Advanced Calibration for

Process Simulation User

Guide
Version P-2019.03, March 2019
Copyright and Proprietary Information Notice
© 2019 Synopsys, Inc. This Synopsys software and all associated documentation are proprietary to Synopsys, Inc. and may only be
used pursuant to the terms and conditions of a written license agreement with Synopsys, Inc. All other use, reproduction,
modification, or distribution of the Synopsys software or the associated documentation is strictly prohibited.
Destination Control Statement
All technical data contained in this publication is subject to the export control laws of the United States of America.
Disclosure to nationals of other countries contrary to United States law is prohibited. It is the reader’s responsibility to
determine the applicable regulations and to comply with them.
Disclaimer
SYNOPSYS, INC., AND ITS LICENSORS MAKE NO WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, WITH
REGARD TO THIS MATERIAL, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
Trademarks
Synopsys and certain Synopsys product names are trademarks of Synopsys, as set forth at
https://www.synopsys.com/company/legal/trademarks-brands.html.
All other product or company names may be trademarks of their respective owners.
Free and Open-Source Licensing Notices
If applicable, Free and Open-Source Software (FOSS) licensing notices are available in the product installation.
Third-Party Links
Any links to third-party websites included in this document are for your convenience only. Synopsys does not endorse
and is not responsible for such websites and their practices, including privacy practices, availability, and content.

www.synopsys.com

ii Advanced Calibration for Process Simulation User Guide

P-2019.03
 Contents
About This Guide xiii
Related Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv
Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiv
Customer Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Accessing SolvNet. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Contacting Synopsys Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Contacting Your Local TCAD Support Team Directly. . . . . . . . . . . . . . . . . . . . . . . . xv

Part I Advanced Calibration in Sentaurus Process 1

Chapter 1 Using Advanced Calibration File of Sentaurus Process 3

Location of Advanced Calibration File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Using Advanced Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Optional Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Additional Calibration by Users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Earlier Versions of Advanced Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Advanced Calibration File for Kinetic Monte Carlo Simulations . . . . . . . . . . . . . . . . . . . 8
Sentaurus Workbench Splits: Saving in TDR Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

Chapter 2 Advanced Calibration for Silicon, SiGe, and Germanium 9

Part 1: Basic Model Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Diffusion Models in Silicon and Germanium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Dopant Cluster Models in Silicon and Germanium. . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Defect Cluster Models in Silicon and Germanium . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
Poisson Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Channeling Dose in Analytic Implantations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
Numeric Solver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Summary of Model Switches. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Part 2: Constant Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
Basic Point-Defect Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Bulk Parameters for Free Interstitials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Bulk Parameters for Free Vacancies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Bulk Recombination of Point Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Boundary Conditions for Point Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
Oxidation-Enhanced Diffusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

Advanced Calibration for Process Simulation User Guide iii

P-2019.03
Contents

Transient-Enhanced Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Boron Diffusion and Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Boron Diffusion Coefficient. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
Effect of Fluorine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Boron Clustering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Boron Dose Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
Fluorine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Arsenic Diffusion and Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Arsenic Diffusivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Arsenic Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Arsenic Dose Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
Phosphorus Diffusion and Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Phosphorus Diffusivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Phosphorus Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
Phosphorus Dose Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Arsenic–Phosphorus Co-Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Indium Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Antimony Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Carbon Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Carbon Diffusivity in Silicon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Carbon Clustering in Silicon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
Carbon in Germanium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Intrinsic Carrier Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Massoud Model Parameters for Wet Oxidation of Silicon. . . . . . . . . . . . . . . . . . . 36
Parameters for Wet and Dry Oxidation of SiGe. . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Smoothing of Amorphous–Crystalline Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Selecting Implantation Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Effect of Germanium and Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
Arguments of SiGe_and_Stress_Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Ge Chemical Effect (Ge_Chem_Eff) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Stress Effect (Stress_Eff) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Segregation at Si–SiGe Interface (Segreg_Model) . . . . . . . . . . . . . . . . . . . . . . . . 40
Effect of Strained Overlayers (Strained_Overlayer) . . . . . . . . . . . . . . . . . . . . . . . 40
Implementation of SiGe_and_Stress_Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Chemical SiGe Alloy Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Stress Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Segregation at Si–SiGe Interface (Segreg_Model) . . . . . . . . . . . . . . . . . . . . . . . . 53
Effect of Strained Overlayers (Strained_Overlayers). . . . . . . . . . . . . . . . . . . . . . . 55
Part 3: Initial Conditions After Ion Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
User-Defined Defect Initialization. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

iv Advanced Calibration for Process Simulation User Guide

P-2019.03
 Contents

impPostProcess_AdvCal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Scaling Factors for Point Defects and Damage . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Values for Initial Dopant Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Sum of As-Implanted Point Defects and Crystal Damage . . . . . . . . . . . . . . . . . . . 60
Subroutines for Setting ifactor and dfactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
ifactor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
vfactor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
dfactor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Thermal Implantations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Scope of Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Damage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Initial Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Implantation Preprocessing and Postprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Analytic Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Monte Carlo Implantation (General) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
Carborane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Part 4: Comprehensive and Slow Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
Interstitial Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Boron–Interstitial Clusters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Arsenic Parameters in AdvancedModels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Phosphorus Parameters in AdvancedModels. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
ChargedCluster Model for Indium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Fluorine Diffusion and Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
Carbon Diffusion and Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Nitrogen Diffusion and Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Solid Phase Epitaxial Regrowth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
Recrystallization Speed . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
Doping Redistribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Melting Laser Anneal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
Settings for Simulation Optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Meshing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Thermodynamics of Silicon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Silicon Absorptivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
MLA Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Limitations of MLA Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Part 5: Accelerating Simulations for Power Technologies . . . . . . . . . . . . . . . . . . . . . . . . 89
Using AdvancedPowerDeviceMode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Variant 1: Explicitly Reverting to Standard Models . . . . . . . . . . . . . . . . . . . . . . . 90
Variant 2: Automatically Reverting to Standard Models . . . . . . . . . . . . . . . . . . . . 91
Contents of AdvancedPowerDeviceMode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Increased Time Steps, Deposition Steps, and Temperature Steps . . . . . . . . . . . . . 91

Advanced Calibration for Process Simulation User Guide v

P-2019.03
Contents

Simplified Physics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Speedup Methods Not Included in AdvancedPowerDeviceMode . . . . . . . . . . . . . 93
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93

Chapter 3 Guidelines for Additional Calibration 107

Accuracy and Limitations of Advanced Calibration of Sentaurus Process . . . . . . . . . . 107
Error Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Point Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Bulk Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
Surface Boundary Conditions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Oxidation-Enhanced Diffusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Clusters of Interstitials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
Vacancy Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Boron Diffusion and Clustering. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Diffusion and Pairing in Silicon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Hopping Length (for ChargedReact Model) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
Effect of Fluorine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
Boron Clustering and Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Boron Dose Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
Arsenic Diffusion and Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
Phosphorus Diffusion and Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Phosphorus Diffusion in Silicon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Phosphorus Activation in Silicon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Phosphorus Dose Loss at Oxide–Silicon Interfaces . . . . . . . . . . . . . . . . . . . . . . . 123
Co-Diffusion of Arsenic and Phosphorus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Indium Diffusion and Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Nonamorphizing Condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Amorphizing Ion Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
SPER Model Usage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Antimony Diffusion and Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Carbon–Interstitial Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Carbon–Boron Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Impact of Carbon on Hole Mobility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Molecular Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
Analytic Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Sentaurus MC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Fluorine Diffusion and Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Nitrogen Diffusion and Clustering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
Diffusion in Strained Silicon and Silicon Germanium . . . . . . . . . . . . . . . . . . . . . . . 138
As-Implanted Dopant Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

vi Advanced Calibration for Process Simulation User Guide

P-2019.03
 Contents

Coimplantation Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140

Preamorphization Implantation Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Cold and Hot Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Amorphization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Channeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
Dose Loss at Silicon–Oxide Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
Calibration of ThreePhaseSegregation Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
Oxidation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
Bird’s Beak in CMOS Devices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
Diffusion in Polysilicon and Out-Diffusion From Polysilicon . . . . . . . . . . . . . . . . . 146
Dopant Penetration Through Gate Oxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Diffusion and Activation in Germanium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Performing Additional Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
Calibration Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Loading a User Calibration File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Lateral Diffusion Along Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
Example of a User Calibration File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
Recommended Numeric Settings for Monte Carlo Implantation . . . . . . . . . . . . . . . . . . 152
Monte Carlo Pocket Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
MC Source/Drain Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Additional Calibration for Power Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Meshing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Recommendations for Bulk Refinement With Adaptive Meshing. . . . . . . . . . . . 156
Refinement at Material Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
Calibration of Wet Etching Rate Modification by Ion Implantation . . . . . . . . . . . . . . . 158
Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Application. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

Chapter 4 Advanced Calibration for 4H-SiC Process Simulation 165

Content of Advanced Calibration File for 4H-SiC Simulation. . . . . . . . . . . . . . . . . . . . 165
Part 1: Basic Model Switches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Poisson Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Dopant Cluster Models in 4H-SiC. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
Dopant Transport at the Oxide–SiC Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Boron Evaporation From the SiC Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
Part 2: Settings for Monte Carlo Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
Impact of Wafer Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

Advanced Calibration for Process Simulation User Guide vii

P-2019.03
Contents

Recommendations for Fine-Tuning of Amorphization . . . . . . . . . . . . . . . . . . . . . . . 172

Illustration of Calibration Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
Part 3: Initial Conditions After Ion Implantation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

Part II Advanced Calibration in Sentaurus Process Kinetic Monte Carlo 177

Chapter 5 Using Advanced Calibration File of Sentaurus Process KMC 179

Location of Advanced Calibration File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Using Advanced Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
Additional Calibration by Users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

Chapter 6 Contents of Advanced Calibration of Sentaurus Process KMC 183

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Supported Materials. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
Part 1: Model Parameters for Implantation Damage and Point Defects . . . . . . . . . . . . . 185
Amorphization and Recrystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Amorphous Silicon and Germanium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Diffusion, Generation, and Recombination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
Charge States . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
Extended Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
SiGe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Linear Germanium Correction Factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Silicon and Germanium Parameter Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . 193
Stress Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Amorphization and Recrystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Band Gap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Diffusion, Generation, and Recombination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Extended Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
Part 2: Model Parameters for Impurities. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
Boron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
Arsenic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Phosphorus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
Indium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Fluorine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
Nitrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
Clusters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
Boron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208

viii Advanced Calibration for Process Simulation User Guide

P-2019.03
 Contents

Arsenic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
Phosphorus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
Arsenic and Phosphorus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
Indium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Boron and Carbon. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Fluorine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
Nitrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
Segregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
Boron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Arsenic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Phosphorus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Indium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
Fluorine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
Nitrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
Recrystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
Boron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
Arsenic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
Phosphorus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
Indium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
Fluorine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
Nitrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Epitaxy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Boron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Arsenic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Phosphorus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
Carbon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
SiGe . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
Linear Germanium Correction Factors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
Silicon and Germanium Parameter Interpolation . . . . . . . . . . . . . . . . . . . . . . . . . 219
Stress Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
Boron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Arsenic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
Phosphorus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
Indium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Section 4: Model Parameters for Epitaxial Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Coordinations.Planes Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Coordinations Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224
Coordinations.Reactions Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

Advanced Calibration for Process Simulation User Guide ix

P-2019.03
Contents

Epitaxial Growth. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225

Etching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228

Chapter 7 Guidelines for Additional Calibration 239

Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC . . . . . 239
Damage and Point Defects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
Amorphization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
Recrystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Diffusion, Generation, and Recombination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Extended Defects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
Impurities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Diffusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
Activation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
Dose Loss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Recrystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
Stress and SiGe Effects. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 248
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

Part III Advanced Calibration for Mechanics Simulations 253

Chapter 8 Using Advanced Calibration File for Mechanics Simulations 255

Location of Advanced Calibration File. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
Using Advanced Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
Earlier Versions of Advanced Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256

Chapter 9 Contents of Advanced Calibration for Mechanics Simulations 259

Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
Switches for Interpolation in Mole Fraction–Dependent Mechanical Models . . . . . 259
Parameters for Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
Suppression of Dilatational Viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
Mole Fraction–Dependent Mechanics Parameters for SiGe . . . . . . . . . . . . . . . . . . . 261
Cubic Crystal Anisotropy for Silicon and Germanium . . . . . . . . . . . . . . . . . . . . . . . 262
Temperature Dependency of Stiffness Coefficients for Silicon and Germanium . . . 262
Isotropic Elastic Moduli for Germanium. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
Polysilicon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
Viscosity of Oxide and Nitride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
Isotropic Moduli for Titanium and Titanium Silicide . . . . . . . . . . . . . . . . . . . . . . . . 264
Amorphous Germanium Oxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

x Advanced Calibration for Process Simulation User Guide

P-2019.03
 Contents

Titanium Nitride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264

Hafnium Oxide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
Silicon Carbide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266

Advanced Calibration for Process Simulation User Guide xi

P-2019.03
Contents

xii Advanced Calibration for Process Simulation User Guide

P-2019.03
 About This Guide
Synopsys is working continually on improving the simulation models and optimizing the
model parameters for the latest technology nodes. This effort is based on long-standing
experience of model calibration for customers and a comprehensive, growing database of
state-of-the-art secondary ion mass spectroscopy (SIMS) profiles. The variety of partners and
data ensures that systematic and random errors in experimental work are minimized in this
model representation. Advanced Calibration provides users with a set of parameters that have
been calibrated to deep-submicron CMOS technology, including ultrashallow junction (USJ)
formation, surface dose loss, and channel and halo dopant implantation and diffusion.

Sentaurus™ Process offers the Tcl-based scripting language Alagator for the implementation
of diffusion and reaction models. This allows users to implement models or to model
extensions. This possibility is also used in the Advanced Calibration of Sentaurus Process: The
Advanced Calibration file of Sentaurus Process contains model selections, parameter
specifications, and some model extensions. This file can be sourced at the beginning of a
Sentaurus Process simulation. In analogy, the Advanced Calibration file of Sentaurus Process
Kinetic Monte Carlo contains model selections and parameter specifications. This file can be
sourced at the beginning of a Sentaurus Process simulation in atomistic mode as well.

The Advanced Calibration of Sentaurus Interconnect is based on the Advanced Calibration of

Sentaurus Process, in particular, the part for mechanics simulations. This file can be sourced at
the beginning of a Sentaurus Interconnect simulation.

Current and future efforts of Synopsys are focused on the integration of the Advanced
Calibration in the process simulators Sentaurus Process, Sentaurus Process Kinetic Monte
Carlo, and Sentaurus Interconnect, and on further improvements of its accuracy.

This user guide explains the Advanced Calibration files for the process simulators Sentaurus
Process, Sentaurus Process Kinetic Monte Carlo, and Sentaurus Interconnect. It is intended for
users who are familiar with Sentaurus Process and want to obtain a higher accuracy in process
simulation. For detailed information about these process simulators, refer to the Sentaurus™
Process User Guide and the Sentaurus™ Interconnect User Guide.

The user guide is divided into the following parts:

■ Part I: Advanced Calibration in Sentaurus Process
The chapters in this part describe the contents and the use of the Advanced Calibration file
of Sentaurus Process. They describe the use of Advanced Calibration for silicon, silicon
germanium, germanium, and silicon carbide. In addition, they explain the accuracy and
limitations of the Advanced Calibration of Sentaurus Process and provides guidelines for
additional calibration.

Advanced Calibration for Process Simulation User Guide xiii

P-2019.03
About This Guide
Related Publications

■ Part II: Advanced Calibration in Sentaurus Process Kinetic Monte Carlo

The chapters in this part describe the contents and the use of the Advanced Calibration file
of Sentaurus Process Kinetic Monte Carlo. They explain the accuracy and limitations of the
Advanced Calibration of Sentaurus Process Kinetic Monte Carlo and provides guidelines
for additional calibration.
■ Part III: Advanced Calibration for Mechanics Simulations
The chapters in this part describe the contents and the use of the Advanced Calibration file
for mechanics simulations for Sentaurus Process and Sentaurus Interconnect.

Related Publications
For additional information, see:
■ The TCAD Sentaurus release notes, available on the Synopsys SolvNet® support site (see
Accessing SolvNet on page xv).
■ Documentation available on SolvNet at https://solvnet.synopsys.com/DocsOnWeb.

Conventions
The following conventions are used in Synopsys documentation.

Convention Description

Blue text Identifies a cross-reference (only on the screen).

Bold text Identifies a selectable icon, button, menu, or tab. It also indicates the name of a field or an
option.

Courier font Identifies text that is displayed on the screen or that the user must type. It identifies the
names of files, directories, paths, parameters, keywords, and variables.

Italicized text Used for emphasis, the titles of books and journals, and non-English words. It also
identifies components of an equation or a formula, a placeholder, or an identifier.

Key+Key Indicates keyboard actions, for example, Ctrl+I (press the I key while pressing the Control
key).

Menu > Command Indicates a menu command, for example, File > New (from the File menu, select New).

xiv Advanced Calibration for Process Simulation User Guide

P-2019.03
 About This Guide
Customer Support

Customer Support
Customer support is available through the Synopsys SolvNet customer support website and by
contacting the Synopsys support center.

Accessing SolvNet
The SolvNet support site includes an electronic knowledge base of technical articles and
answers to frequently asked questions about Synopsys tools. The site also gives you access to
a wide range of Synopsys online services, which include downloading software, viewing
documentation, and entering a call to the Support Center.

To access the SolvNet site:

1. Go to the web page at https://solvnet.synopsys.com.
2. If prompted, enter your user name and password. (If you do not have a Synopsys user name
and password, follow the instructions to register.)

If you need help using the site, click Help on the menu bar.

Contacting Synopsys Support

If you have problems, questions, or suggestions, you can contact Synopsys support in the
following ways:
■ Go to the Synopsys Global Support Centers site on synopsys.com. There you can find
e-mail addresses and telephone numbers for Synopsys support centers throughout the
world.
■ Go to either the Synopsys SolvNet site or the Synopsys Global Support Centers site and
open a case online (Synopsys user name and password required).

Contacting Your Local TCAD Support Team Directly

Send an e-mail message to:
■ [email protected] from within North America and South America
■ [email protected] from within Europe
■ [email protected] from within Asia Pacific (China, Taiwan, Singapore,
Malaysia, India, Australia)

Advanced Calibration for Process Simulation User Guide xv

P-2019.03
About This Guide
Customer Support

■ [email protected] from Korea

■ [email protected] from Japan

xvi Advanced Calibration for Process Simulation User Guide

P-2019.03
Part I Advanced Calibration in Sentaurus
Process

This part of the Advanced Calibration for Process Simulation User Guide contains the
following chapters:

Chapter 1 Using Advanced Calibration File of Sentaurus Process on page 3

Chapter 2 Advanced Calibration for Silicon, SiGe, and Germanium on page 9

Chapter 3 Guidelines for Additional Calibration on page 107

Chapter 4 Advanced Calibration for 4H-SiC Process Simulation on page 165

CHAPTER 1 Using Advanced Calibration File of
Sentaurus Process

This chapter gives a brief introduction to the use of Advanced

Calibration in a process simulation with Sentaurus Process.

Advanced Calibration is a selection of models and parameters, which is recommended by

Synopsys to be used for accurate process simulation. In Sentaurus Process, this selection of
models and parameters is contained in a text file, which can be opened with any standard text
editor.

By sourcing the Advanced Calibration file at the beginning of a process simulation, the
standard calibration of Synopsys is selected. If needed, you can change or extend the Advanced
Calibration. This can be performed by sourcing an additional calibration file, which contains
the required parameter changes, or by editing the Advanced Calibration file with a text editor.

Location of Advanced Calibration File

The Advanced Calibration file is the ultimate product of Synopsys’ calibration efforts. For each
release of Sentaurus Process, there is a new Advanced Calibration file that includes the best
and latest set of models and parameters. To ensure backward compatibility, previous Advanced
Calibration files are still available.

The files for the Advanced Calibration of Sentaurus Process in this release are located at:
$STROOT/tcad/$STRELEASE/lib/sprocess/TclLib/AdvCal

The default file is named AdvCal_2019.03.fps. It represents the first version of Advanced
Calibration P-2019.03. Older versions of the Advanced Calibration file can be found in the
same directory. For example, the file AdvCal_2018.06.fps contains the Advanced
Calibration file for Version O-2018.06 and is available for backward compatibility.

The default Advanced Calibration parameter file contains parameters for Si, SiGe and Ge-
based processes. An additional parameter file AdvCal_SiC_2019.03.fps targets process
simulation of semiconductor devices based on 4H-SiC.

Advanced Calibration for Process Simulation User Guide 3

P-2019.03
1: Using Advanced Calibration File of Sentaurus Process
Using Advanced Calibration

Using Advanced Calibration

To use the Advanced Calibration of Sentaurus Process, at the beginning of the input file, insert
the line:
AdvancedCalibration

or, better:
AdvancedCalibration 2019.03

Alternatively, this file can be sourced by using:

source $AdvCalDir/AdvCal_2019.03.fps

The procedure AdvancedCalibration has two optional parameters <version> and

<material>. The order of arguments is not important. The allowed values for <material>
are Si, SiGe, Ge, 4H-SiC, and SiC:
■ Si, SiGe, and Ge as well as no material value will call the default Advanced Calibration
parameter file for Si, SiGe, and Ge materials.
■ 4H-SiC and SiC will call the Advanced Calibration file for 4H-SiC.

For example, the following command calls the Advanced Calibration file for 4H-SiC
(AdvCal_SiC_2019.03.fps):
AdvancedCalibration 2019.03 4H-SiC

Optional Modules
AdvCal_2019.03.fps includes a base set of models and parameters, and several optional
modules, which are not switched on automatically. Each module can be selected by a single
command after loading Advanced Calibration.

Three modules are useful for selected technologies, and their corresponding commands are as
follows:
■ SiGe_and_Stress_Effect switches on the impact of Ge and stress on dopant diffusion
and activation. It is recommended for PMOS devices with SiGe pockets and also can be
used for HBT devices with SiGe layers. See Effect of Germanium and Stress on page 38.
■ AdvancedPowerDeviceMode is used to speed up the process simulation for many types
of power device. For simulation of power devices, see Part 5: Accelerating Simulations for
Power Technologies on page 89 and Additional Calibration for Power Technologies on
page 155.

