ion Implantation Damage ................................................. 220 MAX.SCALE .................................................................200
Monte Carlo .................................................................. 207 MOM.SCALE .................................................................201
Multi-Layer ................................................................... 200 RP.SCALE ....................................................................200
Stopping Powers in Amorphous Materials and Range
Validation ..................................................................... 222 N
Two-Dimensional Implant Profiles ...................................... 202
Nuclear Scattering ......................................................207, 208
K O
Klaassen bandgap narrowing model ......................................... 78
Operation Modes .................................................................23
Optical System ..................................................................268
L
Ostwald ripening ................................................................138
Linear Rate Constant OXIDATION models ...........................................................356
Chlorine Dependence ...................................................... 183
Oxidation Models ...............................................................173
Doping Dependence ....................................................... 184
Analytical ......................................................................188
Orientation Dependence .................................................. 181
Compress .....................................................................176
Pressure Dependence ..................................................... 182
Linear Rate Constant ......................................................179
Lithography ...................................................................... 277 Mixed Ambient ...............................................................187
LPCVD ............................................................................ 245 Numerical .....................................................................175
Parabolic Rate Constant ..................................................186
M Viscous ........................................................................177
MaskViews Oxidation Simulation Recommendations .................................188
Generating Masks in Athena ............................................... 99 Growing Thin Oxides .......................................................188
Initial Rectangular Grid ...................................................... 92 Implantation Through Thermally-Grown Oxides and Dopant
Loss During Subsequent Annealing ....................................190
Medium Injection
Oxidation Enhanced Diffusion (OED) ..........................131, 191
Bandgap Narrowing Effects ................................................ 78
Oxidation Retarded Diffusion (ORD) ...................................191
Poly-emitter work function .................................................. 77
Modified Gaussian Function (MGF) ........................................ 205
P
Monte Carlo Etching Model
Incoming Ions and Neutrals .............................................. 250 Pair Diffusion ....................................................................127
Ion and Neutral Fluxes .................................................... 251 Parabolic Rate Constant
Polymer Fluxes .............................................................. 253 Chlorine Dependence ......................................................187
Rates .......................................................................... 253 Pressure Dependence .....................................................186
Monte Carlo Implant Models Pearson Differential Equation ...............................................196
Amorphous Material ........................................................ 217 Photoresist Bake ................................................................273
Crystalline Material ......................................................... 217 Physically-Based Simulation ...................................................16
Damage Accumulation Model ............................................ 212
Plasma Etching Model
Electronic Stopping ......................................................... 210
Dopant Enhanced Etching ................................................249
Implantation Geometry .................................................... 212
Implantation Geometry in SiC .................................... 214–216 PLS Diffusion Model Examples .............................................166
Interatomic Potential ....................................................... 209 PLS Diffusion Models ..........................................................151
Nuclear Stopping ........................................................... 208 point defect .......................................................................127
Physical Problems .......................................................... 207 Proximity Printing
Solution ....................................................................... 208 Lithography ...................................................................277
Statistical Sampling ........................................................ 218 Simulation Method ..........................................................279
Monte Carlo Implant Module ................................................. 328 Theory of ..............................................................278–279