4 Advanced Calibration for Process Simulation User Guide

P-2019.03
 1: Using Advanced Calibration File of Sentaurus Process
Additional Calibration by Users

■ AdvancedMLAModel is recommended for modeling melting laser anneal. See Melting

Laser Anneal on page 82.

Other modules switch on physical models that are more complex and more time-consuming
alternatives to the Advanced Calibration default models:
■ AdvancedFluorineModel models the impact of fluorine on transient-enhanced
diffusion and on boron dose loss. It is recommended for processes that include atomic F
implantations. It also can be considered for use in processes including high-dose BF2
implantations. See Fluorine Diffusion and Clustering on page 74.
■ AdvancedNitrogenModel switches on equations for the diffusion and clustering of
nitrogen in silicon. It can be considered for use in processes that include nitrogen
implantations. See Nitrogen Diffusion and Clustering on page 75.
■ AdvancedModels is used to switch on complex physical models for various clustering
phenomena in silicon, such as interstitial clusters, boron–interstitial clusters, and fluorine
clusters. AdvancedModels is used mainly for the purpose of fundamental research. See
Part 4: Comprehensive and Slow Models on page 68.
■ AdvancedSPERModel switches on the solid phase epitaxial regrowth (SPER) model and
sets the calibrated parameters for it. AdvancedSPERModel adds complexity to the
simulation of anneals after amorphizing implantations1. See Solid Phase Epitaxial
Regrowth on page 76.

Additional Calibration by Users

Advanced Calibration is based on the assumption that all parameters that are not changed in
the parameter files are the default parameters of Sentaurus Process. To use the Advanced
Calibration file AdvCal_2019.03.fps, it must be sourced before the real process
description.

After sourcing AdvCal_2019.03.fps, you can change the model switches or parameter
values of the physical models. This should ideally be performed by experienced users with a
good understanding of the diffusion models of Sentaurus Process.

1. After amorphizing implantations, the amorphized regions recrystallize by SPER in subsequent thermal anneals. By
default, Sentaurus Process assumes that SPER is completed instantaneously at the beginning of thermal annealing.
Amorphization is only taken into account by setting special initial conditions for dopants and point defects in
amorphized semiconductor regions at the beginning of diffusion. The SPER model simulates the movement of the
amorphous–crystalline transition region during SPER and the dopant redistribution during such a process. In
particular, the SPER model simulates diffusion in amorphous silicon and the snow plow effect during SPER. The
snow plow effect is the redistribution of certain impurity species (for example, indium), which prefer to stay at the
amorphous side of the amorphous–crystalline interface, towards the surface.

Advanced Calibration for Process Simulation User Guide 5

P-2019.03
1: Using Advanced Calibration File of Sentaurus Process
Earlier Versions of Advanced Calibration

For the process simulation of silicon technology, Advanced Calibration is usually the best
starting point. You can further increase the accuracy for a certain technology by additional fine-
tuning of a few physical parameters.

The best way to perform this is to put all additional calibration in a user calibration file, for
example, my_calibration.fps. This file includes the commands to select optional modules
of Advanced Calibration such as AdvancedModels or AdvancedPowerDeviceMode, and it
includes all project-specific changes to the physical models or parameters with respect to
Advanced Calibration.

In the process simulation file, at the beginning of the process simulation, insert the lines:
AdvancedCalibration 2019.03
source ./my_calibration.fps

This approach allows you to:

■ Separate completely the calibration and the process description.
■ Use the Advanced Calibration file as a starting point.
■ Summarize all project-specific calibration in a short and clear text file.

Detailed information about how to perform additional calibration is given in Chapter 3 on

page 107.

Earlier Versions of Advanced Calibration

You can source earlier versions of the Advanced Calibration file by inserting, for example, the
line:
AdvancedCalibration 2018.06

This is converted internally to:

source $AdvCalDir/AdvCal_2018.06.fps

Table 1 on page 7 lists earlier versions of the Advanced Calibration file that can be loaded with
Sentaurus Process Version P-2019.03.

This possibility is available to provide backward compatibility. You can run simulations with
the latest version of Sentaurus Process, but the simulations can still be based on an old
calibration. For new TCAD projects, it is recommended to load the latest version of Advanced
Calibration.

6 Advanced Calibration for Process Simulation User Guide

P-2019.03
 1: Using Advanced Calibration File of Sentaurus Process
Earlier Versions of Advanced Calibration

Table 1 Earlier versions of Advanced Calibration file and their corresponding commands
Advanced Calibration file Corresponding command

AdvCal_2018.06.fps AdvancedCalibration 2018.06

AdvCal_2017.09.fps AdvancedCalibration 2017.09

AdvCal_2016.12.fps AdvancedCalibration 2016.12

AdvCal_2016.03.fps AdvancedCalibration 2016.03

AdvCal_2015.06.fps AdvancedCalibration 2015.06

AdvCal_2014.09.fps AdvancedCalibration 2014.09

AdvCal_2013.12.fps AdvancedCalibration 2013.12

AdvCal_2013.03.fps AdvancedCalibration 2013.03

AdvCal_2012.06.fps AdvancedCalibration 2012.06

AdvCal_2011.09.fps AdvancedCalibration 2011.09

AdvCal_2010.12.fps AdvancedCalibration 2010.12

AdvCal_2010.03.fps AdvancedCalibration 2010.03

AdvCal_2009.06.fps AdvancedCalibration 2009.06

AdvCal_2008.09.fps AdvancedCalibration 2008.09

AdvCal_2007.12.fps AdvancedCalibration 2007.12

AdvCal_2007.03.fps AdvancedCalibration 2007.03

AdvCal_2006.06.fps AdvancedCalibration 2006.06

AdvCal_2005.10.fps AdvancedCalibration 2005.10

The original versions of the earlier Advanced Calibration files cannot be used in the latest
version of Sentaurus Process, due to changes in the source code and the model library of
Sentaurus Process, which affect the functionality of the old files. Therefore, Synopsys has
adapted the earlier Advanced Calibration files to cope with those changes. Modifications have
been undertaken in such a way that the choice of physical models and parameters is still the
one from the corresponding release. The AdvancedCalibration command always loads the
modified versions.

Most earlier versions of Advanced Calibration contain the Compatibility command. For
example, AdvCal_2018.06.fps contains the command Compatibility 2018.06, which
applies the default parameters and model settings of Sentaurus Process Version O-2018.06
before setting the Advanced Calibration models and parameters.

As an exception, AdvCal_2010.12.fps does not contain the Compatibility 2010.12

command. This command switches off the model switch pdbSet Mechanics

Advanced Calibration for Process Simulation User Guide 7

P-2019.03
1: Using Advanced Calibration File of Sentaurus Process
Advanced Calibration File for Kinetic Monte Carlo Simulations

Total.Concentration.Model 1, which was introduced in Sentaurus Process Version

F-2011.09 and is considered an important improvement for stress calculation. Users who want
a more complete backward compatibility with Version E-2010.12 must insert the command
Compatibility 2010.12 before loading AdvCal_2010.12.fps.

Advanced Calibration File for Kinetic Monte Carlo

Simulations
An Advanced Calibration file is available for simulations with the kinetic Monte Carlo mode
of Sentaurus Process. The use and contents of this file are described in Chapter 5 on page 179
and Chapter 6 on page 183.

Sentaurus Workbench Splits: Saving in TDR Format

Sentaurus Process can be used within Sentaurus Workbench projects. If split commands of
Sentaurus Workbench are used inside the input file of Sentaurus Process, at each split
command, the structure with all the data fields is saved in a TDR format file. In a subsequent
tool instance, the process simulation starts by loading the previously saved structure.

Sentaurus Process does not always save and load the complete status of the process simulation.
In particular, the definitions of Alagator terms and solution commands are only saved in TDR
format if the keyword store is used in the term definition command lines, and Tcl procedures
are only saved if they have been defined by the command fproc (rather than proc).
Furthermore, entries in the parameter database are only saved if the TDR format is used for
saving and loading.

The Advanced Calibration file contains the definitions of terms and procedures. In the latest
file, the definitions of terms, solutions, and procedures are performed in such a way that they
are saved to the TDR format and are reloaded. No additional attention is needed.

In an earlier version of the file (AdvCal_2010.03.fps), a series of terms inside the procedure
SiGe_and_Stress_Effect was modified with the command MultiplyTerm. Since
AddToTerm, SubFromTerm, and MultiplyTerm modify terms without the keyword store,
the corresponding term modifications were not stored in the TDR format. Therefore, the
command MultiplyTerm is no longer used in newer versions of Advanced Calibration.

In early versions of the Advanced Calibration files (AdvCal_2007.12.fps and before), the
definitions of terms and procedures are not saved. Therefore, when using these versions of
Advanced Calibration files, you must ensure that the files are loaded at the beginning of each
part of a split process simulation.

8 Advanced Calibration for Process Simulation User Guide

P-2019.03
CHAPTER 2 Advanced Calibration for Silicon,
SiGe, and Germanium

This chapter explains how to use Advanced Calibration of Sentaurus

Process for silicon, silicon germanium (SiGe), and germanium and
documents the origin of the parameter values.

The focus of Advanced Calibration is monocrystalline silicon, germanium, and SiGe for all Ge
mole fractions. The calibration for silicon and SiGe with low Ge mole fraction ( ≤ 0.5 ) is the
most mature and reliable. On the other hand, the one for pure Ge is less mature and reliable,
and the one for SiGe with high Ge mole fraction (> 0.5) is the least mature and reliable. Many
model equations and model parameters are taken from reliable publications. In addition, a
rigorous calibration has been performed by Synopsys, based on a SIMS database.

The book Intrinsic Point Defects, Impurities, and Their Diffusion in Silicon by Pichler [1] is a
good reference source for parameter values. It refers to more than 3000 scientific papers and
gives a comprehensive overview of the experimental data available for the calibration of
fundamental parameters for diffusion in silicon. For many relevant parameters, Pichler
compares the results from many authors, which can be used to estimate the error bars of the
parameter values.

The Advanced Calibration file AdvCal_2019.03.fps is divided into the following parts,
which contain numbered sections and are executed in sequence:
■ Part 1: Basic model switches
■ Part 2: Constant parameters
■ Part 3: Initial conditions after ion implantation
■ Part 4: Comprehensive and slow models
■ Part 5: Accelerating simulations for power technologies

Part 1: Basic Model Switches

In Sentaurus Process, Advanced Calibration covers several alternatives for diffusion and
activation models. Some models are relatively simple, such as the ChargedPair model for
dopant diffusion or the Transient activation of dopants. Other models are more sophisticated
(for example, the ChargedReact model for dopant diffusion and the ChargedCluster
model for dopant activation) but require more equations to be solved in the diffusion solver and,
therefore, require more CPU time. These different models coexist in Advanced Calibration so

Advanced Calibration for Process Simulation User Guide 9

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 1: Basic Model Switches

that, in simple limiting situations (for example, in thermal equilibrium for low dopant
concentration), different models give the same results. In more complex situations, for
example, during transient-enhanced diffusion (TED), the more complex models will give better
results.

The best choice of fundamental models depends on the problem to be solved. Part 1 of the
Advanced Calibration represents a choice that is recommended by Synopsys for most
applications. For most devices, the modeling of TED and dopant activation is important, and it
is necessary to use some models that describe the underlying physics accurately.

It is often useful to reduce the number of equations to be solved in order to save CPU time. On
the other hand, it might be sometimes necessary to select models that are more sophisticated
than the default choice, even at the cost of increasing the CPU time. In this chapter, the possible
changes with respect to the default model switches are explained.

The most elegant way to change a basic model switch is adding a corresponding line at the
beginning of a project-specific or user-specific calibration file, which is sourced after loading
the Advanced Calibration file. In this way, you can track the differences to the default
suggestions of Synopsys.

CPU time is an important issue for the process simulation of power device fabrication, which
often includes a large number of thermal anneals. A summary of the methods to speed up the
simulation of power device processes is given in Additional Calibration for Power
Technologies on page 155.

Part 4 of the Advanced Calibration file contains the procedure AdvancedModels, which
offers an option to switch to a consistently calibrated set of state-of-the-art models for dopant
and defect clustering with a single command line. This option is recommended for fundamental
research and also can be considered to be used in very advanced CMOS technology. It is
described in Part 4: Comprehensive and Slow Models on page 68.

Diffusion Models in Silicon and Germanium

See section 1.1 of AdvCal_2019.03.fps.

The default choice is the pair diffusion model ChargedPair. The dopants diffuse only through
dopant-defect pairs, where defects can be either interstitials or vacancies. All charge states of
defects and dopant-defect pairs are taken into account, and the concentration of pairs is
assumed to be in local equilibrium with unpaired dopants and defects.

A more sophisticated alternative is the ChargedReact model, a so-called five-stream model,

which is selected by:
pdbSet Si Dopant DiffModel ChargedReact

10 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 1: Basic Model Switches

Here, the diffusion of dopants is simulated through dopant-defect pairs. In contrast to the
ChargedPair model, the simplifying assumption of local equilibrium between pairs and
unpaired dopants is omitted. Instead, the kinetics of pair formation and dissolution is taken into
account. This model needs more CPU time than the ChargedPair model, because additional
equations need to be solved for each dopant. It is possible to select the ChargedReact model
individually for some dopants. For example, it might be reasonable to select it only for boron
but not for other dopants. This can be performed by adding the line:
pdbSet Si Boron DiffModel ChargedReact

It is possible to use the ChargedReact model for some dopants and the ChargedPair model
for all other dopants. In contrast, it is not recommended to mix the ChargedFermi model with
either of the ChargedReact or ChargedPair model, because the treatment of point defects
would become inconsistent.

Dopant Cluster Models in Silicon and Germanium

See section 1.2 of AdvCal_2019.03.fps.

These models govern the dopant activation during thermal annealing. The simplest and fastest
model is None, which means that there are no dopant clusters. This model is recommended for
dopants for which clustering has no influence. This is typically the case when the maximum
concentration of a dopant is far below the solid solubility. For example, in an NMOS simulation
12 –2
with a very low indium dose for the channel implantation (for example, 10 cm ), it is
reasonable to set the indium activation model to None, to speed up the simulation.

For the dopant impurities boron, indium, arsenic, phosphorus, and antimony in silicon as well
as for the dopant impurities boron, arsenic, phosphorus, antimony, and carbon in germanium,
the activation model Transient is used as the default. In this model, dopants can be bound in
clusters, which consist only of dopants of one species. The equilibrium distribution of dopants
into clusters and substitutional impurities is governed by the solid Solubility; the rate at
which the equilibrium is reached is governed by the parameter CluRate. Both Solubility
and CluRate are Arrhenius-type constants with individual parameters for each dopant.

Special models exist for boron, arsenic, phosphorus, indium, carbon, and fluorine clustering in
silicon.

Boron can form so-called boron–interstitial clusters (BICs) together with silicon interstitials.
BICs exist in various sizes, as BmIn ‘molecules’ inside silicon, which grow or evaporate by the
incorporation or emission of silicon interstitials or boron-interstitial pairs. The BIC model can
be selected by using:
pdbSet Si Boron ActiveModel ChargedCluster

Advanced Calibration for Process Simulation User Guide 11

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 1: Basic Model Switches

The BIC model is not used by default because the solution of individual equations for all BICs
is numerically expensive. Furthermore, the BIC model should only be used in combination
with the Full model for interstitial clusters, which uses more equations than the 1Moment
model for interstitial clusters, which is the Advanced Calibration default. The recommended
way to use the BIC model is to execute the procedure AdvancedModels defined in part 4 of
the Advanced Calibration file (see Part 4: Comprehensive and Slow Models on page 68). The
selection Boron ActiveModel Transient gives satisfactory results in many situations.

Arsenic and phosphorus can form clusters together with point defects, which are so-called
arsenic–vacancy (As–V), phosphorus–vacancy (P–V), and phosphorus–interstitial (P–I)
clusters. To take these into account, you can switch on the ChargedCluster model for silicon
using:
pdbSet Si Arsenic ActiveModel ChargedCluster
pdbSet Si Phosphorus ActiveModel ChargedCluster

In the case of arsenic, a family of four different As–V clusters and pure As clusters will form.
In the case of phosphorus, two different P–V clusters, one pure P cluster, and one P–I cluster
are modeled. The models can be used with both the 1Moment and Full models for interstitial
clusters. The formation and dissolution of As–V, P–V, and P–I clusters change the local
concentration of silicon point defects (interstitials and vacancies).

The older Cluster model, in which As4V is the only type of As cluster, is not recommended
to be used, as it sometimes overestimates the impact of As-cluster formation and dissolution
on the point-defect concentration in silicon. Instead, the default choice for As clustering is the
Transient model, in which As3 is the only As cluster and which is easy to understand and
calibrate.

The activation model ChargedCluster can also be applied to simulate formation of indium
clusters. This model is invoked in the procedure AdvancedModels (see Part 4:
Comprehensive and Slow Models on page 68).

Carbon can form carbon–interstitial clusters in silicon, with a similar reaction chain as for
BICs. The formation of carbon–interstitial clusters is activated by:
pdbSet Si Carbon ActiveModel NeutralCluster

Solving a transient equation for the formation and dissolution of Ge–B pairs in silicon is not
considered necessary. Instead, in cases where the chemical effect of Ge on B diffusion needs
to be taken into account, you can select a calibrated modification of B diffusivity in the
presence of germanium by using:
SiGe_and_Stress_Effect 1 1 1 0

immediately after sourcing the Advanced Calibration file. This is explained in Effect of
Germanium and Stress on page 38.

12 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 1: Basic Model Switches

Defect Cluster Models in Silicon and Germanium

See section 1.3 of AdvCal_2019.03.fps.

For silicon and germanium, interstitial clustering is described by the 1Moment cluster model.
In this model, the capturing and release of interstitials from {311} defects is described
according to a publication by Rafferty et al. [2]. This model uses only a single equation to
describe the time evolution of interstitial clusters and is considered a good compromise
between accuracy and computation speed.

A complex silicon interstitial-clustering model, including small interstitial clusters, {311}

defects, and dislocation loops, is used in the AdvancedModels set and described in Interstitial
Clusters on page 69.

Vacancy clusters are not simulated by default in Advanced Calibration because their modeling
is not needed for regular processes. If vacancy clusters are relevant, for example, in processes
that include vacancy engineering, that is, the creation of a vacancy-rich region by high-energy
implantation, you have the option of three different models for vacancy clusters in silicon.

The simplest model is switched on by:

pdbSet Si Vac ClusterModel 1Moment

In this case, the nucleation, growth, and dissolution of vacancy clustering are modeled with
arbitrary calibrated parameters, which is analogous to the standard model for interstitial
clusters.

A more comprehensive model as proposed in the ATOMICS research project [3] is enabled by:
pdbSet Si Vac ClusterModel Full

By default, if the Full model is selected for vacancy clusters, Sentaurus Process will solve
seven equations for small vacancy clusters (V2–V8), with calibrated parameters including
binding energies based on ab initio simulations [3].

For numeric efficiency, an alternative calibration of the Full model using fewer equations can
be selected by using:
pdbSet Si Vac ClusterModel Full
pdbSet Si Vac MultiClusterModel Full {2Moment}
pdbSetDouble Si Vac CL.Size 3
pdbSetDoubleArray Si V3 kfV { 0 {[expr 4*3.1415*2.97e-8*[pdbGet Si Vac D 0]]} }
pdbSetDoubleArray Si V3 krV { 0 0 }
pdbSetDoubleArray Si V3 kfI { 0 0 }
pdbSetDoubleArray Si V3 krI { 0 0 }

Advanced Calibration for Process Simulation User Guide 13

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 1: Basic Model Switches

In this case, one equation for the small vacancy clusters V2 and two equations for the voids
(DVoid for the concentration of V-clusters, and CVoid for the total concentrations of vacancies
in these clusters) are solved with arbitrary calibrated parameters.

Poisson Equation
See section 1.4 of AdvCal_2019.03.fps.

In Advanced Calibration, the Poisson equation for the electrical potential is solved for both
silicon and germanium. Alternatively, you can switch off the Poisson equation with the
command:
pdbSetBoolean Si Potential Poisson 0

In this case, local charge neutrality is assumed and the number of partial differential equations
is reduced by one. In most situations, local charge neutrality gives approximately the same
results as the Poisson equation. At p-n junctions, the assumption of charge neutrality gives a
sharper peak of the electric field than the Poisson equation, which results in slightly sharper
kinks of dopant profiles at p-n junctions.

Modern submicron CMOS devices have very thin dielectrics. Therefore, the electrostatic
interaction between the gate and the channel region in such devices is strong. This interaction
results in the presence of an additional potential at semiconductor surfaces (under the gate) and
strong electric fields in the semiconductor near semiconductor–oxide interfaces. To take this
electrostatic interaction into account, the Poisson equation with proper boundary conditions
can be solved in all materials, including dielectrics, especially under the gate [4]. Ideally, also
quantum corrections to the distribution of electrons and holes can be taken into account in the
simulation [4]. These effects, with a small but noticeable impact on CMOS device
characteristics [4], are not included in the current release of Advanced Calibration for
Sentaurus Process.

Channeling Dose in Analytic Implantations

See section 1.5 of AdvCal_2019.03.fps.

For analytic implantations, the switch ChanDoseInterpolation 1 selects the correct

method of interpolation of the ion-channeling dose between the tabulated values of the
Default tables.

The coimplantation model [5] for damage accumulation is switched on. This model provides a
description of ion channeling for successive ion implantations.

14 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 1: Basic Model Switches

Boundary Conditions
See section 1.6 of AdvCal_2019.03.fps.

In the pair diffusion model, the segregation of dopants at silicon and germanium surfaces
involves the capture or creation of dopant–defect pairs at the silicon side of the interface.

In the pair segregation model used in Advanced Calibration, when a dopant–defect pair
diffuses to an interface between semiconductor and another material, the dopant can enter the
other material (or, for three-phase segregation, the interface layer), whereas the point defect
remains on the semiconductor side of the interface.

The following selection means that the point defect released can have any charge state (and not
only a neutral charge state):
pdbSet Ox_Si Boundary UseUnpairedTotalInt 1
pdbSet Gas_Si Boundary UseUnpairedTotalInt 1
pdbSet Nit_Si Boundary UseUnpairedTotalInt 1

Similarly, for the opposite segregation reaction, when a dopant–defect pair is formed at the
silicon side of the interface, a point defect with any charge state can be consumed at the silicon
side of the interface.

As a consequence of this selection, the time at which segregation equilibrium is reached in

highly doped regions, where most point defects are charged, is decreased. The segregation
equilibrium itself is not affected. Although the name of the Boolean parameter is
UseUnpairedTotalInt, the selection is applied to both interstitials and vacancies.

The selection:
pdbSetSwitch Ox_Si I Surf.Recomb.Vel Normalized

and the corresponding lines for vacancies and other interfaces allow the generation and
recombination of point defects at silicon surfaces in all charge states.

For B, As, and P at Si–SiO2, Si–Si3N4, Ge–SiO2, and Ge–GeO2 interfaces, the three-phase
segregation model is the default in Advanced Calibration. For In and Sb, the simpler
segregation model is the default.

Advanced Calibration for Process Simulation User Guide 15

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 1: Basic Model Switches

Numeric Solver
See section 1.7 of AdvCal_2019.03.fps.

By default, the direct solver PARDISO is used for 1D and 2D simulations, and the iterative
solver ILS is used for 3D simulations. For better performance, Advanced Calibration selects
ILS to solve the linear systems also in 2D. Specific parameters can be set for ILS in 2D. While
nd is selected by default for ILS.symmOrdering as the optimum for multithreaded
calculations, you should consider switching to mmd for single-thread calculations.

Summary of Model Switches

Table 2 and Table 3 on page 17 summarize the default model switches and all the alternatives
supported by Advanced Calibration. For all supported model switches, the corresponding
calibrated parameters are included in the Advanced Calibration file (AdvCal_2019.03.fps)
and are ready to be applied automatically when alternative models are selected. The procedure
AdvancedModels, which switches on several more complex models at the same time, is
explained in Part 4: Comprehensive and Slow Models on page 68.

Table 2 Model switches for silicon in Advanced Calibration

Model Default Supported alternatives

pdbSet Si Dopant DiffModel ChargedPair ChargedReact1

pdbSet Si Boron ActiveModel Transient None

pdbSet Si Indium ActiveModel Transient None

pdbSet Si Arsenic ActiveModel Transient None, ChargedCluster

pdbSet Si Phosphorus ActiveModel Transient None, ChargedCluster

pdbSet Si Antimony ActiveModel Transient None

pdbSet Si Germanium ActiveModel None

pdbSet Si Carbon ActiveModel NeutralCluster None

pdbSet Si Int ClusterModel 1Moment

pdbSet Si Vac ClusterModel None 1Moment, Full

pdbSet Si Potential Poisson 1 0

pdbSet ImplantData UseCoImplant 1 0

pdbSet Ox_Si Boundary UseUnpairedTotalInt 1

1. If the basic choice is ChargedPair, it is possible to select ChargedReact for individual dopants.

16 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Table 3 Model switches for germanium in Advanced Calibration

Model Default Supported alternatives

pdbSet Ge Dopant DiffModel ChargedPair ChargedReact1

pdbSet Ge Boron ActiveModel Transient None

pdbSet Ge Arsenic ActiveModel Transient None

pdbSet Ge Phosphorus ActiveModel Transient None

pdbSet Ge Antimony ActiveModel Transient None

pdbSet Ge Carbon ActiveModel Transient None

pdbSet Ge Int ClusterModel 1Moment None

pdbSet Ge Vac ClusterModel None

pdbSet Ge Potential Poisson 1 0

pdbSet ImplantData UseCoImplant 1 0

pdbSet GeOx_Ge Boundary UseUnpairedTotalInt 1

1. If the basic choice is ChargedPair, it is possible to select ChargedReact for individual dopants.

Part 2: Constant Parameters

This part of the Advanced Calibration file contains the parameters for the diffusion and reaction
equations, which are set at the beginning of the process simulation and remain valid for all
process steps until the end of the simulation. The parameters are set for all alternatives listed
in Table 2 on page 16. This allows you to select any of the alternatives models with all
corresponding parameters by using a single command line, which can be ideally placed in a
user calibration file, which is sourced immediately after sourcing AdvCal_2019.03.fps.

Many parameters are taken from either the literature of carefully designed experiments or the
publication by Pichler [1], which gives an outstanding, comprehensive overview on the
publications of impurity diffusion and activation in silicon. Other parameters have been
calibrated based on the SIMS database of Synopsys.

Model parameters, which depend on particular ion implantation steps, are included in the third
part of the Advanced Calibration file and are described in Part 3: Initial Conditions After Ion
Implantation on page 55. Examples of these are the number of point defects generated by ion
implantation, which can depend on the implantation conditions.

Advanced Calibration for Process Simulation User Guide 17

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Basic Point-Defect Parameters

See section 2.1 of AdvCal_2019.03.fps.

The bulk parameters for interstitials and vacancies (sections 2.1.1–2.1.3 of

AdvCal_2019.03.fps) are the most fundamental parameters in the pair diffusion model.
They have been carefully selected from the literature. Any change affects not only the diffusion
of point defects, but also the diffusion of all dopant species that diffuse in dopant-defect pairs.
Changing the point-defect parameters with every new technology calibration would make it
difficult to compare the results of different calibration projects. Therefore, it is strongly
recommended that these parameters are not changed in any way.

In principle, this is also true for the surface boundary conditions (BCs) for point defects
(section 2.1.4 of AdvCal_2019.03.fps). It is advisable not to change them because the
calibration of all models for TED and the diffusion of all dopants would be affected. However,
the BCs depend on the capping material and the local concentration of impurities. For
polysilicon and oxynitride, the BCs can depend on the details of the process flow. Therefore,
in practice, the surface recombination lengths of point defects can be considered to be
calibration parameters for the fine-tuning of process simulation.

Oxidation and nitridation cause the injection of interstitials and vacancies, respectively, at the
exposed surface. A calibration of interstitial injection has been performed for dry oxidation.
For nitridation and wet oxidation, the surface boundary conditions for point defects are less
reliable than for inert atmosphere and can be considered to be calibration parameters for the
fine-tuning of diffusion processes.

Bulk Parameters for Free Interstitials

See section 2.1.1 of AdvCal_2019.03.fps.

For silicon, the diffusivity of interstitials Di is taken from Bracht et al. [6]. The equilibrium
concentration Cstar is chosen such that the product Di*Cstar has the value
25 –1 –1
1.59 × 10 × exp ( – 4.702 eV/kT ) cm s . This is a reasonable compromise between
conflicting suggestions in the literature [1][7][8] and is in acceptable agreement with various
‘clean’ data on silicon isotope diffusion and dopant diffusion in silicon that has been published
[6][8][9]. The same value for Di*Cstar was also used in [10].

The charge distribution for free interstitials and vacancies was taken from method.advanced
of the Synopsys Taurus™ TSUPREM-4™ process simulator and is based on various
publications [11][12][13]. During calibration, a small change with respect to the TSUPREM-4
parameters has been introduced for the relative abundance of negatively charged vacancies.

18 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

For germanium, there is a lack of experimental data on self-interstitial properties. For the
diffusivity Di of interstitials in germanium, a migration energy of 1.6 eV is assumed, which is
0.4 eV higher than calculation results for uncharged interstitials published by Vanhellemont et
al. [14]. The equilibrium concentration Cstar is estimated based on the formation energy
(2.78 eV), which is 0.4 eV less than the calculated value of Vanhellemont et al. [15]. This
choice is partially motivated by the parameter choice for pure silicon, where the activation
energy for high-temperature interstitial migration, derived from experiments by Bracht [6], is
higher than the value calculated with ab initio methods. For simplicity, the prefactors for the
diffusivity and the equilibrium concentration have the same values as in silicon.

Bulk Parameters for Free Vacancies

See section 2.1.2 of AdvCal_2019.03.fps.

For silicon, the diffusivity of vacancies Dv is taken from [6]. The equilibrium concentration
Cstar is chosen such that Dv*Cstar corresponds to the value from [6] at 1014.25°C . The
activation energy for Dv*Cstar (4.14 eV) is taken from [7].

For germanium, the equilibrium concentration of vacancies Cstar is based on [14], but with
the formation energy reduced by 0.4 eV. The diffusion barrier of vacancies Dv is derived from
the vacancy equilibrium concentration and the experimental vacancy-mediated self-diffusion
( –2 ) –1
coefficient of germanium ( 13.6 × exp ( – 3.09 eV/kT ) ) cm s following [16]. The resulting
migration barrier is 1.14 eV.

Bulk Recombination of Point Defects

See section 2.1.3 of AdvCal_2019.03.fps.

It is assumed that the bulk recombination is diffusion limited. Furthermore, the recombination
of interstitials and vacancies, which are both positively or both negatively charged, is assumed
to be suppressed by electrostatic repulsion.

Boundary Conditions for Point Defects

See section 2.1.4 of AdvCal_2019.03.fps.

Natural boundaries for both vacancies and interstitials are assumed. The surface recombination
length is 1 nm for Si–SiO2, Ge–SiO2, and Ge–GeO2 boundaries, and 10 nm for Si–SiN
boundaries. For gas–silicon and gas–germanium boundaries, which are used during epitaxy,
the surface recombination for vacancies is strongly reduced.

Advanced Calibration for Process Simulation User Guide 19

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Oxidation-Enhanced Diffusion

See section 2.1.5 of AdvCal_2019.03.fps.

During oxidation, there is an additional flux of interstitials into silicon. The rate of interstitial
injection by oxidation is proportional to the parameter theta and depends on the velocity v of
the moving Si–SiO2 interface and the electron concentration at the silicon side of the interface
by the factor:
( 1 + Gpow ) mm + m + 1 + p + pp
v × ---------------------------------------------------------------------------------------------------------------------------------------------------------
2 –1 –2
- (1)
mm × ( n ⁄ n i ) + m × ( n ⁄ n i ) + 1 + p × ( n ⁄ n i ) + pp × ( n ⁄ n i )

where theta, Gpow, mm, m, p, and pp are defined in AdvCal_2019.03.fps.

For dry oxidation of silicon, the values of theta and Gpow at Oxide_Silicon were
calibrated with experimental data from [17]1 and [18] for low-doped silicon. The values of mm,
m, p, and pp have been calibrated with data from ultrashallow junction (USJ) formation in dry,
oxidizing atmosphere. They can be modified for the purpose of fine-tuning oxidation-enhanced
diffusion for high surface doping. For wet atmosphere (partial pressure of H 2 O > 0 ), a smaller
value of theta has been calibrated from corresponding SIMS data.

Interstitial injection during dry oxidation appears to be suppressed in the case of high mole-
fraction SiGe [19]. Therefore, the value of theta at Oxide_Germanium, the upper mole-
fraction limit for SiGe–SiO2 interfaces, is set to a five orders of magnitude lower value for
oxidation in general compared to the value at Oxide_Silicon for dry oxidation. For the
calibration of oxidation of SiGe for all mole fractions, see SiGe Oxidation on page 46.

NOTE No calibration parameters for germanium oxidation-enhanced diffusion

are included in Advanced Calibration. By default, no interstitial
injection at GeOxide_Germanium is assumed.

In addition to interstitial injection into silicon, the boundary condition for vacancies at the
moving Si–SiO2 interface during oxidation is altered. The equilibrium concentration of
vacancies defined for the moving interface is lower compared to the bulk. This effect is
implemented by the term VacInterfaceCStarFactorOED and is calibrated based on dopant
SIMS and SiGe interdiffusion profiles of oxidation experiments.

1. The reference [17] includes data for oxidation-enhanced diffusion (OED) of P and B. A higher diffusion
enhancement was reported for P than for B, which was ascribed to a 20% vacancy component of boron diffusion.
Recent experiments indicate that the vacancy component for B diffusion should be much less than 20% (for an
overview, see [1]). Therefore, instead of ascribing the lower OED of boron to diffusion of B–V pairs, the calibration
of theta with data from [17] was performed under the assumption that the observed difference between the OED
of P and B was mainly due to experimental inaccuracies. Giving equal weight to the P and B data, Synopsys obtained
a 10% reduction of theta with respect to using only the P data for the extraction of theta.

20 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Transient-Enhanced Diffusion
See section 2.2 of AdvCal_2019.03.fps.

The model of Rafferty et al. [2] is used to simulate the evaporation of silicon interstitials from
{311} defects. The reaction rates for the capture and evaporation of interstitials have been
calibrated with transmission electron microscope (TEM) data on the dissolution of {311}
defects published by Stolk et al. [20] and Saleh et al. [21].

As an initial condition, it is assumed that all interstitials generated by ion implantation are
bound in {311} clusters (InitPercent = 1.0).

The selected model gives accurate results for the dissolution of {311} defects, as illustrated in
Figure 15 on page 110. However, note that the model is too simple to describe the initial phase
of ultrahigh interstitial supersaturation after ion implantation, which was reported by Cowern
et al. [9] and is ascribed to the formation and dissolution of small interstitial clusters, and which
is illustrated in Figure 6 on page 70. In addition, the model underestimates the stability of
interstitial clusters in situations where dislocation loops form and where most of the excess
interstitials are bound to dislocation loops rather than {311} defects. This might happen, for
example, after amorphizing implantations into silicon, as illustrated in Figure 7 on page 71.

In situations where TED is not governed by {311} defects, but rather by small clusters or
dislocation loops, the model is less accurate. A more comprehensive model for silicon
interstitial clusters has been calibrated by Zographos et al. [22]. This is switched on if you
execute the procedure AdvancedModels, defined in part 4 of the Advanced Calibration file.

The model of Rafferty et al. [2] also is used to simulate the evolution of extended interstitial
defects in germanium. The reaction rates for the capture and evaporation of interstitials have
been calibrated [23] with experimental data from Napolitani et al. [24], where B diffusion
events are correlated quantitatively with the measured positive strain associated with the end
of range (EOR) damage.

Advanced Calibration for Process Simulation User Guide 21

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Boron Diffusion and Activation

See section 2.3 of AdvCal_2019.03.fps.

Boron Diffusion Coefficient

The macroscopic values for the boron diffusivity in silicon are based on the literature [1], and
the diffusion of boron is assumed to be only interstitial mediated:
pdbSetDoubleArray Si B Int D { 0 {[Arr 0.123 3.57]}
1 {[expr [Arr 3.71 3.67]+[Arr 2.5e-6 2.5]]}
2 {[Arr 39.8 4.37]} }

Unlike in [1], the diffusivity of B using B–I+ pairs is a sum of two Arrhenius expressions. The
second, with an Arrhenius energy of 2.5 eV, is introduced to increase the diffusivity of B for
low temperatures (T < 840°C ) in comparison to the B diffusivity suggested in [1]. It should be
noted that the scientific literature [1] gives inconsistent values for B diffusivity at low
temperatures. Following TSUPREM-4, the pairing constants are chosen such that the
diffusivity of boron–interstitial pairs has the same order of magnitude as the diffusivity of
unpaired interstitials.

The hopping length of B is taken from Giles et al. [25] and is based on B marker layer diffusion
data in the temperature range of 500°C – 800°C . It is only relevant if the ChargedReact
model is switched on for boron. In this case, the hopping length has an influence on the length
of the tail of the profile.

The Boolean switch Kick.Out.Rate.Based.On.Lambda is set to 1. With this setting,

relevant only when the ChargedCluster model is selected, the average hopping length
between formation and dissolution of B–I pairs equals the parameter lambda, unless you scale
the kick-out reaction rate by the user-defined term React<dopant><defect>Factor. (With
Kick.Out.Rate.Based.On.Lambda 0, which was the only option in Sentaurus Process
Version K-2015.06 and earlier, this is not strictly the case in strained Si and SiGe, and in the
presence of diffusion enhancement factors.) To preserve backward compatibility for unstrained
Si, the term ReactBoronIntFactor is defined as "1.0/BoronDiffFactor". In this way,
the migration length of B–I pairs in unstrained Si regions of high F concentration is the same
as in AdvCal_2015.06.fps.

The macroscopic values for boron diffusivity in germanium are based on the literature [26], and
the diffusion of boron is assumed to be only interstitial mediated [23]. For consistency with the
diffusivity in silicon, a negligible contribution of B–I++ pairs is included as well. Boron
undergoes very little intrinsic or transient-enhanced diffusion in germanium.

22 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Effect of Fluorine

It is known that boron diffusion in silicon can be reduced by the presence of fluorine. The main
reason for this is that F–V clusters, which form after ion implantation, catch excess interstitials,
which are also present in silicon after implantation [27]. A complete physics-based model for
the interactions between B, I, and F must be very complex, because F atoms are redistributed
during solid phase epitaxial regrowth of amorphized layers. Instead, in the Advanced
Calibration, a simpler approach is used by default.

It is assumed that F atoms are immobile after ion implantation and that the presence of F atoms
reduces directly the diffusivity of B atoms by a factor (BoronDiffFactor), which depends
20
on the F concentration. This factor is close to 1 for F concentrations smaller than 1 × 10 and
becomes important only for very high F concentrations. It has been calibrated by comparisons
of USJ boron SIMS profiles, which were made by boron implantation and annealing, and BF2
implantation and annealing, respectively. This simple approach is not very predictive. Since it
assumes a local B–F interaction, it is not suitable to study USJ formation after Ge+F+B
cocktail implantations.

A sophisticated F–V clustering model is available. The model is invoked by the procedures
AdvancedFluorineModel and AdvancedModels (see Part 4: Comprehensive and Slow
Models on page 68). This model describes the fluorine effect on boron diffusion in silicon more
accurately.

For germanium, no effect of fluorine on boron diffusion is assumed.

Boron Clustering
The Transient cluster model is used for silicon and germanium. Four boron atoms form a
cluster. Interstitials are not built into B clusters.

For silicon, the solid solubility value is a fit to the collection of literature data compiled by
Pichler [1]. Two Arrhenius functions are combined: one covers the range T < 1000°C and the
other, T > 1000°C . The active boron reaches the solid solubility if the total B concentration
reaches TotSolubility, which is defined in Advanced Calibration as three times the B solid
solubility. The rate at which the equilibrium between active and clustered B is reached is given
by the parameter CluRate. CluRate has been calibrated by Synopsys using experimental data
from the Synopsys SIMS database. With the obtained value, a significant amount of boron
SIMS data can be reproduced, including ultrashallow junction profiles. However, in some
situations, it is recommended to tune CluRate to improve the accuracy of fitting SIMS data or
sheet resistance data in the process window of interest. In particular, in the presence of F, it
might be necessary to reduce CluRate.

Advanced Calibration for Process Simulation User Guide 23

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Accelerated boron-cluster formation at the initial annealing stage after implantation is taken
into account by introducing the dependency of the forward-clustering rate on the interstitial
supersaturation:
term Si name=BoronTClusterForwardFac store add eqn="(Int/EqInt)^2.2"

This basically means that interstitials serve as a catalyst for boron deactivation.

For the initial activation of boron after implantation, the basic assumption is that it is given by
a small value in crystalline silicon (AcInit) and by a higher value (AmInit) in recrystallized
areas. In the Synopsys calibration, the initial activation of boron in recrystallized regions is
smaller than the measured values reported, for example, by Colombeau et al. [28]. This is
necessary because, with the chosen clustering model for boron, using initial activation levels
20 –3
higher than 1 × 10 cm results in an overestimated transient-enhanced diffusion of boron for
a number of SIMS data in the Synopsys database, where ultralow energy implantation was
followed by low-temperature annealing.

For germanium, the solid solubility, the clustering rate, and the initial activation levels after
implantation (AcInit and AmInit) are calibrated against published data [26][29][30] and data
provided by AMAT-VSE1. For long-time anneals (hours at 900°C), boron shows a very low
solid solubility [26], which was used in earlier calibration work [23]. However, higher
activation levels (similar to boron activation in silicon) have been observed after implantation
and short-time anneals.

The apparent discrepancy between long-time and short-time anneals might indicate that
different mechanisms are responsible for B deactivation in Ge, that is:
■ The formation of small B clusters as in silicon.
■ A different mechanism that requires a higher thermal budget.

In the Transient model, only a single type of B clusters with a single deactivation reaction
is assumed. In AdvCal_2019.03.fps, the solubility of boron in Ge has been calibrated to a
high value typical for post-implantation anneal with small or medium thermal budgets (up to
several minutes at 860°C). As for silicon, the term BoronTClusterForwardFac for the
dependency of the forward-clustering rate on the interstitial supersaturation is defined for
germanium.

For the initial activation in the case of in situ boron-doped epitaxial growth of silicon and
20 –3
germanium, an activation level of 4 × 10 cm is assumed and defined by the EpiInit
parameter.

If the BIC (ChargedCluster) model is used for silicon, good results for B activation and
deactivation can be obtained with realistic assumptions for AcInit and AmInit. This is taken
into account in the procedure AdvancedModels in part 4 of the Advanced Calibration file, in

1. Applied Materials - Varian Semiconductor Equipment.

24 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

which the BIC model is switched on. The calibrated cluster parameters for the BIC model are
contained in section 2.3.2.3 of AdvCal_2019.03.fps.

Boron Dose Loss

For B, As, and P, the three-phase segregation model is used for dose loss modeling.

In Sentaurus Process, by default, these dopants do not share trap sites at the interface with other
dopants. For the ChargedPair model and the ChargedReact diffusion model, the flux of
dopants from silicon or germanium into the interface layer is proportional to the concentration
of dopant–defect pairs on the silicon side of the interface, and the out-diffusion from the
interface to silicon or germanium is proportional to the concentration of point defects on the
silicon side of the interface.

The number CMax of trap sites at the interface and the trapping and emission rates of the three-
phase segregation model have been calibrated by Synopsys, based on a collection of boron
SIMS profiles.

The diffusivity of trapped B atoms along the Si–SiO2, Ge–SiO2, and Ge–GeO2 interfaces is set
to zero. It can be set to a value higher than zero for increasing the lateral diffusion in 2D or 3D
simulations [31].
21 –3
In oxide, the boron diffusivity is increased for very high B concentrations (> 1 × 10 cm )
and for high F concentrations. This diffusion enhancement has been calibrated by Synopsys,
21 –3
based on SIMS profiles. For very high B concentration in oxide (> 1 × 10 cm ), out-
diffusion into the gas contributes to the dose loss. The out-diffusion rate has been calibrated by
Synopsys, using SIMS data.