R T
Rapid Thermal Anneals (RTA) Technical Support ................................................................12
Epitaxy ...........................................................................65 TonyPlot ..........................................................................401
Notes .............................................................................62 Trap Equation ...................................................................136
Oxidation ........................................................................63
Two Dimensional Model
REFLOW Model ................................................................255 Defect Diffusion ..............................................................143
RTA Diffusion Modeling .......................................................145 Dislocation Loop Based Bulk Recombination ........................142
Dopants .......................................................................134
S Interstitial Generation ......................................................138
Interstitials ....................................................................135
Second Order Fick’s Equation ...............................................129
Recombination at Interfaces ..............................................138
SiGe Process Simulation Vacancies .....................................................................141
Deposit ..........................................................................80
Two-Dimensional Implant Profiles
DIFFUSE ................................................................81, 307
Convolution Method ........................................................202
MATERIAL ....................................................................349
Depth-Independent Lateral Distribution ................................203
Material ..........................................................................80
Gaussian Lateral Distribution Function ................................203
METHOD ................................................................80, 352
Implant Parameters in the Moments Statement .....................206
SiGe/SiGeC Simulation .......................................................228 Lateral Standard Deviation ................................................204
Silicidation Model ...............................................................193 Non-Gaussian Lateral Distribution Functions .........................205
Simulation ........................................................................279 Parabolic Approximation of Depth-Dependent Lateral Distribution ....
Solid Solubility ...................................................................413 204
Using PRINT.MOM for Extraction of Spatial Moments .............207
SSUPREM4 ......................................................................126
Calibrating ATHENA for a Typical Bipolar Process Flow ............74
Changing the Method Statement During the Process Flow .........54 U
Choosing an Appropriate Model ...........................................53 userimp ............................................................................206
Modelling the Correct Substrate Depth ..................................55
Using Athena With Other Silvaco Software .................................15
Multiple Models ................................................................52
process steps ..................................................................52
Rapid Thermal Anneals (RTA) .............................................62 V
SiGe Process Simulation ....................................................80 vacancy flux expression .......................................................141
SSuprem4 ..........................................................................52
Standard Examples ..............................................................20
Statistical Sampling
rare event trajectory splitting technique ................................218
rare events ....................................................................218
restart ..........................................................................218
trajectory splitting ...........................................................218
std-tables .........................................................................206
Stress Models ...........................................................230–233
Diffusion Enhancement/Retardation ....................................233
Dopant-Induced Intrinsic ..................................................232
History .................................................................233, 358
String Algorithm .................................................................236
Structure Exposure
Development .................................................................115
Post Development Bake ...................................................115
Post Exposure Bake ........................................................115
Structure Manipulation Tools
ATHENA In 1D Mode ........................................................84
Stretch ...........................................................................82
Structure FLIP .................................................................82
Surface Recombination
CNET ..........................................................................159
SVDP implant models, see Analytic Implant Models ...................198

Silvaco TCAD Athena User Manual

Uploaded by

Copyright:

Available Formats

Silvaco TCAD Athena User Manual

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Silvaco TCAD Athena User Manual

Uploaded by

Copyright:

Available Formats

What is the document about?

What is the document about?

What software does it mention?

What software does it mention?

Athena

The information contained in this document is subject to change without notice.

2 Athena User’s Manual

Font Style/Convention Description Example

• This represents a list of items or • Bullet A

3 Athena User’s Manual

4 Athena User’s Manual

2.6.5 The Base Current Profile – Low Injection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

5 Athena User’s Manual

3.5 Ion Implantation Models. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195

6 Athena User’s Manual

7 Athena User’s Manual

6.12 CPULOG. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 301

8 Athena User’s Manual

6.62 STRIP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397

9 Athena User’s Manual

C.7 Changes in the INITIALIZE statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424

10 Athena User’s Manual

1.1 Athena Overview

1.1.1 Using This Manual

1.1.2 Technical Support

12 Athena User’s Manual

1.2 Athena Features and Capabilities

Table 1-1 Athena Features and Capabilities

Bake • Time and temperature bake specification.

13 Athena User’s Manual

Table 1-1 Athena Features and Capabilities

Etch • Extensive geometric etch capability.

14 Athena User’s Manual

Table 1-1 Athena Features and Capabilities

Implantation • Experimentally verified Pearson and dual Pearson analytical

1.2.1 Using Athena With Other Silvaco Software

15 Athena User’s Manual

1.2.2 The Value Of Physically-Based Simulation

16 Athena User’s Manual

2.1 Getting Started

2.1.1 Running Athena Under DeckBuild

18 Athena User’s Manual

Figure 2-1 Main DeckBuild Window

19 Athena User’s Manual

2.1.2 Loading And Running Athena Standard Examples

Note: To access the examples in PC, select HelpExamples.

20 Athena User’s Manual

Figure 2-2 DeckBuild Examples Window

21 Athena User’s Manual

22 Athena User’s Manual

2.2 Operation Modes

2.2.1 Interactive Mode With DeckBuild

2.2.2 Batch Mode With DeckBuild

23 Athena User’s Manual

2.2.3 No Windows Batch Mode With DeckBuild

2.2.4 Running Athena inside DeckBuild

24 Athena User’s Manual

2.2.5 Running Athena In Standalone Mode (Without DeckBuild)

2.2.6 Running Athena 1D (One-Dimensional Version of Athena)

25 Athena User’s Manual