Spacer oxides that are formed by a TEOS process and capped by a SiN layer can contain a high
concentration of hydrogen, which increases the boron diffusivity in oxide. This case is not
taken into account in the Advanced Calibration file.

Out-diffusion from bare silicon surfaces (without an oxide layer between silicon and the gas
ambient) has not been calibrated yet.

Dose loss parameters also are provided for nitride–silicon interfaces. These have the same
values as for oxide–silicon interfaces and will be considered only as a starting point for a
calibration to be performed by users.

Based on the experimental data available, boron shows no significant dose loss at the Ge–GeO2
interface. The trapping rate at the semiconductor side is set to a lower value than for Si–SiO2.

Advanced Calibration for Process Simulation User Guide 25

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Fluorine
See section 2.4 of AdvCal_2019.03.fps.

As previously explained, Advanced Calibration offers two modeling approaches for the effects
of fluorine in silicon: a very simple approach and a sophisticated approach.

In the very simple approach, it is assumed that fluorine is completely immobile after ion
implantation. Three empirical effects of fluorine on B diffusion are implemented in section 2.3
of AdvCal_2019.03.fps:
■ Fluorine reduces directly the B diffusivity in Si (using BoronDiffFactor).
■ Fluorine increases the B diffusivity in oxide and, therefore, increases the dose loss.
■ Fluorine increases the B out-diffusion from oxide to the gas and, therefore, increases the
B dose loss.

The effect of fluorine on boron diffusion and dose loss has been calibrated with SIMS data
from BF2 implantation and annealing. For all data, B and F have approximately the same spatial
distribution after ion implantation. The calibration is valid only for this particular situation.
Experimental data where fluorine was implanted separately was not taken into account in the
calibration. When F is implanted separately from B, the simple approach for the influence of
F on B diffusion is not predictive.

The sophisticated physics-based fluorine model is explained in Part 4: Comprehensive and

Slow Models on page 68.

For germanium, the very simple approach, with the assumption that fluorine is completely
immobile after ion implantation, is also chosen by default. No effect on boron diffusion is
assumed.

Arsenic Diffusion and Activation

See section 2.5 of AdvCal_2019.03.fps.

Arsenic Diffusivity

The values for the diffusivity of arsenic in silicon have been calibrated by Synopsys based on
20 –3
SIMS data. For regions with high As concentration (> 2.0 × 10 cm ), the diffusivity of As
is increased sharply, following the measurements by Larsen et al. [32] and using a formula that
is based on the percolation theory and lattice Monte Carlo simulations [33]. It is assumed that
a high concentration of P increases the diffusivity of As–V pairs in a similar way as a high

26 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

concentration of As using percolation. Both vacancy and interstitial components of the

diffusion coefficients were calibrated to obtain better simulation results in the high As
concentration region. Following TSUPREM-4, the pairing constants are chosen such that the
diffusivity of arsenic-defect pairs has the same order of magnitude as the diffusivity of unpaired
defects.

The macroscopic value for arsenic diffusivity in germanium is calibrated [23] based on
different data [34][35][36][37][38], and the diffusion of arsenic is assumed to be dominated by
As+V– – pairs. For consistency with the diffusivity in silicon, As–I and As–V pairs with
different charge states are included as well. In pure germanium, As diffusion is dominated by
diffusion of As+V– – pairs. However, other arsenic-defect pair diffusivities are relevant as well,
in particular, for mole fraction–interpolated diffusivities in SiGe. Under extrinsic doping, the
diffusion of arsenic is strongly enhanced.

Arsenic Clusters
Arsenic forms clusters with vacancies in silicon [1][39] and germanium [40]. However, the
activation model Transient is the default one for simplicity. It is assumed that three arsenic
atoms form an As3 cluster. The values of the solid solubility and the clustering rate have been
calibrated by Synopsys.

For silicon, the calibration includes an Arrhenius break at 880°C for CluRate to allow for
good accuracy of As deactivation at low-temperature processes:
pdbSet Si As CluRate {[ArrBreak 7.4974e-5 1.40 5.51647e3 3.2 880]}

The parameter KcEq.From.Default.Formula is set to 1. This option is relevant for mole-

fraction interpolation in SiGe. With the option switched on, the equilibrium active
concentration corresponds to the solubility specified in SiGe exactly as intended.

The impact of phosphorus on arsenic activation in silicon is taken into account by an empirical
expression for the cluster dissolution rate:
term Si name=ArsenicTClusterBackwardFac store add eqn= "1.0+PActive/1e20"

A more complex As–V clustering (ChargedCluster) model can be selected as well to

simulate arsenic activation in silicon. The kinetics of As cluster formation is described with a
family of four neutral clusters: As2, As2V, As3, and As3V. The model reaction pathway is shown
in Figure 1 on page 28.

Advanced Calibration for Process Simulation User Guide 27

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

As As2 As
As33

V AsV V, I AsI AsV V

AsV As2V As3V

Figure 1 Reaction pathway for As–V clustering model

In this model, as for the BIC model, the most important calibration parameters are the
formation energies (ClusterFormE). The most stable cluster type at high arsenic
concentration is As3.

For the initial activation of arsenic after implantation, the basic assumption is that it is given by
16 –3
a small value of 2 × 10 cm in nonamorphized crystalline silicon (AcInit) and by a higher
20 –3
value of 2 × 10 cm (AmInit) in recrystallized silicon.

For germanium, the solid solubility, the clustering rate, and the initial activation levels after
implantation (AcInit and AmInit) are calibrated [23] based on a collection of data [35][36]
[37].

For the initial activation in the case of in situ arsenic-doped epitaxial growth of silicon and
20 –3
germanium, an activation level of 2 × 10 cm is assumed and defined by the parameter
EpiInit.

Arsenic Dose Loss

Arsenic dose loss is modeled by the three-phase segregation model, following the experimental
work of Kasnavi et al. [41] and the model of Oh and Ward [31]. In the Advanced Calibration,
arsenic atoms do not share interface trap sites with other dopants such as P. The parameters for
As dose loss have been calibrated by Synopsys based on SIMS data. The parameter CMax,
2
which gives the maximum concentration of As atoms that can be stored per cm of the
interface, is used to fine-tune the As dose loss. In the ChargedPair model and the
ChargedReact model, the segregation rate is proportional to the concentration of As–I and
As–V pairs.

28 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Phosphorus Diffusion and Activation

See section 2.6 of AdvCal_2019.03.fps.

Phosphorus Diffusivity

In silicon, phosphorus diffuses predominantly through interstitials at high temperatures. The

diffusivity of P–I pairs has a similar value as the fit to literature data in [1], but with an
increased relative contribution of P+I0 pairs and a reduced contribution of P+I– pairs. The
diffusivity of P through P–V pairs is smaller. It is relevant only at very high P concentration
and is most relevant when C co-doping is used.

The macroscopic value for phosphorus diffusivity in germanium is calibrated based on

different data [34][35][36][37][42][43][44][45][46], and the diffusion of phosphorus is
assumed to be dominated by P+V– – pairs. For consistency with the diffusivity in silicon and for
the purpose of mole-fraction interpolation of diffusivities in SiGe, P–I and P–V pairs with
different charge states are included as well. Under extrinsic doping, the diffusion of phosphorus
is strongly enhanced due to the increased abundance of mobile P+V– – pairs.

Following TSUPREM-4, the pairing constants are chosen such that the diffusivity of
phosphorus-defect pairs equals approximately the diffusivity of unpaired defects.

Phosphorus Clusters

Phosphorus forms clusters with vacancies in silicon [1][39] and germanium [47], and in
addition with interstitials in silicon [1][39]. However, the activation model Transient is the
default one for simplicity. The model parameters for the formation and dissolution of P clusters
have been calibrated by Synopsys, based on SIMS data at high concentrations. It is assumed
that three P atoms can form a P3 cluster.

It should be mentioned that the clustering model underestimates the stability of P complexes
22 –3
for extremely high P concentrations (> 10 cm ) in silicon near the surface, which can be
15 –2
obtained after high-dose implantation (for example, 5 × 10 cm ) with low energy (for
example, 2 keV).

A more complex P–V and P–I clustering (ChargedCluster) model can be selected as well
to simulate phosphorus activation in silicon. The kinetics of P cluster formation is described
with a family of four neutral clusters: P2, P2V, P3V, and P2I. In this model, as for the BIC model,
the most important calibration parameters are the formation energies (ClusterFormE). Note
that the complex model does not have a higher accuracy than the default one in general.

Advanced Calibration for Process Simulation User Guide 29

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

For the initial activation of phosphorus after implantation, the basic assumption is that it is
18 –3
given by a small value of 4 × 10 cm in nonamorphized crystalline silicon (AcInit) and by
20 –3
a higher value of 3.0 × 10 cm (AmInit) in recrystallized areas.

For germanium, the solid solubility, the clustering rate, and the initial activation levels after
implantation (AcInit and AmInit) are calibrated based on a collection of data [35][36]
[37][42][43][44][45][46].

For the initial activation in the case of in situ phosphorus-doped epitaxial growth of silicon and
20 –3
germanium, an activation level of 2 × 10 cm is assumed and defined by the EpiInit
parameter.

As–P Clusters in Silicon

A simple model for As–P clusters in silicon is implemented in section 2.6.2.3 of

AdvCal_2019.03.fps as an alternative approach to reduce P diffusion in the region with
high As concentration. It is not switched on by default because the calibration is not reliable
and because the direct modification of P diffusivity and As diffusivity in co-doped regions
provides a superior overall accuracy. After sourcing the Advanced Calibration file, you can
switch on the As–P cluster model with the command:
Use_As3P_clusters

The command Use_As3P_clusters is a procedure defined in section 2.6.2.3 of

AdvCal_2019.03.fps, which instructs Sentaurus Process to solve for the mixed cluster
As3P. The parameters As3P_k1 and As3P_k2, defined in section 2.6.2.3, govern the formation
and dissolution rate of such clusters.

NOTE The As–P clustering model must only be used when using the
Transient cluster model for As.

Phosphorus Dose Loss

Phosphorus dose loss is described by the three-phase segregation model. Phosphorus atoms can
be incorporated into the silicon–SiO2 interface.

In addition, it is assumed that two P atoms located at the silicon–SiO2 interface can form P2
pairs. The concentration of pairs increases quadratically with the concentration of unpaired P
atoms trapped at the interface. The calibration of P and P2 trapping at the interface is performed
17 –3
in such a way that, for low P concentrations at the silicon side of the interface (< 10 cm ),
20 –3
unpaired P atoms govern the P dose loss. For high P concentrations (> 10 cm ) and inert
anneals (Si–SiO2 interface not moving), a considerable fraction of phosphorus trapped at the
interface is bound in P2 pairs. For oxidation (Si–SiO2 interface moving), the P2 pairs are
unstable even at high P concentration.

30 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

This dose loss model was developed due to the need to calibrate, with a consistent set of
parameters, the P dose loss for low and high interface concentrations, using the Synopsys SIMS
database. The model is supported by the results of first-principles calculations on the
mechanism of P segregation at the Si–SiO2 interface [48]. At the Ge–GeO2 interface, P shows
strong dose loss also at low concentrations [49].

The interface trap density, emission, and trapping rates, and the pair formation and dissolution
rates have been calibrated by Synopsys based on phosphorus SIMS data ranging from
ultrashallow junction formation to long-time oxidation.

The last lines of section 2.6.3 of AdvCal_2019.03.fps contain the Alagator implementation
of the P2 pair trap formation at Si–SiO2 interfaces.

NOTE When using Segregation boundary conditions at Si–SiO2, Si–Si3N4,

or Ge–SiO2 interfaces instead of ThreePhaseSegregation boundary
conditions, switch off the solution P2trap:
pdbSet Ox_Si P BoundaryCondition Segregation
solution name=P2trap nosolve store

Arsenic–Phosphorus Co-Diffusion
The physics of P and As diffusion is very complex if P and As are implanted and annealed
14 –2
together, in particular, for high As implantation doses (> 10 cm ). The following empirical
approaches are included in the Advanced Calibration file to achieve a good accuracy for As–P
co-diffusion in silicon:
■ The P diffusivity is modified as a function of arsenic concentration using the terms
PhosphorusIntDiffFactor and PhosphorusVacDiffFactor. These terms, if
defined, are multiplied by the diffusivity of P–I and P–V pairs, respectively. The diffusivity
decreases as follows (section 2.6.1.4 of AdvCal_2019.03.fps):
term Si name=PhosphorusIntDiffFactorDopant store add \
eqn= {([Arr 3.846e21 0.5]/([Arr 3.846e21 0.5]+Arsenic))}
term Si name=PhosphorusVacDiffFactorDopant store add \
eqn= {([Arr 2.4e22 0.5]/([Arr 2.4e22 0.5]+Arsenic))}
■ The pressure effects are redefined inside the SiGe_and_Stress_Effect procedure
(section 2.14.2 of AdvCal_2019.03.fps):
term Si name=PhosphorusIntDiffFactorPressure store add eqn= 1.0
term Si name=PhosphorusVacDiffFactorPressure store add eqn= 1.0
term Si name=PhosphorusIntDiffFactor store add \
eqn= {PhosphorusIntDiffFactorDopant * PhosphorusIntDiffFactorPressure}
term Si name=PhosphorusVacDiffFactor store add \
eqn= {PhosphorusVacDiffFactorDopant * PhosphorusVacDiffFactorPressure}

Advanced Calibration for Process Simulation User Guide 31

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

■ In this implementation, PhosphorusIntDiffFactor is defined as a product of two

terms. The factor PhosphorusIntDiffFactorDopant describes the modification of
P–I diffusivity in the presence of As. The factor PhosphorusIntDiffFactorPressure
can be used to describe the impact of pressure on P–I diffusivity. It is 1.0 by default and can
be redefined to include the impact of pressure if you switch on stress effects in a process
simulation with the SiGe_and_Stress_Effect procedure (see Stress Effect
(Stress_Eff) on page 39). This implementation allows you to separately fine-tune the
impact of dopants and the impact of stress on the diffusivity of P–I pairs.
■ Similarly, the terms PhosphorusVacDiffFactor, ArsenicVacDiffFactor, and
ArsenicIntDiffFactor also are defined as a product of two terms.
■ The As diffusivity is modified as a function of arsenic and phosphorus concentration using
the terms ArsenicVacDiffFactor and ArsenicIntDiffFactor. The diffusivity
increases as follows (section 2.5.1.1 of AdvCal_2019.03.fps):
term Si name=ArsenicIntDiffFactorDopant store add \
eqn= {(1.0+((PActive*0.6+AsActive)/[Arr 1.8e21 0.25])^3.5)}
term Si name=ArsenicVacDiffFactorDopant store add \
eqn= {(1.0+[ArrBreak 1.0 0 5.1920595e-7 -1.65 1050] \
*((PActive*0.3+AsActive)/2.1e20)^3.0)}
■ The pressure effects are redefined inside SiGe_and_Stress_Effect (section 2.14.2 of
AdvCal_2019.03.fps):
term Si name=ArsenicIntDiffFactorPressure store add eqn= 1.0
term Si name=ArsenicVacDiffFactorPressure store add eqn= 1.0
term Si name=ArsenicIntDiffFactor store add \
eqn= {ArsenicIntDiffFactorDopant * ArsenicIntDiffFactorPressure}
term Si name=ArsenicVacDiffFactor store add \
eqn= {ArsenicVacDiffFactorDopant * ArsenicVacDiffFactorPressure}
■ The dissolution of As clusters is modified in the presence of substitutional P using the term
ArsenicTClusterBackwardFac (section 2.5.2.1 of AdvCal_2019.03.fps):
term Si name=ArsenicTClusterBackwardFac store add eqn="1.0+PActive/1e20"

Depending on the process window of interest (window of As dose, As energy, P dose, P energy,
and annealing conditions), you might need some additional fine-tuning of the P or As
parameters to achieve a good fit between simulation and experimental data.

For more information, see Co-Diffusion of Arsenic and Phosphorus on page 123.

32 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Indium Parameters
See section 2.7 of AdvCal_2019.03.fps.

For silicon, the diffusivity values have been obtained by Synopsys from calibration of SIMS
data. The pairing constants are chosen such that the diffusivity of In–I pairs is approximately
equal to the diffusivity of free interstitials. The indium solid solubility and the clustering rate
have been calibrated by Synopsys.

The dose loss of indium during annealing is diffusion limited. Almost all indium atoms, which
diffuse to the Si–SiO2 interface, are built into the oxide. This is reflected by a very low
segregation coefficient. Furthermore, it is assumed that indium evaporates at the oxide–gas
surface.

The indium diffusion and dose loss is well calibrated for typical indium channel or halo
implantations below the amorphization dose. For high indium doses (typically
13 –2
> 5 × 10 cm ) and for the annealing of preamorphized wafers, the modeling of indium is not
accurate for the following reason: During solid phase epitaxial regrowth (SPER), indium atoms
are pushed towards the surface, due to a segregation effect between the crystalline and
amorphous phases of silicon [50][51]. This segregation increases the overall dose loss of
indium dramatically. However, the SPER is not modeled by default in the Advanced
Calibration of Sentaurus Process, but it can be enabled by the procedure
AdvancedSPERModel (see Solid Phase Epitaxial Regrowth on page 76).

Indium diffusion, clustering, and segregation in germanium has not yet been calibrated by
Synopsys. Therefore, the same parameter values are assumed in germanium as in silicon.

Antimony Parameters
See section 2.8 of AdvCal_2019.03.fps.

The macroscopic values for antimony diffusivity in silicon are based on the literature [1].
Antimony diffuses through Sb–V pairs. In highly doped regions (antimony concentration
20
> 2.0 × 10 ), the diffusivity is enhanced as observed by Larsen et al. [32].

The macroscopic values for antimony diffusivity in germanium are calibrated based on a single
literature source [37], and the diffusion of antimony is assumed to be dominated by Sb+V– –
pairs. For consistency with the diffusivity in silicon, Sb–V pairs with different charge states are
included as well. Their diffusivity is small in pure Ge, but is relevant for the mole-fraction
interpolation of Sb–V diffusivities in SiGe.

Advanced Calibration for Process Simulation User Guide 33

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

The pairing constants are chosen such that the Sb–defect pair diffusivity has a similar value as
the diffusivity of the unpaired defect.

The cluster parameters and the interface segregation have been calibrated by Synopsys based
on SIMS data and sheet resistance data. No ultrashallow junction Sb profiles in silicon have
been used for the calibration, therefore, the model parameters are not expected to be predictive
for Sb ultrashallow junction formation. The diffusivity of Sb in SiO2 is taken from Aoyama et
al. [52], and the ones in Si3N4 and GeO2 are assumed to be the same.

Carbon Diffusion
See section 2.9 of AdvCal_2019.03.fps.

Carbon Diffusivity in Silicon

The macroscopic diffusivity of carbon in silicon is taken from the literature [1]. The diffusivity
of C–I pairs has been calibrated by Synopsys. The Frank–Turnbull mechanism is switched off.
Carbon interstitials are only formed by the kick-out mechanism.

Carbon Clustering in Silicon

Four types of carbon–interstitial cluster are taken into account: C2, C2I, C3I2, and C3I3. The
following reactions for the formation and dissolution of clusters are considered:
■ C–I + C <=> C2I
■ C2 + I <=> C2I
■ C2I + C–I <=> C3I2
■ C3I2 <=> C3I3 + V

The clustering rates, together with the diffusivity of C–I pairs, have been calibrated by
Synopsys using SIMS data from marker layer experiments [53][54][55] and from data on
ultrashallow junction formation following Ge+C+B [56] and Ge+C+BF2 ‘cocktail’
implantations. The hopping length of C–I pairs is calibrated in the Sentaurus Process defaults.
It is assumed that, in regions that are amorphized by ion implantation and recrystallized by
20 –3
solid phase epitaxy, carbon is in the substitutional state up to concentrations of 3.0 ×10 cm
and else in C2 clusters immediately after the recrystallization. In contrast, in nonamorphized
regions, carbon is assumed to be mostly in C3I2 clusters at the beginning of thermal annealing.

During the formation of C–I clusters, the concentration of free interstitials is reduced and
vacancies are created. As a consequence, B diffusion is retarded and Sb diffusion is enhanced.
The increase of the solid solubility of boron in regions of high carbon concentration [56] is not
taken into account in the Advanced Calibration.

34 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

The C–I clustering model allows you to obtain accurate results also for the analysis of
phosphorus ultrashallow junction formation by Si+C+P ‘cocktail’ implantations and
subsequent spike annealing [57][58].

Carbon in Germanium

According to the literature [59], carbon diffuses very slowly through interstitials in germanium.
Carbon diffusion is modeled by the ChargedPair model, but it is assumed to be uncharged.

Detailed information or experimental data on the solubility or substitutionality of carbon in

germanium after implantation, SPER, or annealing is currently missing in the literature.
Therefore, the current assumption is that carbon has a high solubility and initial activation after
implantation (AcInit and AmInit). The Transient cluster model is used for carbon
clustering, with two carbon atoms assumed to form a cluster without interstitials.

Based on atomistic simulation [60][61] and comparison with experiments [59], carbon forms
mixed clusters with arsenic or phosphorus and vacancies, which reduce dopant diffusion.
Mixed As–C–V and P–C–V clusters are modeled by the ComplexCluster model. The
clustering and dissolution rates are calibrated [23] based on literature data [45][59].

Based on the experimental data available, carbon shows no significant dose loss at the
Ge–GeO2 interface. Therefore, the simple Segregation model is used with balanced
segregation.

Intrinsic Carrier Concentration

See section 2.10 of AdvCal_2019.03.fps.

For silicon, the intrinsic carrier concentration ni is taken from Morin and Maita [62]. As an
alternative, a simplified formula (ni_MM_simple) has been prepared. This simplified formula
is frequently used in other silicon process simulators.

The intrinsic carrier concentration is based on the one from silicon and corrections for band
gap and effective density-of-states for SiGe with Ge mole fraction 1 (see Impact of Stress on
Electrostatic Potential on page 47).

Advanced Calibration for Process Simulation User Guide 35

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Oxidation
See section 2.11 of AdvCal_2019.03.fps.

By default, the parameters of the Deal–Grove model [63] for thick oxides and the parameters
of the Massoud model for the initial regime of the oxidation are well calibrated for silicon.
Only the Massoud model parameters for wet oxidation are redefined in Advanced Calibration,
as described in Massoud Model Parameters for Wet Oxidation of Silicon.

For germanium oxidation, only dry oxidation of Ge (100) and (111) surfaces [64] has been
calibrated, and the calibration parameters are available as Sentaurus Process defaults.

SiGe oxidation has been calibrated and calibration parameters are included in Advanced
Calibration, as described in Parameters for Wet and Dry Oxidation of SiGe on page 36.

Massoud Model Parameters for Wet Oxidation of Silicon

Advanced Calibration contains parameters for the Massoud model for wet (H2O) atmosphere.
According to this model, for very thin oxides (thickness < 15 nm), the silicon oxidation rate is
increased with respect to the Deal–Grove model for oxidation.

Without Advanced Calibration, the parameters C0 and CW default to zero for wet oxidation,
which means that the Massoud model is effectively switched off and underestimates the oxide
growth of thin oxides in wet atmosphere.

For silicon surfaces with a (100) crystal orientation, the calibration gives an immediate
accuracy of ± 15% for oxidation rates in wet or mixed atmosphere. For other crystal
orientations, less data was available for calibration and the parameters are less reliable. If fine-
tuning is required, you need to consider adjusting the parameters C0 ( ± 25% ) and L0 ( ± 25% ).
If this is not sufficient, you can consider other parameters for calibration such as the initial
oxide thickness, or doping and pressure dependency of the oxidation rate. For the pressure
dependency of the Massoud correction term, the parameter value MassoudPress=1.0 is
proposed. If necessary, slightly smaller values (for example, 0.8 or 0.9) can be considered for
fine-tuning.

Parameters for Wet and Dry Oxidation of SiGe

Advanced Calibration contains parameters for the Deal–Grove model and the Massoud model
for wet and dry oxidation of SiGe. The parameters for Oxide_Germanium, the upper mole-
fraction limit for SiGe–SiO2 interfaces, are included in section 2.11.2 and section 2.11.3 of
AdvCal_2019.03.fps, and are effective when SiGe parameter interpolation is enabled (see
Effect of Germanium and Stress on page 38).

36 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

For details on the calibration of oxidation rates, see SiGe Oxidation on page 46.

Smoothing of Amorphous–Crystalline Interface

See section 2.12 of AdvCal_2019.03.fps.

After ion implantation some silicon regions can be amorphized. By default, the solid phase
epitaxial regrowth is not simulated explicitly in Advanced Calibration, but the initial conditions
for point defects and dopant activation are set differently in amorphized or crystalline regions.

At the amorphous–crystalline interface, there is a sharp step in the initial conditions for the
diffusion solver. Unless the mesh is very fine at the amorphous–crystalline interface, this might
lead to a numeric error in the result. To reduce this error, Sentaurus Process can smooth the
initial point-defect concentration at the amorphous–crystalline interface with the parameter
AmorpGamma. Synopsys has investigated typical situations and has found that the value
AmorpGamma=0.8 often gives the smallest numeric error.

In simulations with coarse meshes (mesh spacing of 2–4 nm across amorphous–crystalline

interfaces), the smallest numeric error is obtained with even smaller values of AmorpGamma,
for example, AmorpGamma=0.6.

Selecting Implantation Tables

See section 2.13 of AdvCal_2019.03.fps.

For analytic implantation of BF2 and indium, improved tables are switched on by the lines:
pdbSet ImplantData BF2 TableVersion 2008.09
pdbSet ImplantData In TableVersion 2008.09

For high-energy boron implantations, an improved table is available, but it is not switched on
by default. To select it for energy > 200 keV, set the switch:
pdbSetBoolean ImplantData B UseHighEnergyBoronTable 1

For details, refer to the corresponding explanations in Implantation Preprocessing and

Postprocessing on page 65.

Advanced Calibration for Process Simulation User Guide 37

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Effect of Germanium and Stress

See section 2.14 of AdvCal_2019.03.fps.

Strained and relaxed SiGe, strained-silicon layers, and process-induced stress are used widely
in state-of-the-art silicon process technology. Section 2.14 of AdvCal_2019.03.fps is
dedicated to the effects of Ge concentration and stress on the diffusion, clustering, and
segregation of dopants and point defects. Some models and part of the calibration are
summarized in [65].

NOTE In the Advanced Calibration models for the impact of Ge on dopant and
defect diffusion, it is assumed that germanium is treated as an impurity
in silicon (rather than being present in a separate ‘SiGe’ material).

The calibration focuses on the full Ge mole fraction range from 0% to

100%. However, it is less predictive for very high concentrations of Ge
(> 50%). Most of the experimental data available for calibration covers
only low Ge mole fractions (< 40%) or pure Ge (100%).

The calibration of stress and Ge effects is based on a relatively small set

of measured data, many of them published marker layer experiments,
which focus on equilibrium conditions. The models have not yet been
tested rigorously against SIMS data for nonequilibrium annealing
conditions and against electrical data from real-device fabrication
processes.

The bandgap narrowing due Ge content is always taken into account.

The diffusivity for Ge in silicon and oxide [66] are defined for constant diffusion models.
However, the Ge diffusion is only assumed to be constant if no SiGe effects are enabled as
described in the following.

The calibration of additional physical effects related to Ge or stress is contained in the

procedure SiGe_and_Stress_Effect. It has four Boolean arguments: Ge_Chem_Eff,
Stress_Eff, Segreg_Model, and Strained_Overlayer. Each of these, if set to 1, will
switch on additional physical effects. By default, none of these additional effects is switched
on. This saves CPU time in situations where the effects can be neglected. After sourcing
AdvCal_2019.03.fps, you can switch on the required effects by a single command.

For example, the following command switches on the first three effects (chemical effect of
germanium, stress effects in bulk silicon, and segregation at Si–SiGe interfaces), but not the
fourth effect (strained overlayer):
SiGe_and_Stress_Effect 1 1 1 0

38 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Arguments of SiGe_and_Stress_Effect
Before describing in detail the impact of SiGe_and_Stress_Effect, the impact of each
argument is summarized individually.

Ge Chemical Effect (Ge_Chem_Eff)

Switching on Ge_Chem_Eff enables SiGe parameter interpolation between the silicon and
germanium parameters.

NOTE Interpolation is performed only in silicon regions where the maximum

Ge concentration exceeds the threshold concentration Min.Conv.Conc
defined by the parameter pdbGet SiGe Ge Min.Conv.Conc. By
default, Min.Conv.Conc is equal to 5e20.

Germanium content dependency of the following parameters is taken into account in

AdvCal_2019.03.fps:
■ Band gap, carrier effective density-of-states (DOS), intrinsic carrier concentration, and
permittivity
■ Point-defect diffusivities, equilibrium concentrations, and surface recombination rates
■ Dissolution rate of {311} defects, and rate of {311} defect transformation to dislocation
loops
■ B, As, P, and Sb diffusivity, hopping length, solubility, clustering rate, dose loss, and initial
conditions after implantation

The value of Ge_Chem_Eff controls the definition of the xMoleFraction field saved into the
TDR file and controls whether the new or the old syntax is used for interpolation of mechanical
parameters and the lattice mismatch stress simulation. For details, see Overview on page 259.

In addition, an equation for Ge diffusion in SiGe is defined, which takes into account the
dependency of Si or Ge interdiffusion on the point-defect concentration.

Stress Effect (Stress_Eff)

Switching on Stress_Eff causes additional physics to be taken into account. The stress effect
is included for the following parameters:
■ Band gap
■ Point-defect equilibrium concentration and diffusivity
■ Dissolution rate of {311} defects, and rate of {311} defect transformation to dislocation
loops

Advanced Calibration for Process Simulation User Guide 39

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

■ B, As, P, Sb, In, and Ge diffusivity

■ B, As, P, and Sb solid solubility
■ Formation energy of BICs and As–V clusters

Except for the additional bandgap narrowing, which depends on individual strain-tensor
components, only the hydrostatic pressure is used for the modification of the diffusion and
activation models. Anisotropic effects, which would require the use of diffusivity tensors, are
2
not included in Advanced Calibration. The unit of pressure in Sentaurus Process is dyn/cm
2
( 10 dyn/cm = 1 Pa).

Sentaurus Process takes into account various stress-causing mechanisms (see Sentaurus™
Process User Guide). One of these is compressive stress from lattice mismatch due to the
presence of impurities. The lattice mismatch due to Ge and C is taken into account, with the
parameters adjusted according to the literature (see Mole Fraction–Dependent Mechanics
Parameters for SiGe on page 261). In typical cases, the impact of carbon-induced strain on
dopant diffusion and activation is small, but not negligible. Stress from lattice mismatch due to
high concentrations of other stress-causing impurities, such as B or extended defects, has not
been taken into account in the calibration of Advanced Calibration.

This additional stress source can be switched on with the strain_profile command, as
shown below, but – since all models are calibrated without this additional stress source –
switching on stress from impurities other than Ge might require a recalibration of other process
model parameters, such as B solid solubility:
strain_profile Si species=Boron strain= { 0 -0.30 } ratio= { 0 1 }

Segregation at Si–SiGe Interface (Segreg_Model)

Boron and phosphorus have been found to segregate at Si–SiGe interfaces, in experiments
using epitaxial layers of strained SiGe embedded in Si, or strained silicon embedded in relaxed
SiGe. Boron segregates into SiGe; whereas, phosphorus is pushed out of the SiGe region. A
theoretical analysis [67] indicates that the dominant driving force for this segregation is the
pressure gradient in the case of boron, and the difference in band structure between Si and SiGe
in the case of phosphorus. The same but weaker effect as for phosphorus is assumed for arsenic
as well.

Effect of Strained Overlayers (Strained_Overlayer)

The model developed by Cowern [68] predicts an influence of overlayer strain on the point-
defect concentration in the silicon. That effect was confirmed by experiments showing a strong
dopant diffusion change under high-level stressed nitride layers [69][70].

40 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

The change of equilibrium concentration due to overlayer strain can be expressed by the factor
exp ( P ov V f ⁄ kT ) , in which P ov is the hydrostatic pressure in the overlayer, and V f is the
volume of an atom in silicon. This effect is switched on by setting the argument
Strained_Overlayer in the procedure SiGe_and_Stress_Effect to 1.

Since the impact of strained overlayers is not fully understood yet (especially in the situation
where thin oxides are in between silicon and strained overlayers), the procedure
SiGe_and_Stress_Effect is most widely applied with the arguments 1 1 1 0:
SiGe_and_Stress_Effect 1 1 1 0

Implementation of SiGe_and_Stress_Effect

Chemical SiGe Alloy Effects

If Ge_Chem_Eff is equal to 1, then the SiGe parameter interpolation between the silicon and
germanium parameters is enabled by:
pdbSet Si Skip.Parameter.Interpolation 0
pdbSet Ge Skip.Parameter.Interpolation 0

This also switches on the following mapping of Ge concentration to the Ge mole fraction in Si:
C Ge
xMoleFraction = 2 ------------------------------------------------------------------------------------------- (2)
2
LD Si + LD Si + 4C Ge ( LD Ge – LD Si )

This complex formula (and a similar one for Ge material) replaces the simpler formula:
C Ge
xMoleFraction = ------------ (3)
LD Si

which is used for Skip.Parameter.Interpolation 1. xMoleFraction is used for

parameter interpolation in SiGe in process and device simulation models. The complex
formula gives slightly higher xMoleFraction for the same Ge concentration, due to the
different lattice constants of Si and Ge.

Impact of Germanium on Electrostatic Potential

The energy bandgap change along with mole fraction shows a very nonlinear curve due to the
transition between the X-valley ( x < 0.8 ) and L-valley ( x > 0.8 ) [71]. With enabled SiGe
parameter interpolation, the parameter delEg.Function for the generic interpolation method
is used for SiGe to model such nonlinearity.

Advanced Calibration for Process Simulation User Guide 41

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

The following default bandgap narrowing expression, fitting experimental data based on
optical measurements from [71], is used:
2 2
ΔEg = ( 0.3758 × ( 0.7795 – x ) – 0.22835 ) – 0.2406 × exp ( – ( ( 1 – x ) ⁄ 0.1176 ) ) (4)

Additional effects are taken into account as follows:

■ The permittivity of Ge-doped silicon is calculated by the following formula, in which xGe
is the Ge concentration in Si, ε Si = 11.7 and ε Ge = 15.8 :
ε = ( 1 – x Ge ) × εSi + x Ge × ε Ge (5)

■ The mole fraction dependency of the product of the DOS in the conduction and valence
bands, N C × N V , is assumed to be linear for computational efficiency ( 1 – x Ge × 0.705 ) ,
which is a good approximation of the following nonlinear factor taken from [72]:
3⁄4
( 1 – x Ge × 0.47 ⁄ 0.81 ) (6)

NOTE Since the Potential Use.DOS model is disabled, the (small) change
of electron affinity between Si and SiGe is not taken into account in a
precise way. For simplicity, it is assumed that the midgap energy level
does not depend on the Ge concentration.

Impact of Germanium on Point Defects

In contrast to III–V compound semiconductors, which have well-defined stoichiometry with

distinct lattice sites for group III and group V atoms, SiGe allows arbitrary stoichiometry and
random ordering of Si and Ge atoms. As a consequence, in SiGe, there is no need to
differentiate between Si-like and Ge-like interstitials, or to define Si-type and Ge-type
vacancies. Instead, it is fully sufficient to consider the density of ‘interstitials’ (extra (Si or Ge)
atoms in the lattice) and ‘vacancies’ (vacant lattice sites) in the model.

In the literature, there is no agreement on the impact of Ge on point-defect parameters except

for the vacancy transport capacity and equilibrium concentration. Computational studies of
vacancies in SiGe have found an ~1.0 eV vacancy formation energy decrease in Ge [73][74]
[75]. However, in Advanced Calibration, a decrease of the V formation energy of only 0.39 eV
is assumed in germanium. The main difference of the transport capacities for silicon and
germanium is modeled by different vacancy diffusivities. The interpolation between silicon
and germanium is parabolic in logarithmic scale and was calibrated based on SiGe
interdiffusion experiments [76].

Both the interstitial equilibrium concentration and the interstitial diffusivity in Ge are assumed
to be larger than in silicon. In SiGe, a parabolic mole-fraction interpolation in logarithmic scale
is applied for interstitial diffusivity, as well as for the equilibrium concentration.

42 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

The I–V bulk recombination in SiGe is increased by a mole fraction–dependent factor

BulkRecFac, which equals 1.0 for pure Si and pure Ge, and has its maximum value for
Si50Ge50. This increased I–V recombination in SiGe has been found to be beneficial for the
modeling of phosphorus transient-enhanced diffusion in SiGe after low-dose or high-dose P
implantation. I–V recombination reduces the P diffusion tail from P–I diffusion most
efficiently for high-dose P implantations.

To ensure the same surface recombination length of point defects for Si, SiGe, and Ge surfaces,
the SiGe interpolation of the surface recombination rate follows the interpolation of the point-
defect diffusivities.

Impact of Germanium on Extended Defect Parameters

The {311} interstitial clusters are less stable in the presence of Ge, and the transformation of
{311} defects into dislocation loops is faster [3][77][78]. The effects have been calibrated by
Synopsys based on experimental data generated from the ATOMICS research project [3].

For the Moment model for I-clusters, the parameters for growth (Ikfi and Ikfc) and
dissolution (Ikr) are interpolated between Si and Ge, parabolic, and in logarithmic scale.

Since the Full model for I-clusters is not defined for Ge, the parameter adjustment prefactors
C311DiffIntFactor and CLoopTransfer are used instead of parameter interpolation
between Si and Ge. The impact of Ge is taken into account only if Ge_Chem_Effect=1:
term Si name= C311DissIntFactor store add \
eqn= "(1.0+2.0*xMoleFraction+0.5*xMoleFraction*xMoleFraction)"
term Si name= CLoopTransfer store add \
eqn= "(1.0+25.*xMoleFraction+18.*xMoleFraction*xMoleFraction)"

Note that the terms for C311DissIntFactor and CLoopTransfer as listed above do not
include the effect of stress, which is included in the Advanced Calibration file (see Impact of
Stress on Extended Defect Parameters on page 49).

SiGe Interdiffusion

The interdiffusion model [79][80] is switched on to describe interdiffusion of Si and Ge. Its
parameters were extracted and calibrated against data from [81][82][83][84][85]. While the
(dominant) interdiffusivity from vacancies can be extracted reliably, the parameter error might
be larger for the smaller component of interdiffusivity from the diffusion of interstitials. Data
for interdiffusion during oxidation [86] was used to test assumptions on the parameters for
SiGe interdiffusion using interstitials.

The biggest uncertainties in calibration exist for highly n-doped SiGe, where interdiffusion is
believed to have a dominant contribution from P–V or As–V pairs, which is not yet included
in the calibration, due to lack of experimental data suitable for calibration. To include

Advanced Calibration for Process Simulation User Guide 43

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

contributions from dopant–vacancy pair diffusion in SiGe interdiffusion, you can define the
term InterVDiffFactor, which is 1.0 by default, as a function of dopant concentration. This
factor is then multiplied to the interdiffusion caused by V diffusion in SiGe.

By default, for simplicity, no drift component is assumed in SiGe interdiffusion. As Ge atoms

are larger in size than Si atoms, you might expect that a gradient in pressure adds a drift term
to SiGe interdiffusion. To switch on this drift term, insert the following lines after calling the
procedure SiGe_and_Stress_Effect:
pdbSet Si Ge Interdiffusion.Drift 1
pdbSet Ge Si Interdiffusion.Drift 1
pdbSetDouble SiGe Vac D.Inter.X2 {[Arr 1.2e5 1.52]}

Here, the first two lines switch on the drift contribution to interdiffusion, which effectively
increases Si and Ge interdiffusion because the pressure gradient is, in many cases (and
precisely in all 1D cases), parallel to the gradient of Ge concentration. The third line sets the
interdiffusivity in SiGe using vacancies to a value smaller than the one calibrated for
Interdiffusion.Drift 0. With this reduced value, the published interdiffusivity data used
for model calibration matches Interdiffusion.Drift 1.

Germanium Effect on Dopant Diffusivity

Arsenic, Sb, and Ge diffusion in relaxed SiGe is much faster than in Si [87][88][89]. A
nonlinear dependency of the vacancy-mediated diffusivity activation energy on the Ge content
was first measured for Sb by Larsen and Kringhoj [87]. In Advanced Calibration, it is assumed
that the vacancy components of the As, P, and Sb diffusivities show a similar trend with Ge
mole fraction and temperature, as the diffusivity of free vacancies. Therefore, the calibration
of the corresponding parameters for parabolic mole-fraction interpolation started from the
values for vacancies and has been fine-tuned with the help of SIMS data provided by
AMAT-VSE1 [76].

There is a lack of published experimental work devoted to P diffusion in SiGe, especially for
high Ge mole fractions. The marker layer experiments performed by Zangenberg [89] indicate
an increase of diffusivity with small Ge contents up to ~25%, and saturation or even a little
decrease of Ge content of 40%. However, P diffusivity in pure Ge [23] is much higher than the
one in SiGe with Ge content ≤ 40% . The parameters for logarithmic mole-fraction
interpolation of P–V and P–I diffusivities have been calibrated against dedicated data provided
to Synopsys by AMAT-VSE [76].

According to most experimental work [90][91][92], interstitial-meditated B diffusion in

relaxed SiGe is retarded with increasing Ge content up to ~50%. For higher Ge contents, the
boron diffusion is again enhanced [93]. Therefore, a linear expression for the diffusivity
activation energy as derived by Ahn [94] for small Ge mole fractions is not adequate, and a

1. Applied Materials - Varian Semiconductor Equipment.

44 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

parabolic one is used for the full Ge mole fraction range instead. Finally, the retardation effect
of F on B diffusion decreases linearly with increasing Ge mole fraction.

Table 4 summarizes the SiGe interpolation for the diffusivity components for different dopants.

Table 4 Germanium chemical effect on dopant diffusivity components

Diffusivity component Logarithmic parabolic interpolation Comments on data used for calibration
factor

Sb–V ln ( [Arr 1.5e12 3.22] ) × x Ge × ( 1 – x Ge ) Marker layer experiments [87].

Dedicated experiments from AMAT-VSE.

As–I 0 Marker layer experiments [88].

USJ experiments [95][96][97].

As–V ln ( [Arr 9e10 3.22] ) × x Ge × ( 1 – x Ge ) Marker layer experiments [88].

USJ experiments [95][96][97].
Dedicated experiments from AMAT-VSE.
–3
B–I ln ( 8.0 × 10 ) × x Ge × ( 1 – x Ge ) Marker layer experiments [90].
USJ experiments [93][98].
Dedicated experiments from AMAT-VSE.

P–I ln ( [Arr 8e13 2.13] ) × x Ge × ( 1 – x Ge ) Marker layer experiments [89].

USJ experiments [95].
Dedicated experiments from AMAT-VSE [76].

P–V ln ( [Arr 3e13 3.46] ) × x Ge × ( 1 – x Ge ) Marker layer experiments [89].

USJ experiments [95].
Dedicated experiments from AMAT-VSE [76].

Germanium Effect on Dopant Activation

The Solubility of boron is slightly higher in SiGe than in pure Si. For the Solubility of
boron in SiGe, parabolic mole-fraction interpolation is used in logarithmic scale. For the
Solubility of As, P, and Sb in SiGe, parabolic mole-fraction interpolation is used in linear
scale.

The dopant clustering reaction rates CluRate are interpolated parabolically in logarithmic
scale.

Germanium Effect on Interface Segregation

The interface segregation has been calibrated based on SIMS. For B, As, and P, the
ThreePhaseSegregation parameters for trapping and emission rates as well as the interface
trap number CMax are interpolated between Si and Ge in logarithmic scale. In addition,
parabolic interpolation is used for CMax for As and P. The interstitial fraction of dopant
trapping in equilibrium is interpolated in linear scale.

Advanced Calibration for Process Simulation User Guide 45

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Germanium Effect on Initial Conditions After Implantation

For the initial activation parameters AmInit and AcInit, linear interpolation is used for SiGe
in general. In the case of boron and BF2, logarithmic interpolation is used inside the
aminit_Boron and aminit_BF2 procedures, which define the initial activation in
amorphized regions.

The ifactor for SiGe of both As and P is interpolated linearly between Si and Ge. The
interpolation is defined in the ifactor_arsenic and ifactor_phosphorus procedures.
Similarly, mcdfactor for B, As, and P is subject to linear mole-fraction interpolation.

SiGe Oxidation

During oxidation of SiGe, Ge enhances oxidation rates and is completely rejected from the
oxide, so that it piles up at the SiO2–SiGe interface [99]. This effect is exploited by the Ge
condensation technique to enrich the Ge mole fraction of SiGe layers on insulators by
oxidation [100]. The selective Si consumption by oxidation at the SiO2–SiGe interface is
compensated by the thermal diffusion of Si from the bulk SiGe layer. However, when the Ge
composition of the SiGe layer exceeds approximately 80%, the Si flux by thermal diffusion
from the bulk SiGe layer cannot compensate the Si consumption by oxidation, resulting in the
simultaneous oxidation of Si and Ge [101].

The parameters for dry SiGe oxidation were extracted and calibrated against literature data for
oxide growth rates [100][101][102][103][104][105][106][107] and Ge pileup profiles [101]
[103][106][107]. The oxidation rate depends on the Ge mole fraction on the SiGe side of the
SiO2–SiGe interface, which itself strongly depends on SiGe interdiffusion, which was first
calibrated mainly based on inert anneals (see SiGe Interdiffusion on page 43). For strong Ge
pileup by oxidation leading to high Ge mole fractions and compressive strain, SiGe
interdiffusion is enhanced, which again leads to a lower and wider Ge pileup. Unfortunately,
this interdependency is difficult to calibrate because the experimental Ge profiles available
have, in general, limited accuracy. In addition, some uncertainties in the calibration exist for
the influence of oxidation on the strain by SiGe lattice mismatch and on SiGe interdiffusion.
Moreover, the self-limiting oxidation effect due to melting of the Ge pileup layer at the
oxide–SiGe interface [108], or other causes [109], has not been included in the calibration.

The selective Si oxidation in general and the combined Si and Ge oxidation only at high Ge
–3
mole fraction (Ge concentration > 3.8e22 cm ) have been implemented by terms for the
segregation ratio and the transfer rate for Ge at the SiO2–SiGe interface, which depend on the
oxidation rate (ReactionSpeed) and Ge concentration in silicon (Germanium_Silicon):
term Ox /Si name=GermaniumSegregationFactor store add \
eqn= {(1.0+ReactionSpeed*[Arr 1.0e26 2.0] * \
(Germanium_Silicon>3.8e22 ? 0.0 : (1.0-Germanium_Silicon/3.8e22)) * \
(Germanium_Silicon>1.0e18 ? 1.0 : (Germanium_Silicon/1.0e18)))}

46 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

term Ox /Si name=GermaniumTransferFactor store add \

eqn= {(1.0+ReactionSpeed*[Arr 1.0e35 4.0] * \
(Germanium_Silicon>3.8e22 ? 0.0 : (1.0-Germanium_Silicon/3.8e22)) * \
(Germanium_Silicon>1.0e18 ? 1.0 : (Germanium_Silicon/1.0e18)))}

In addition, the parameters for wet SiGe oxidation were extracted and calibrated against
literature data for oxide growth rates [110][111]. The accuracy is limited in general.

Oxidation-enhanced diffusion is suppressed already for small Ge mole fractions [19]. This
effect is achieved by logarithmic scale interpolation of the interstitial injection parameter
theta between a high value for Oxide_Silicon and a low value for Oxide_Germanium
(see Oxidation-Enhanced Diffusion on page 20).

Stress Effects

To keep the implementation short, the abbreviations Vt_i, kT_i, PR, and PR_Si are defined.
The pressure abbreviations PR and PR_Si are limited by a maximum-allowed value of 4 GPa,
which is defined in section 2.14 by the following parameter:
pdbSetDouble SiGe Mechanics MaxPressure 4.0e10

The definition of PR is:

set maxP [pdbGet SiGe Mechanics MaxPressure]
if { $Stress_Eff } { set PR \
((Pressure>$maxP)?$maxP:((Pressure<(-$maxP))?(-$maxP):Pressure))
} else { set PR 0.0 }

See section 2.14.2 of AdvCal_2019.03.fps.

Impact of Stress on Electrostatic Potential

If Stress_Eff is equal to 1, bandgap narrowing due to stress is calculated according to the

deformation potential theory as published by [112]. If Ge_Chem_Eff is switched on as well,
the deformation potential constants are interpolated linearly between values for pure silicon
and values for pure germanium. If Ge_Chem_Eff is switched off, the silicon parameters are
used. The deformation potential constants for Si and Ge proposed by Van de Walle [113] were
chosen for Advanced Calibration.

The model is switched on using the command:

pdbSet Si Potential niMod StrainDependent

Advanced Calibration for Process Simulation User Guide 47

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

Impact of Stress on Point-Defect Parameters

The diffusivities and equilibrium concentrations of interstitials and vacancies are modified in
the presence of hydrostatic pressure using the terms IntDiffFactor, VacDiffFactor,
IntCStarFactor, and VacCStarFactor, which are defined for both silicon and
germanium, and are interpolated for SiGe:
pdbSetDouble Si Int idf -0.142e-23
pdbSetDouble Si Int icf -1.211e-23
pdbSetDouble Si Vac vdf 1.195e-23
pdbSetDouble Si Vac vcf 1.324e-23
pdbSetDouble Ge Int idf -0.142e-23
pdbSetDouble Ge Int icf -2.800e-23
pdbSetDouble Ge Vac vdf 8.000e-24
pdbSetDouble Ge Vac vcf 1.324e-23
pdbSetDouble SiGe Vac vdf.X2 1.9e-23
pdbSetDouble SiGe Vac vcf.X2 2.0e-23
term Si name=IntDiffFactor store add \
eqn= "exp($PR*[_MoleFraction::Param Si Int idf]*$kT_i)"
term Si name=VacDiffFactor store add \
eqn= "exp($PR*[_MoleFraction::Param Si Vac vdf]*$kT_i)"
term Si name=IntCStarFactor store add \
eqn= "exp($PR*[_MoleFraction::Param Si Int icf]*$kT_i)"
term Si name=VacCStarFactor store add \
eqn= "exp($PR*[_MoleFraction::Param Si Vac vcf]*$kT_i)"
term Ge name=IntDiffFactor store add \
eqn= "exp($PR*[_MoleFraction::Param Ge Int idf]*$kT_i)"
term Ge name=VacDiffFactor store add \
eqn= "exp($PR*[_MoleFraction::Param Ge Vac vdf]*$kT_i)"
term Ge name=IntCStarFactor store add \
eqn= "exp($PR*[_MoleFraction::Param Ge Int icf]*$kT_i)"
term Ge name=VacCStarFactor store add \
eqn= "exp($PR*[_MoleFraction::Param Ge Vac vcf]*$kT_i)"

In SiGe, the actual definition of the mole fraction–dependent terms is performed inside the
procedure _AlloyCompound, which is defined in section 2.14.5 of AdvCal_2019.03.fps.
(Sentaurus Process calls the _AlloyCompound procedure during diffusion preprocessing if
SiGe is to be modeled as an alloy. Terms including mole fraction–interpolated parameters are
best defined in this procedure.)

Compressive hydrostatic pressure decreases the equilibrium concentration of interstitials and

increases the equilibrium concentration of vacancies. If Stress_Eff=1, the diffusivities of
point defects are modified as a function of pressure. The activation volumes for the above
prefactors were derived from the induced strain values calculated by Diebel [114].

48 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

For the vacancy diffusivity, a parabolic decrease of the activation volume is assumed between
pure Si and pure Ge. The value for pure Ge has been adjusted to match experimental data
reported by Kawamura et al. [115].

The modified equilibrium concentrations are used automatically by Sentaurus Process in the
definition of surface boundary conditions.

The point-defect interface and bulk recombinations are diffusion-limited processes, but the
modified diffusivities are not taken into account automatically in surface boundary conditions
and bulk recombination rates. Therefore, together with changing the diffusivity of interstitials
and vacancies, the surface and bulk recombination rates are adjusted accordingly.

In the drift-diffusion equation for point defects, the pressure gradient is implicitly taken into
account as an additional force, which pushes interstitials towards regions of tensile stress and
vacancies towards regions of compressive stress.

Impact of Stress on Extended Defect Parameters

The {311} interstitial clusters are less stable in the presence of compressive hydrostatic
pressure, and the transformation of {311} defects into dislocation loops is faster [3][77][78].

The corresponding parameter adjustment prefactors IClusterDissIntFactor (used in the

1Moment model for I-clusters), as well as C311DiffIntFactor and CLoopTransfer (used
in the Full model for I-clusters), have been calibrated by Synopsys based on experimental
data generated within the ATOMICS research project [3].

The impact of stress is taken into account only if Stress_Eff=1:

term Si name=IClusterDissIntFactor store add eqn= "exp($PR*2.6e-23*$kT_i)"
term Si name=C311DissIntFactor store add \
eqn= "(1.0+2.0*xMoleFraction+0.5*xMoleFraction*xMoleFraction)* \
exp($PR*1.6e-23*$kT_i)"
term Si name=CLoopTransfer store add \
eqn= "(1.0+25.*xMoleFraction+18.*xMoleFraction*xMoleFraction)* \
exp($PR*3.3e-23*$kT_i)"

Note that the terms for C311DissIntFactor and CLoopTransfer as listed above also
include the effect of Ge as included in the Advanced Calibration file (see Impact of Germanium
on Extended Defect Parameters on page 43).

Impact of Pressure on Dopant Diffusivity

The diffusivity of dopants changes as a function of hydrostatic pressure. With increasing

compressive pressure, diffusion using dopant–interstitial pairs is reduced, and diffusion using
dopant–vacancy pairs is increased.

Advanced Calibration for Process Simulation User Guide 49

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

The following formula for the macroscopic diffusivity DAX using AX dopant–defect pairs is
used:
D AX ( P ) = D AX ( P = 0 ) exp ( – P V AX ⁄ kT ) (7)

where P is the hydrostatic pressure (positive sign means compressive pressure), and V AX is
the activation volume. The activation volume values for different dopants are listed in Table 5.

Table 5 Activation volumes for dopant diffusivity components in silicon

Diffusivity component prefactor Activation Comments on the data used
volume [cm3]

AntimonyVacDiffFactor -2.334e-23 Ab initio calculations of the strain impact by Ahn [94].

ArsenicIntDiffFactorPressure 1.746e-23 Ab initio calculations of the strain impact by Ahn [94].

ArsenicVacDiffFactorPressure -1.800e-23 Ab initio calculations of the strain impact by Ahn [94],

slightly reduced during calibration against experimental
data.

BoronIntDiffFactor 2.394e-23 Average value from ab initio calculations of the strain

impact for different diffusion directions by Diebel [114].

GermaniumDiffFactor -3.000e-23 Marker layer experiments by Zangenberg [89].

IndiumIntDiffFactor 0.936e-23 Ab initio calculations of the strain impact by Ahn [94].

PhosphorusIntDiffFactorPressure 2.200e-23 Average value from ab initio calculations of the strain

impact for different diffusion directions by Ahn [94],
slightly reduced during calibration against experimental
data.

PhosphorusVacDiffFactorPressure VacDiffFactor*
VacCStarFactor

The pressure dependency of diffusivities by different dopant–defect pairs is illustrated in

Figure 2 on page 51. The stress effect on dopant diffusivity is superimposed on the chemical
effect of Ge in a multiplicative way.

The stress effect is simulated by the activation energy correction using diffusivity prefactors.
For example, in the case of boron, it is performed by the term BoronIntDiffFactor, which
is defined in silicon:
term Si name=BoronIntDiffFactor store add eqn= "exp(-2.394e-23*$PR*$kT_i)"

During assembly of the diffusion equations, Sentaurus Process checks each dopant and
material to assess whether such diffusion factors exist. The diffusivity through
dopant–interstitial or dopant–vacancy pairs then is multiplied by the corresponding diffusion
enhancement factors.

50 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

P−V
As_V
Diffusivity Enhancement due to Stress

101 P−I
As−I
−
In−I
B−I

100

10−1

−10 1
Pressure [GPa]

Figure 2 Enhancement of macroscopic diffusivity using selected dopant–defect pairs as a

function of hydrostatic pressure, at T = 900oC in pure silicon

Impact of Pressure on Dopant Activation

If Stress_Eff is switched on, the solid solubility of dopants depends on the strain. In general,
for compressive strain, the solubility of atoms smaller than Si increases; whereas, the solubility
of larger atoms decreases. In the Transient model, the stress effect is taken into account by
introducing pressure-dependent correction factors to Solubility and TotalSolubility:
S ( P ) = S ( P = 0 ) exp ( – P V/kT ) (8)

where S is the dopant solid solubility, and V is the activation volume. The activation volume
values for different dopants are presented in Table 6. The same volumes were chosen for
Solubility and TotalSolubility.

Table 6 Activation volumes for dopant solid solubility

Dopant Activation Comments on the data used
volume [cm3]

Antimony 5.796e-24 Theoretical value from ab initio calculations of the induced strains by Ahn et
al. [116] multiplied by a factor of 0.6 according to experimental data [3].

Arsenic 1.062e-24 Ab initio calculations of the induced strains by Ahn et al. [116].

Boron -3.636e-24 Theoretical value from ab initio calculations of the induced strains by Ahn et
al. [116] multiplied by a factor of 0.2 according to experimental data [3].

Phosphorus -4.596e-24 Ab initio calculations of the induced strains by Ahn et al. [116].

Advanced Calibration for Process Simulation User Guide 51

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

The formulas for As and P are taken from ab initio calculations [116]. The impact of pressure
on B solubility is chosen to be only 20% of the theoretical prediction in [116], as this better fits
the experimental work performed in the ATOMICS research project [3] and experimental data
from B ultrashallow junction formation with different variants of Ge preamorphization. The
impact of pressure on Sb solubility is chosen to be 60% of the theoretical calculation in [116].
The relative change of solid solubility for different dopants as a function of pressure is shown
in Figure 3.

As
B
P
Sb
Solid Solubility Change due to Stress

100

-1 0 1
Pressure [GPa]

Figure 3 Relative change of solid solubility of As, B, P, and Sb as a function of pressure,

at T = 900oC

An example of the definition of B pressure-dependent solid solubility is:

pdbSetString Si B SS.Factor "exp(3.636e-24*$PR*$kT_i)"
pdbSetString Si B Total.SS.Factor "exp(3.636e-24*$PR*$kT_i)"

Since the emission rate for the silicon side in the three-phase segregation models is
proportional to the solid solubility, corresponding modification should also be included in the
boundary condition. For example, it was performed by the following line for B:
pdbSetString Si B Side.SS.Factor "exp(3.636e-24*$PR_Si*$kT_i)"

NOTE Arbitrary Alagator expressions for the dopant solid solubility prefactors
can be defined by users in Sentaurus Process. The names of terms used
for the solid solubility, the total solid solubility, and the emission rate
correction are SS.Factor, Total.SS.Factor, and
Side.SS.Factor, respectively.

52 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

In the ChargedCluster model, the solid solubility is not defined explicitly but results from
the formation energies of the various clusters. Therefore, the formation energies of the clusters
are modified as a function of pressure. This has been performed in such a way that the overall
effect of pressure on dopant activation is similar as in the Transient model. The calibration
is arbitrary, because not enough experimental data is available to calibrate the effect of pressure
on each BIC or AsV cluster individually.

Segregation at Si–SiGe Interface (Segreg_Model)

Boron and P have been found to segregate at Si–SiGe interfaces, in experiments using epitaxial
layers of strained SiGe embedded in Si, or strained Si embedded in relaxed SiGe. Boron
segregates into SiGe; whereas, P is pushed out of the SiGe region. A theoretical analysis [67]
indicates that the dominant driving force for this segregation is the pressure gradient in the case
of B, and the difference in band structure between Si and SiGe in the case of P.

If Segreg_Model is switched on, an additional driving force for the drift-diffusion of B, As,
and P is taken into account by using the PDependent model for both Si and Ge (not shown
here):
if { $Segreg_Model } {
pdbSetSwitch Si B StressModel PDependent
term Si name=BoronIntSSFactor store add \
eqn= "exp($Vt_i*(0.5*(Pressure/5.11e10)))"
term Si name=BoronIntSPFactor store add \
eqn= "exp($Vt_i*(0.5*(Pressure/5.11e10)))"

pdbSetSwitch Si AsStressModel PDependent

term Si name=ArsenicIntSSFactor store add \
eqn= "exp($Vt_i*(-0.2*xMoleFraction))"
term Si name=ArsenicIntSPFactor store add \
eqn= "exp($Vt_i*(-0.2*xMoleFraction))"
term Si name=ArsenicVacSSFactor store add \
eqn= "exp($Vt_i*(-0.2*xMoleFraction))"
term Si name=ArsenicVacSPFactor store add \
eqn= "exp($Vt_i*(-0.2*xMoleFraction))"

pdbSetSwitch Si P StressModel PDependent

term Si name=PhosphorusIntSSFactor store add \
eqn= "exp($Vt_i*(-0.6*xMoleFraction))"
term Si name=PhosphorusIntSPFactor store add \
eqn= "exp($Vt_i*(-0.6*xMoleFraction))"
term Si name=PhosphorusVacSSFactor store add \
eqn= "exp($Vt_i*(-0.6*xMoleFraction))"
term Si name=PhosphorusVacSPFactor store add \
eqn= "exp($Vt_i*(-0.6*xMoleFraction))"
}

Advanced Calibration for Process Simulation User Guide 53

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 2: Constant Parameters

In this calibration, the potentials behind the additional driving forces can be expressed as:
E mech,B = – 0.5 × Pressure ⁄ 5.11e10 (9)

E chem,As = 0.2 × x Ge (10)

E chem,P = 0.6 × x Ge (11)

where:
■ E mech,B is the mechanical potential of substitutional boron.
■ E chem,As and E chem,P are the chemical potential of substitutional As and P, respectively, in
eV.
■ Pressure is the hydrostatic pressure in dyn/cm2.
■ x Ge is the fractional Ge concentration.

These formulas have been calibrated to experimental data [117][118], as illustrated in Figure 4.
It should be noted that all measured data, which can be directly used for calibration, is extracted
from plane wafer experiments with SiGe marker layers. In these wafers, the biaxial stress is
approximately proportional to the Ge concentration. At present, a clear separation, whether the
segregation is caused by a mechanical or an electrochemical potential, is not possible from
available experimental data.

Ge, after anneal, simulated

B, after anneal, SIMS 4e+21
1019

Germanium Concentration [cm–3]

B, after anneal, simulated
Boron Concentration [cm−3]

B, initial
3e+21

1018 2e+21

1e+21

1017 0
0 0.2 0.4 0.6 0.8
Depth [µm]
Figure 4 B diffusion (56 hours at 850oC) in SiGe marker layers; SIMS data is extracted
from [117] and simulations are performed with "SiGe_and_Stress_Effect 1 1 1 0"

54 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

Effect of Strained Overlayers (Strained_Overlayers)

The model developed by Cowern [68] predicts an influence of overlayer strain on the point-
defect concentration in the silicon. That effect was confirmed by experiments showing a strong
dopant diffusion change under high-level stressed nitride layers [69][70].

The change of equilibrium concentration due to overlayer strain can be expressed by the factor
exp ( P ov V f ⁄ kT ) , in which P ov is the hydrostatic pressure in the overlayer, and V f is the
volume of an atom in Si. This effect is switched on by setting the argument
Strained_Overlayers in the procedure SiGe_and_Stress_Effect to 1.

Unfortunately, it is not clear yet from experiments how intermediate layers such as an oxide
layer between strained nitride and silicon influence the point-defect boundary conditions. With
the current model implementation, a thin, unstrained, oxide layer would shield very effectively
the impact of a strained SiN overlayer, in contrast to what has been observed in experiment [3].
Therefore, you need to be very careful in using this model in 2D or 3D process simulations,
which usually include native oxide layer deposition steps. In addition, as thermal expansion is
another source of stress, the temperature at which native oxide layers are deposited in the
process simulation can have a significant impact on point-defect concentrations at
oxide–silicon boundaries, and inaccuracies in thermal expansion coefficients at elevated
temperatures will translate into inaccurate calculations of point-defect concentrations.

NOTE It is recommended to define the same internal stress values in the native
oxide as in the nitride to obtain the strained overlayer effect on the point-
defect concentration in silicon.

Part 3: Initial Conditions After Ion Implantation

In this part of the Advanced Calibration file, the initial conditions for diffusion data fields after
ion implantation are set immediately after implantation or immediately before the diffusion. In
particular, the concentration of point defects or point-defect clusters present after ion
implantation is specified dependent on the implantation species, energy, and dose.
Furthermore, a framework is provided for a user-defined specification of initial electrical
activation of as-implanted dopants.

In general, the initial conditions are important for annealing steps with small thermal budget.
This is typically the case in deep-submicron technology.

The contents of part 3 of the Advanced Calibration file are described in the next sections.

Advanced Calibration for Process Simulation User Guide 55

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

User-Defined Defect Initialization

See sections 3.1, 3.4, and 3.5 of AdvCal_2019.03.fps.

The following command is selected:

pdbSet ImplantData defect.model user.defined

Using this command, Sentaurus Process expects the updating of data fields after ion
implantation to be performed by the Alagator procedure UserPointDefectModel, which is
defined in section 3.5. The UserPointDefectModel procedure is called once after each ion
implantation1 to perform the following tasks:
■ It calls the impPostProcess_AdvCal procedure, implemented in section 3.3 of
AdvCal_2019.03.fps. This procedure updates the data fields and parameters for
initialization of defect and dopant data fields for all materials after ion implantation and
before diffusion, depending on implantation conditions. The same physical parameters are
applied to analytic and Monte Carlo implantations.
■ For Monte Carlo implantation, it calls additional procedures. First, it calls the procedure
ctrimDamageModels_AdvCal in the case of Crystal-TRIM implantation, which is
implemented in section 3.4 of AdvCal_2019.03.fps. This is needed for storing the
Monte Carlo amorphization data field. Second, it calls the procedure CoImpPostProcess
if the Coimplant model is switched on. (For analytic implantations, CoImpPostProcess
is already in the regular implantation postprocessing procedures of Sentaurus Process and
does not need to be called again in UserPointDefectModel.)
■ For Monte Carlo implantation of molecules, UserPointDefectModel is called once for
each atom species. In the special case of BF2 implantation, care has been taken that
impPostProcess_AdvCal is called only once (with the argument Name=Boron, but not
with the argument Name=Fluorine). For all other molecular species with more than one
atom species, such as B18H22 or C2B10H12, the procedure impPostProcess_AdvCal is
called once for each atom species. This might cause (a) the damage field to be added
several times and (b) excess interstitials to be added for each implantation ion species: (a)
can be compensated by setting a small value of the corresponding mcdfactor procedure
for molecular species, and (b) might be wrong for molecules that include H atoms, because
H is likely to remain in interstitial positions (like F). This is performed automatically in the
case of C2B10H12 in the _AdvCal::ImpPreProcess and _AdvCal::MCPostProcess
procedures, implemented in section 3.6 of AdvCal_2019.03.fps, as explained in
Implantation Preprocessing and Postprocessing on page 65. For implantations of
molecules other than BF2 and C2B10H12, such as B18H22, it is recommended to subtract
excess interstitials generated by H as shown in Molecular Implantation on page 134.

1. For Monte Carlo implantation of the molecular species BF2, the procedure UserPointDefectModel is
actually called once for each element (boron and fluorine). This special case is taken into account in the
implementation of UserPointDefectModel.

56 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

All data-field processing for dopants and point defects is performed in the Alagator procedure
named impPostProcess_AdvCal.

impPostProcess_AdvCal
See section 3.3 of AdvCal_2019.03.fps.

Scaling Factors for Point Defects and Damage

See section 3.3.1 of AdvCal_2019.03.fps.

Point Defects

The amount of point defects generated by ion implantation is calculated with the Advanced
Calibration ‘+x’ model. The point defects are located at the same position as the as-implanted
ions. The interstitial concentration originating from ion implantation is the as-implanted
dopant profile multiplied by ifactor. The vacancy concentration is the as-implanted dopant
profile multiplied by vfactor.

Unless specified directly in the implant command, ifactor and vfactor are calculated in
the following way. If a procedure ifactor_${mat}_${Species} (for example,
ifactor_Germanium_Arsenic) is defined for the current implantation species into a
specific material, ifactor for this specific material is calculated in this procedure as a
function of the implantation energy and dose. Otherwise, if a procedure ifactor_$Species
(for example, ifactor_Arsenic) is defined for the current implantation species (for any
material), ifactor is calculated in this procedure as a function of the implantation energy and
dose. Otherwise, ifactor has the value 1. Similarly, if a procedure vfactor_$Species is
defined for the current implantation species (for a specific material or any material), vfactor
(for this specific material or any material) is calculated in this procedure as a function of energy
and dose. Otherwise, vfactor has the value 0.

In section 3.2 of AdvCal_2019.03.fps, procedures are defined for the calculation of

ifactor after implantation of As, B, BF2, C, C2B10H12, Ge, In, P, or Si ions. vfactor equals
0 for all implantation species. In Advanced Calibration, ifactor is calibrated for pure Si but
is not defined to be silicon specific. Therefore, in the case of other materials, the same
ifactor as for Si is used, unless a material-specific ifactor is defined, for example, for
ifactor_Germanium_Germanium and ifactor_Germanium_Silicon.

The concept behind this type of implementation is that it is very convenient for users to fine-
tune ifactor or vfactor for any species as a function of implantation energy and dose,
according to their needs.

Advanced Calibration for Process Simulation User Guide 57

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

For this purpose, it is sufficient to (re-)define a very short procedure ifactor_$Species or

vfactor_$Species. This can be performed ideally after sourcing the Advanced Calibration
file.

The return value of ifactor_$Species and ifactor_${mat}_${Species} is usually a

number, but it is also possible to define an expression as the return value.

For example, after sourcing AdvCal_2019.03.fps, you can define the following procedure
to create one excess interstitial per B atom in pure Si, and a slightly reduced number of
interstitials in regions that have a high concentration of Ge before the B implantation:
fproc ifactor_Boron {Energy Dose} {
return "(1e22/(Germanium+1e22))"
}

You can specify ifactor directly in the implant command. In that case, the directly
specified values are used instead of the return values of ifactor_$Species or
ifactor_${mat}_${Species}.

With Advanced Calibration loaded, you cannot define ifactor or vfactor with pdbSet
commands such as:
pdbSet <material> <dopant> IFactor <value>

The value of the pdb parameter IFactor is ignored, and ifactor is calculated as described
in the text.

Similarly, the parameter database entries for VFactor, MCIFactor, MCVFactor, and
DFactor are ignored when Advanced Calibration is used.

Damage

The crystal damage is proportional to the concentration of displaced atoms in silicon, including
Frenkel pairs. It is scaled by dfactor. The damage field is used by Sentaurus Process to
determine whether a region is amorphous or crystalline at the beginning of the annealing.

By default, the damage model (AmModel) is Damage. In this case, the damage field is
calculated during ion implantation, using the damage accumulation model of Sentaurus MC or
Crystal-TRIM for Monte Carlo ion implantation or analytic damage profiles suggested by
Hobler and Selberherr [119] for analytic ion implantation. The damage field is scaled by
dfactor. For analytic implantation, dfactor can depend on implantation energy and dose if
a procedure dfactor_$Species or dfactor_${mat}_${Species} is defined for the
implantation species.

Similarly, for Monte Carlo implantation, dfactor can depend on implantation energy and
dose if a procedure mcdfactor_$Species or mcdfactor_${mat}_${Species} is

58 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

defined. These procedures are defined in section 3.2.2 of AdvCal_2019.03.fps. For most
species, the procedure mcdfactor_$Species returns a constant value. The calibration of
damage scaling for Monte Carlo implantation has been performed for Sentaurus MC.

To change the damage scaling for Monte Carlo implantation, you have the following options:
■ Redefine the mcdfactor_$Species or mcdfactor_${mat}_${Species} procedure.
This is the best method.
■ Define the parameter MCDFactor for a material and species. By default, MCDFactor is
undefined for all materials and all species. If you define it, MCDFactor will be used instead
of the return value of the mcdfactor procedures defined in the Advanced Calibration file.
■ Specify the mc.dfactor argument directly in the implant command. In that case, for
this particular implantation, the specified value will be used.

You can use a ‘+x’ damage model by defining a procedure AmModel_$Species or

AmModel_${mat}_${Species}, which gives the return value ‘+1’. In this case, the crystal
damage is the product of the as-implanted dopant profile and dfactor. The ‘+x’ damage
model can be helpful to describe situations where buried amorphous layers are formed after
medium-dose implantation of heavy ions (for example, indium implantation with 100 keV,
13 –2
6 × 10 cm ). With the ‘+x’ damage model, it is necessary to define very high values for
dfactor to adjust the measured and simulated amorphization.

Values for Initial Dopant Activation

See section 3.3.2 of AdvCal_2019.03.fps.

AcInit and AmInit determine the initial activation of dopants after implantation. For regions
with an as-implanted dopant concentration higher than AcInit (AmInit), the surplus dopants
are assumed to be in clusters at the beginning of a diffusion process. Both AcInit and AmInit
are applied in subroutines of the callback procedure diffPreProcess, which is called before
each diffusion simulation. AcInit is used for regions that are not amorphized by ion
implantation, and AmInit is used for regions that are amorphized by ion implantation and
recrystallize at the beginning of the diffusion process.

For most dopants, AcInit and AmInit are defined in part 2 of the calibration file as constant
parameters. Only AcInit for B and BF2 is defined in a similar way to ifactor.

You can define AcInit and AmInit in a similar way to ifactor. If a procedure
acinit_${mat}_${Species} (for example, acinit_Germanium_Arsenic) is defined
for a specific material, AcInit for this specific material is calculated in this procedure as a
function of the implantation energy and dose. Otherwise, if a procedure acinit_$Species
(for example, acinit_Arsenic) is defined (for any material), AcInit is calculated in this
procedure as a function of the implantation energy and dose. The same applies to AmInit.

Advanced Calibration for Process Simulation User Guide 59

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

By using expressions as the return value of the function acinit_$Species,

aminit_$Species, it is possible to define AcInit, AmInit as terms.

Sum of As-Implanted Point Defects and Crystal Damage

See sections 3.3.3 of AdvCal_2019.03.fps.

The data fields Int_LastImp and Vac_LastImp, which are generated during implantation,
are added to the fields Int_Implant and Vac_Implant, respectively, which are used in the
callback procedure diffPreProcess for generating the initial conditions for diffusion. This
enables a correct treatment of several subsequent implantations of the same dopant.

Similarly, the crystal damage from individual implantation steps is added to the field Damage,
which is used during diffusion preprocessing to determine amorphous regions. In Advanced
Calibration, AmModel is Damage for all species by default. Therefore, the damage field
Damage_LastImp is added to Damage. Damage_LastImp has been calculated during ion
implantation using either the damage accumulation model of the Monte Carlo ion implantation
or analytic damage formulas suggested by Hobler and Selberherr [119] for analytic ion
implantation.

Subroutines for Setting ifactor and dfactor

See section 3.2 of AdvCal_2019.03.fps.

In the Advanced Calibration, ifactor, vfactor, dfactor, AcInit, and AmInit can
depend on the material, species, energy, and dose of the ion implantation. All values and
formulas in section 3.2 have been calibrated by Synopsys, using mainly SIMS data. The
guidelines used for the calibration are explained here.

ifactor

Giles [120] observed that using an ifactor of 1.0 gives good results for many situations. An
ifactor of 1.0 means that, after the fast recombination of excess interstitials and vacancies
generated by ion implantation, one interstitial atom survives per implanted ion. In
AdvCal_2019.03.fps, an ifactor of 1.0 is used for the implantation of most species.

14 –2
For high dose (> 1 × 10 cm ) arsenic implantation (into silicon), a higher value of ifactor
is used. In this case, ifactor is set in the procedure ifactor_Arsenic as a steady function
of implantation energy and dose, and can have a value up to 9.0 for high-energy and high-dose
implantation.

60 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

The higher value of ifactor will reflect a consequence of amorphization of the silicon surface
layer by high-dose arsenic implantation. During solid phase epitaxial regrowth (SPER) of the
amorphized surface region, all point defects in that region are eliminated.

Since the interstitials generated by collisions during ion implantation are (on average) located
slightly deeper than the vacancies, a larger fraction of interstitials will be located in the deeper
region inside silicon, which is not amorphized, and will survive the damage annealing by
SPER. After recombination of all of the vacancies and interstitials that survived SPER, there
remains an average number of interstitials per implanted ion higher than 1.0 in the
nonamorphized region. This consideration can be investigated quantitatively by using the full
cascade mode of Monte Carlo ion implantation simulation, as illustrated in Figure 5.
Damage (Amorphization Threshold: 1.15e22)
Interstitials, as-implanted
1022 Vacancies, as-implanted
+x interstitials, ifactor=6.8
Net interstitials in nonamorphized region
1021 Arsenic, as-implanted
Concentration [cm–3]

1020

1019

1018

0 0.02 0.04 0.06

Depth [µm]
15 –2
Figure 5 Derivation of ifactor for As implantation (dose 1 ×10 cm , energy 8 keV) by full-
cascade MC simulation. The upper black solid line and the line with triangles
show the interstitial and vacancy concentration, respectively, as calculated by full-
cascade MC simulation. Near the surface (depth < 0.008 μm), the concentration
of vacancies is higher; in other regions, the concentration of interstitials is higher
than the concentration of vacancies. After SPER, only point defects in the
nonamorphized region (depth > 0.022 μm) survive. The net interstitial
concentration (Int – Vac; thick black line), calculated from the full-cascade MC
results, is similar to the As concentration multiplied by ifactor close to 10. The
dashed line shows the As concentration in the crystalline region, multiplied by
6.8, which is close to the ifactor for this process condition calculated by
AdvCal_2016.03.fps.

ifactor is set to the value 0.5 for BF2 implantation into silicon. This is meant to anticipate
the capability of F to eliminate interstitials, which is not taken into account in the diffusion
models used by Advanced Calibration by default.

Advanced Calibration for Process Simulation User Guide 61

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

If the Boolean parameter AdvCal_IFactorsVFactorsZero is set to 1, the creation of excess

interstitials and vacancies is suppressed. This parameter allows you to suppress transient-
enhanced diffusion (TED) after ion implantation. It can be used to speed up process simulation
for process parts, where TED can be neglected.

vfactor

vfactor is set to 0.0 for all implanted ions.

dfactor

The damage scaling factors for analytic and Sentaurus MC implantation into silicon and
germanium have been calibrated based on a wide range of experimental data, mainly TEM and
RBS data from the literature. They differ from 1 to compensate for the default underestimation
or overestimation of amorphization caused by implantation of these species.

For analytic implantation into silicon, the damage factor is decreased for B, but it is increased
for Si. For As, BF2, Ge, In, and P implantation, dfactor depends on the implantation energy
and dose.

For MC implantation with Sentaurus MC, the damage is scaled by the parameter MCDFactor,
which can be set individually for each implantation species. The values of MCDFactor have
been calibrated for silicon and germanium with experimental data.

Finally, to improve the accuracy for the projected ranges of Ge implantations into silicon by
Sentaurus MC, the empirical correction factor parameter for energy loss per collision
(nucl.cor) has been set to a value less than 1. In addition, the parameters Dacc and Dcrit,
which govern the damage accumulation and amorphization during Monte Carlo ion
implantation with Crystal-TRIM, have been changed for silicon to improve the reproduction of
the depth of amorphous layers after Si implantation into silicon.

Thermal Implantations
See sections 3.2, 3.6, and 3.7 of AdvCal_2019.03.fps.

Thermal implantations involve substrates kept at temperatures that differ from room
temperature. Cold or cryogenic implantations are performed below room temperature. Under
these conditions, there is less dynamic annealing and the implantation amorphizes more
rapidly, leading to deeper amorphous layers and less implantation channeling [121][122][123].
Hot or heated implantations are performed above room temperature. Under these conditions,
there is more dynamic annealing and the implantation amorphizes less rapidly, leading to
shallower amorphous layers and more implantation channeling [124][125].

62 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

Dynamic annealing is lattice self-repair during the implantation. Implantation creates

substantial damage, basically I and V, which can cluster into amorphous pockets. Like single I
and V, the I and V in amorphous pockets can recombine and, therefore, reduce the damage.
This recombination is hindered by an energy barrier, but it still occurs at relatively low
temperatures during implantation.

Different implantation dose rates and temperatures lead to different balances of damage
generation and lattice self-repair during implantation and, therefore, different amorphization
and channeling. In addition, increased channeling due to the lower vibrational amplitude of the
Si lattice at lower implantation temperatures is a much weaker effect than the reduction of
channeling due to stronger amorphization.

These effects can all be simulated by default with Sentaurus Process Kinetic Monte Carlo [126]
(see Amorphization on page 240). However, neither the implantation temperature nor the dose
rate is taken into account by default in Advanced Calibration for Sentaurus Process. To account
for the temperature dependency of implantation, section 3.2.4 contains the procedure
AdvancedThermalImplantModels. On demand, this procedure switches on a consistent
framework and calibration for thermal implantation.

To use this procedure, specify the command AdvancedThermalImplantModels

immediately after loading the Advanced Calibration file. An implantation temperature other
than room temperature must be specified in the implant command:
implant Carbon dose=1.0e+15 energy=6.0 temperature=-100

At elevated temperatures, some dopants (for example, arsenic) diffuse during implantation.
Since diffusion is not modeled in continuum during implantation, you must specify a
corresponding diffuse command immediately after the implant command, with the
diffusion temperature of the implant command and with the diffusion time depending on the
–2 –2 –1
dose (for example, 1e15 cm ) and the average dose rate (for example, 5e13 cm s ):
implant Arsenic dose=1.0e+15 energy=2.0 temperature=450
diffuse temp=450 time=[expr 1.0e+15/5.0e+13]<s>

NOTE It is recommended to use the ChargedReact model for implant

temperature diffusion. In addition, to allow diffusion at the implant
temperature, the minimal diffusion temperature for which diffusion
equations are solved must be set accordingly. In Advanced Calibration,
it is 300°C. The minimal diffusion temperature is defined by:
pdbSetDouble Diffuse minT 300.0

The following sections describe the implementation and calibration of the procedure
AdvancedThermalImplantModels.

Advanced Calibration for Process Simulation User Guide 63

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

Scope of Calibration

In AdvancedThermalImplantModels, a calibration is included only for Sentaurus MC

implantation, which allows a calibration for amorphization and channeling at the same time.
Since analytic implantation profiles are used for room temperature implantations in general,
they might need adjustments for cold and hot implantations, which is beyond the scope of the
AdvancedThermalImplantModels procedure.

The calibration covers thermal implantation from cold (–100°C) to hot (500°C) into silicon,
based on data from AMAT-VSE1 [126] [127] for implanted ions included in the calibration,
which are arsenic, boron, carbon, fluorine, germanium, indium, phosphorus, and silicon.

Thermal implantation into SiGe and Ge is currently not calibrated, meaning the parameters are
independent of the implant temperature. However, the AdvancedThermalImplantModels
procedure is compatible with the SiGe parameter interpolation and the calibration included in
the SiGe_and_Stress_Effect procedure. This means that you can simulate thermal
implantation into SiGe, but it is not necessarily accurate, since the SiGe parameters are linearly
interpolated between the temperature-dependent Si parameters and the Ge parameters for room
temperature.

NOTE Since calibration of thermal implantation is based on a specific smaller

experimental dataset from AMAT-VSE, it is not necessarily identical at
the default room temperature to the default calibration for damage and
initial conditions.

Damage
For Sentaurus MC implantations, the implantation damage is adjusted in section 3.2.4.2 by
changing the dynamic annealing factor, that is, the parameter for the survival rate of the Frenkel
pairs for the noncascade damage model, by either increasing the value for cold implantations
[126] or decreasing the value for hot implantations [127]:
pdbSet Si <species> surv.rat <value>

This not only increases (or decreases) damage during implantation to influence channeling, but
also leads to more (or less) amorphization at the end of the implantation.

Temperature dependencies of the survival rates are defined the implantation callback procedure
AdvancedThermalImpPreProcess, so you can access the current implant temperature.
This callback procedure is defined exclusively in the AdvancedThermalImplantModels
procedure and is called in the procedure _AdvCal::ImpPreProcess defined in section 3.6.

1. Applied Materials - Varian Semiconductor Equipment.

64 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

The surv.rat parameter values for room temperature are also defined outside the callback
procedure. In addition, the temperature-independent damage scaling factors mcdfactor are
adjusted for phosphorus, carbon, and silicon in section 3.2.4.1. The mcdfactor for other
implanted species are aligned with standard Advanced Calibration values.

Initial Conditions

In section 3.2.4.4, the interstitial scaling factors ifactor are defined to be temperature
dependent for arsenic and phosphorus. While the initial activation for recrystallized regions
(aminit) is assumed to be independent of the implant temperature, the initial activation for
crystalline regions (acinit) is defined to be temperature dependent for arsenic, phosphorus,
and boron.

At room temperature, both ifactor and acinit values defined in the callback procedure
AdvancedThermalImpPreProcess are aligned with standard Advanced Calibration values
and with the values from SiGe parameter interpolation included in the procedure
SiGe_and_Stress_Effect.

If you use the ChargedReact model, which is recommended for thermal implantations, the
initial distribution of arsenic and boron in substitutional and interstitial configurations is
defined in section 3.2.4.3 in the callback procedure AdvancedThermalDiffPreProcess.
This procedure is defined exclusively in AdvancedThermalImplantModels and is called in
the _AdvCal::DiffPreProcess procedure defined in section 3.7.

Implantation Preprocessing and Postprocessing

See section 3.6 of AdvCal_2019.03.fps.

Analytic Implantation
12 –2 12 –2
For implantations with dose < 10 cm , ChanDoseScaling=1. The dose of 10 cm
corresponds to the smallest entries in the double-Pearson Default implantation tables. With
ChanDoseScaling=1, for doses smaller than the smallest entries in the table, the channeling
dose scales linearly with the total dose.
12 –2
For dose > 10 cm , ChanDoseScaling is irrelevant for single implantations. For several
subsequent implantations, ChanDoseScaling is sometimes important because the
CoImplant model is used in Advanced Calibration. In the CoImplant model, the damage
from previous implantations is converted into an equivalent dose, as described in the
Sentaurus™ Process User Guide. If this equivalent dose is smaller than the smallest dose in
the table used, the calculation of the channeling dose depends on ChanDoseScaling.

Advanced Calibration for Process Simulation User Guide 65

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

12 –2
For dose > 10 cm , ChanDoseScaling=1 for backward compatibility of results obtained
with the CoImplant model.

Sentaurus Process offers an improved implantation table boron_in_silicon_2012 for

boron implantations with energy > 200 keV. To automatically select this table for
energy > 200 keV, users of Advanced Calibration can set the parameter:
pdbSetBoolean ImplantData B UseHighEnergyBoronTable 1

With UseHighEnergyBoronTable set to 1, the new table is switched on for boron

implantations with energy > 200 keV in the procedure _AdvCal::ImpPreProcess, which is
called at the beginning of executing an implant command. In the procedure
_AdvCal::ImpPostProcess, which is called at the end of executing an implant command,
Sentaurus Process switches back to the Default table for boron implantations, which is
usually preferred for low-energy boron implantations.

The table boron_in_silicon_2012 uses the Taurus table format (and, additionally, for low
energy (energy < 85 keV), it is identical to the Taurus table for boron implantations). In the
Taurus format tables, the CoImplant model is not applied. With the improved table, to take
into account the reduction of high-energy boron ion channeling due to preceding implantations,
you can use the PAI model. To do this, specify the keyword pai in the corresponding implant
command.

Monte Carlo Implantation (General)

The procedures _AdvCal::ImpPreProcess and _AdvCal::MCPostProcess are used for

the energy- and dose-dependent calibration of implantation profiles and amorphization by
Sentaurus MC.

The parameters for B, BF2, H, He, and P in silicon have been adjusted to improve the accuracy
of as-implanted profiles. A more accurate peak position of H, He, and P implantation profiles
is achieved by the calibration of the electronic stopping correction factor, LSS.pre. The
scr.par parameter, which is defined differently from the default for H and He implantation,
has been calibrated to reproduce the width of the peak of as-implanted profiles and the
channeling tail. The calibration for H covers the energy range 5–1500 keV. The calibration for
He was tested against SIMS data from 1000 keV and 2000 keV He implantations. For high-
energy implantation in the channeling direction (tilt=0), the channeling of H and He atoms
is very sensitive to the tilt angle, and the calibration of scr.par is less reliable. The calibration
of the parameters d.sim and nloc.pre helps to reproduce the channeling tail of low-energy
B implantations. These parameters are reset to their default values in the procedure
_AdvCal::MCPostProcess.

For As, BF2, and Ge implantation into silicon, the damage scaling defined in the callback
procedure mcdfactor_$Species depends on the implantation energy and dose. In particular,
for accurate amorphization by high-dose As implantations, a high damage scaling factor is

66 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 3: Initial Conditions After Ion Implantation

required. However, the dose dependency of the damage scaling factor has been introduced to
reduce amorphization by low- and medium-dose As implantations, such as channel or pocket
implantations in CMOS technology.

Carborane

It is recommended to use Sentaurus MC for molecular implantation (see Molecular

Implantation on page 134). In the case of Sentaurus MC implantation of carborane into silicon,
the following settings defined in the procedures _AdvCal::ImpPreProcess and
_AdvCal::MCPostProcess help to obtain accurate results for the amorphization and good
initial conditions:
1. The procedure _AdvCal::ImpPreProcess is called before the implantation. It adjusts
amorphization by carborane implantation. With $Energy as the energy of the carborane
molecule, the following parameter values are set before the carborane implantation:
pdbSet Si B surv.rat [expr (25.0/$Energy+0.25)]
pdbSet Si C surv.rat [expr (25.0/$Energy+0.25)]
pdbSetDouble Si d.sim 0.33
pdbSetDouble Si B nloc.pre 0.37
where surv.rat (‘survival rate’) scales the crystal damage, which is generated by nuclear
collisions during the implantation. In particular for small energies, it must be set to very
high values; otherwise, ion channeling during implantation is overestimated. In addition,
the calibration of the parameters d.sim and nloc.pre helps to further improve the
accuracy for the channeling tail.
2. The procedure _AdvCal::MCPostProcess is called after the MC implantation. In the
case of carborane implantation, it resets surv.rat for B and C to the default values, as
there might be other B or C ion implantations later in the process simulation:
pdbSet Si B surv.rat 0.225 ; # Sentaurus Process default
pdbSet Si C surv.rat 0.45 ; # Sentaurus Process default
3. The procedure _AdvCal::MCPostProcess also switches off the hydrogen solution to
save CPU time in subsequent anneals:
solution name=Hydrogen nosolve store
4. The procedure _AdvCal::MCPostProcess also sets the number of interstitials to be
generated by carborane implantation. Otherwise, Sentaurus Process adds ‘automatically’
one interstitial for each C, B, and H atom. This is not reasonable for H, which probably
comes to rest in an interstitial position and is unlikely to kick out Si atoms from the lattice
site. In addition, one ‘free’ interstitial per C atom might be an overestimation, because 2/3
interstitials per C are already included in C3I2 clusters, which the C atoms are assumed to
form immediately after implantation in the nonamorphized regions.

Advanced Calibration for Process Simulation User Guide 67

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Therefore, a correction of the ‘automatic’ calculation of excess interstitials is implemented:

set ifactor [ifactor_C2B10H12 $Energy $Dose]
sel z = "Int_Implant - $ifactor * Hydrogen_LastImp - 0.667*Carbon_LastImp"
Silicon name=Int_Implant store

The amorphization of silicon by carborane implantation has been calibrated with TEM data
provided by AMAT-VSE. The result of this calibration is summarized in Table 7.

Table 7 lists the thickness of an amorphized silicon layer after carborane implantation at room
temperature. The TEM measurement is compared to Sentaurus Process simulations with
default parameters and after calibration of surv.rat and the damage scaling factor. The
15 –2
process conditions correspond to a B dose of 10 cm and to B energies of 200 eV, 500 eV,
and 1.5 keV. All values are given in Å.

Table 7 Thickness of amorphized silicon layer after carborane implantation at room

temperature
Process TEM Sentaurus MC with AdvCal_2019.03.fps, Sentaurus MC with AdvCal_2019.03.fps,
condition no additional calibration surv.rat and damage scaling factor
calibrated

2.67 keV 1e14 20–25 0 20

6.67 keV 1e14 35–40 20 41

20.0 keV 1e14 70–75 70 74

Part 4: Comprehensive and Slow Models

The fourth part of AdvCal_2019.03.fps contains the procedure AdvancedModels. This
procedure switches on a consistent calibration of some advanced models for silicon, which are
not used by default. To use this procedure, apply the command AdvancedModels
immediately after loading the Advanced Calibration file. This switches on the following
models:
■ The Full model for interstitial clusters. In this model, the kinetics of formation and
dissolution of small interstitial clusters, {311} defects, and dislocation loops is described
by seven equations [22].
■ The ChargedCluster model for boron–interstitial clusters (BICs) and indium clusters.
Boron-clustering kinetics is described by the formation and dissolution of six types of BIC:
B2, B2I, B2I2, B3I, B3I2, and B3I3. Indium clustering is described by three clusters: In2, In2I,
and In2V.
■ The ChargedCluster model for fluorine clusters. This model allows you to simulate the
F effect on boron diffusion in a more accurate way compared to the model used by default.

68 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

In addition, other model parameters are adjusted for As, B, P, and In diffusion. These additional
adjustments have been calibrated with SIMS data. They are needed because the simulation of
transient-enhanced diffusion of these dopants is affected by switching on the Full model for
interstitial clusters instead of the 1Moment model, which is the default in Advanced
Calibration.

The disadvantage of switching on the advanced models is that the total number of equations to
be solved for dopant and defect clustering increases sharply, which leads to a typical increase
of CPU time for annealing by a factor of three or more, in comparison to simulations with the
Advanced Calibration standard models. Another disadvantage is that, due to the larger number
of model parameters, additional calibration is typically more difficult than with the standard
models.

In general, it is not recommended to use only part of the advanced models. In particular, for B
USJ formation, you should not use the ChargedCluster model without switching on the
Full model for interstitials, and vice versa. However, in devices where the maximum
concentration of B or As is small, you might consider selecting a simple clustering model for
this dopant to save CPU time.

The option AdvancedModels is recommended for fundamental research on process

simulation models and for applications where the standard models are not suited to predict all
trends correctly. These can include, for example, the modeling of dopant activation and defect
annealing during low-temperature processes.

The next sections describe the contents of the procedure AdvancedModels.

Interstitial Clusters
See section 4.1 of AdvCal_2019.03.fps.

The advanced interstitial clustering model, first published in [22], is switched on by:
pdbSet Si Int ClusterModel Full
pdbSet Si Int MultiClusterModel Full { 2Moment Loop }

In this model, seven equations are solved to describe the kinetics of self-interstitial clusters:
■ Three data fields (I2, I3, I4) describe small interstitial clusters (SMICs).
■ Two data fields (D311, density of {311} defects, and C311, density of interstitials bound
in {311} defects) describe the presence of {311} defects.
■ Two data fields (DLoop, density of dislocation loops, and CLoop, density of interstitials
bound in dislocation loops) describe dislocation loops.

Advanced Calibration for Process Simulation User Guide 69

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

For a complete description of the model and a comparison to experimental data, refer to the
literature [22].

The differences of this model to the default 1Moment model are illustrated in Figure 6 and
Figure 7 on page 71.

107

106
Interstitial Supersaturation

105

104

103

102
Experiment
101 Simulation: AdvancedModels
Simulation: 1Moment model
100
100 101 102 103 104 105
Time [s]
Figure 6 Time evolution of interstitial supersaturation after low dose 40 keV silicon
implantation, during annealing at 600oC. Experimental data points are taken from
[9]. For short-time anneals (< 1000 s), a high supersaturation is maintained by
dissolution of SMICs (small interstitial clusters). This is reproduced correctly with
the Full model, but not with the 1Moment model. For longer anneal times, the
interstitial supersaturation is maintained by {311} defects. In this situation, the
1Moment model gives similar results to the Full model.

The high supersaturation of interstitials in the initial phase of low-temperature annealing is

important also for USJ formation with spike annealing, since all temperature ramps start at low
temperature. Therefore, if the Full model is used for interstitial clusters, a suitable calibration
of transient-enhanced diffusion of dopants that diffuse together with interstitials (B, P, In)
requires either a dopant–interstitial clustering model or reduced diffusivities of
dopant–interstitial pairs at low temperatures.

The following command defines the initial conditions:

pdbSet Si I2 InitPercent 1.0

Excess interstitials generated by ion implantation are placed in I2 clusters.

70 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

1014
Clustered Interstitials [cm−2]

TEM: Total interstitials in clusters

Simulation: Total interstitials in clusters, Full model
Simulation: Interstitials in clusters, 1Moment model
1013 TEM: Interstitials in loops
Simulation: Interstitials in loops, Full model
TEM: Interstitials in {311} defects
Simulation: Interstitials in {311}, Full model

14
1012 100 keV silicon implant, dose 2x10 , anneal at 800°C

102 103
Time [s]
Figure 7 Time evolution of concentration of interstitials in {311} defects and dislocation
loops after amorphizing silicon implantation, during annealing at 800oC.
Experimental data points (TEM) are extracted from [128]. The Full model offers a
good description of interstitials bound to {311} defects and dislocation loops. In
contrast, the 1Moment model underestimates the stability of interstitial clusters in
situations with dislocation loops.

Boron–Interstitial Clusters
See section 4.2 of AdvCal_2019.03.fps.

Boron–interstitial clusters (BICs) should be used in combination with the Full model for
interstitial clusters. The BIC model is selected by:
pdbSet Si B ActiveModel ChargedCluster

All clusters are assumed to be electrically neutral. The most important model parameters are
the formation energies of the various clusters. These have been optimized by comparison to a
collection of SIMS data for various process conditions. The model reaction pathway is shown
in Figure 8 on page 72.

The BIC model allows you to use realistic initial conditions (AmInit, AcInit, ifactor,
dfactor) for B after ion implantation, in all situations. These are set in section 4.2.2 of
AdvCal_2019.03.fps. The scaling factors for the damage (dfactor) and for the number of
interstitials (ifactor) produced by implantation of BF2 have been extracted from calibrated
reference simulations using the full cascade Monte Carlo engine. The ChargedReact (five-
stream) diffusion model for B is switched on together with the BIC model.

Advanced Calibration for Process Simulation User Guide 71

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

B2
I,V BI

BI
B B2I B3I

BI
I,V I,V

BI
BI B2I2 B3I2

BI I,V
B3I3

Figure 8 Reaction pathway for BIC model

The BIC model describes the activation of B more accurately than the Transient cluster
model. This is most evident for thermal anneals with a low thermal budget, such as low-
temperature rapid thermal annealing (RTA) or millisecond annealing [129].

An example is shown in Figure 9, where the AdvancedModels provides superior accuracy for
RTA at 900°C (left). No substantial improvement is obtained for the spike anneal (right) for
typical CMOS technology.
AdvancedModels AdvancedModels
AdvancedCalibration default AdvancedCalibration default
1020 SIMS 1021 SIMS
Boron Concentration [cm−3]

Boron Concentration [cm−3]

1019 1020

1018 1019

1017 1018

B 20 keV, dose 5e14, 30 s at 900°C B 0.5 keV, dose 1e15, spike anneal 1050°C
1016 1017
0 100 200 300 0 20 40 60
Depth [nm] Depth [nm]

Figure 9 Comparison of simulated B profiles with AdvancedModels (red lines: Full model
for interstitial clusters, ChargedCluster and ChargedReact models for B) and
default Advanced Calibration model (blue lines: 1Moment model for interstitial
clusters, Transient model for B clusters) to SIMS data. (Left) Boron profiles after
RTA at 900oC (SIMS data from [130]). AdvancedModels is significantly more
accurate. (Right) Boron profiles after a spike anneal at 1050oC (SIMS data from
AMAT-VSE). All simulations are performed with AdvCal_2013.12.fps, either with
or without the AdvancedModels option.

72 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Arsenic Parameters in AdvancedModels

See section 4.3 of AdvCal_2019.03.fps.

The five-stream (ChargedReact) diffusion model for arsenic is switched on in

AdvancedModels. Other parameters do not change.

Phosphorus Parameters in AdvancedModels

See section 4.4 of AdvCal_2019.03.fps.

The five-stream (ChargedReact) diffusion model for phosphorus is switched on in

AdvancedModels. Other parameters do not change.

ChargedCluster Model for Indium

See section 4.5 of AdvCal_2019.03.fps.

Three types of indium cluster are taken into account: In2, In2I, and In2V. The following
reactions for the formation and dissolution of clusters are considered:
■ In–I + In<=> In2I
■ In2 + I <=> In2I
■ In2V + I <=> In2

The most stable cluster type is In2V. It is assumed that In2I and In2 clusters are formed at the
very beginning of thermal annealing after implantation.

The five-stream (ChargedReact) diffusion model for In is switched on in AdvancedModels.

The following lines help to improve the convergence of diffusion simulation with indium:
pdbSetDouble Si In Abs.Error 1e5
pdbSetDouble Si In Rel.Error 1e-2

Advanced Calibration for Process Simulation User Guide 73

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Fluorine Diffusion and Clustering

See section 4.6 of AdvCal_2019.03.fps.

The procedure AdvancedFluorineModel defines a physics-based model for fluorine

diffusion and clustering in silicon. It can be used in combination with both models for
interstitial clusters, that is, the default 1Moment model and the full model of the
AdvancedModels set.

Based on the literature [131][132][133], fluorine tends to stay in the interstitial position due to
a strong interstitial fluorine binding. In the context of the five-stream diffusion model
ChargedReact, this means that the neutral component of ChargePair is set to a relatively
high value, leading to a higher interstitial fluorine (FluorineInt) concentration than the
‘substitutional’ fluorine (Fluorine) concentration in general. Moreover, the literature
[131][132][133] indicates that interstitial fluorine prefers to decorate vacancies to form so-
called fluorine–vacancy clusters. In the framework of the ChargedCluster model in which
‘substitutional’ impurities cluster with silicon point defects, these clusters of interstitial
fluorines and vacancies (that is, F3V) result in fluorine–interstitial clusters (that is, F3I2)
through the following relation:
F3V <=> Fi3V <=> F3I3V <=> F3I2 (12)

The allowed cluster types are F2, F2I, and F3I2 with formation energies inherited ab initio [133].
The initial conditions of fluorine after ion implantation are mainly interstitial fluorine for
crystalline silicon and partially clustered in F2I, and F3I2 in recrystallized silicon.

The three-phase segregation model is selected as the Si–SiO2 interface model for fluorine with
parameter values allowing for strong dose loss. In addition, two F atoms located in the Si–SiO2
interface are allowed to cluster with a B atom, resulting in a fluorine-dependent boron dose
loss. In the presence of fluorine, some B atoms in the interface cluster with fluorine, thereby
freeing interface traps for single B atoms. This leads to an increased total number of boron traps
in the interface and, therefore, a stronger boron dose loss. The interface trap density, emission,
and trapping rates, and the cluster formation and dissolution rates have been calibrated by
Synopsys based on SIMS data. The Alagator implementation of the B–F cluster formation at
Si–SiO2 interfaces is part of the Advanced Calibration file.

See Fluorine Diffusion and Clustering on page 136 for additional fine-tuning of the model.

74 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Carbon Diffusion and Clustering

See section 4.7 of AdvCal_2019.03.fps.

Carbon segregation to the end-of-range defects can be modeled by the trapping of carbon by
{311} defects and loops. The corresponding calibrated model can be optionally enabled by
setting:
pdbSetBoolean Si C EORTrap 1

following the AdvancedModels command.

Nitrogen Diffusion and Clustering

See section 4.9 of AdvCal_2019.03.fps.

The procedure AdvancedNitrogenModel defines a physics-based model for nitrogen

diffusion and clustering in silicon. It can be used in combination with both models for
interstitial clusters, that is, the default 1Moment model and the full model of the
AdvancedModels set.

According to the literature [134], nitrogen behaves differently from other group V impurities
in bulk Si. While P, As, and Sb are shallow n-type dopants, substitutional nitrogen Ns is a deep-
level impurity. Therefore, the neutral five-stream diffusion model NeutralReact is selected.
Nitrogen has low solubility due to a strong interstitial nitrogen binding [134]. The stable
nitrogen–interstitial Ni pair shows strong diffusion due to a low migration barrier [134][135].
In addition, substitutional nitrogen forms stable pairs with a vacancy [136]. The reaction Ni +
V <=> Ns has been calibrated to allow for detailed balance with the reaction Ns + I <=> Ni,
meaning that the reaction rates vanish in equilibrium.

Current understanding attributes nitrogen diffusion in silicon not only to the migration of
interstitial nitrogen, but also to the migration of the nitrogen dimer N2I2 [135]. In this
calibration, the nitrogen dimer diffusion is enabled by selecting the Dimer model for nitrogen:
pdbSet Si N Dimer NDimer

and defining a diffusivity for the nitrogen dimer. The nitrogen dimer can form immobile
clusters with both interstitials and vacancies, which are stable to high temperatures [137].
These complexes then can suppress the formation of large vacancy or interstitial clusters.
Again, all reactions have been calibrated to allow for detailed balance.

The ThreePhaseSegregation model describes nitrogen dose loss. At oxide–silicon

interfaces, the parameter Scale.PairSegregation_Silicon is set to 1.0. This means that

Advanced Calibration for Process Simulation User Guide 75

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

no silicon self-interstitials are left behind when Ni pairs cross the interface from silicon to
oxide.

See Nitrogen Diffusion and Clustering on page 137 for additional fine-tuning of the N and N
dimer clustering model.

Solid Phase Epitaxial Regrowth

See section 4.10 of AdvCal_2019.03.fps.

The solid phase epitaxial regrowth (SPER) model in continuum Sentaurus Process simulates
the movement of amorphous–crystalline (a/c) interfaces due to the recrystallization of the
amorphous silicon and the dopant dynamics during such a process. The procedure
AdvancedSPERModel containing the calibration of the SPER phase field model is not enabled
by default. It can be switched on as an option and used in combination with all other models of
Advanced Calibration.

The boundary movement during SPER is described with the specific solution fields, either the
distance field by the level-set method or the phase field by the phase field method. The phase
field method is selected by:
pdbSet Diffuse SPER.Model PhaseField

It uses a consistent mesh structure, so that the phase and the other solutions are coupled
seamlessly into the Scharfetter–Gummel discretization scheme, which improves the
convergence if there is high drift due to an abrupt phase change.

With the calibration parameters of AdvancedSPERModel and a decent fine mesh (1–2 nm
spacing) in the amorphized region, the phase field method has proven to be robust in one and
two dimensions, having no convergence or oscillation problems. The mesh spacing in
amorphized regions should not exceed the phase transition width of 3 nm to allow for accurate
and stable simulation results.

The use of the phase field method is straightforward. In the case of annealing after an
amorphizing implantation, the phase field method assumes no diffusion in a crystalline region
during SPER, so that only recrystallization, diffusion in amorphous regions, and redistribution
at a/c interfaces occur.

As soon as regrowth is completed, the phase field method is disabled and the regular diffusion
initialization is applied. The dopant activation in the regrowth region is performed with the pdb
parameter AmInit or the term ${Sol}AmInit. Finally, the standard equations for crystalline
regions are solved for the remaining annealing time.

76 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Recrystallization Speed

The calibration includes parameters for the recrystallization speed for undoped amorphous
silicon (a-Si) in the (100) orientation as based on the literature [138]. The corresponding
parameters are:
pdbSet Si SPER PhaseTransWidth 0.003
pdbSet Si SPER Lambda.Fac 1.61
pdbSet Si SPER Relax.Rate { [Arr 3.08e8 2.68]/ \
[pdbGet Si SPER PhaseTransWidth]/[pdbGet Si SPER PhaseTransWidth]}

For the definitions of these parameters, refer to the Sentaurus™ Process User Guide. In
general, you do not need to change these parameters. The phase transition width
PhaseTransWidth has been chosen to be 3 nm to guarantee good convergence for regular
meshes used in process simulation, with spacing less than 3 nm in regions of the a/c interface.
Finally, the parameters Relax.Rate and Lambda.Fac have been chosen to give the correct
recrystallization speed for the given PhaseTransWidth.

In addition, retardation of recrystallization by the presence of fluorine is taken into account, but
to a smaller extent than reported in the literature [139]. The change in recrystallization speed
is defined by the factor:
pdbSet Si SPER R.Fac "(0.8+0.2*3.e18/(3.e18+Fluorine))"

Enhancement of the recrystallization speed by both n-type and p-type doping is not included.
According to Olson and Roth [138], the speed of SPER can be enhanced by a factor of
approximately 10 in a-Si with a boron concentration > 2e20 cm–3. To adjust the
recrystallization speed, change the expression of R.Fac.

In addition, the orientation-dependent parameters are set by:

pdbSet Si SPER R.Fac.Aniso { 100 20.0 110 10.0 111 1.0 }
pdbSet Si SPER E.Aniso { 100 0.0 110 0.0 111 0.0 }

This means that the recrystallization in the (100) orientation is twice as fast as in the (110)
orientation, and 20 times as fast as in the (111) direction, following [140]. In the case of
recrystallization of an amorphized pocket on a (100) wafer, recrystallization towards the
surface is faster than in the lateral direction (see Figure 10 on page 78). In the model, the
reduced speed of SPER in the (110) and (111) directions leads to an increased snow plow
effect. Whether dopants are really pushed more for SPER in the (110) direction than for SPER
in the (100) direction has not been investigated yet in experiments. For the very slow regrowth
in the (111) direction, the calibration of doping redistribution during SPER has not been tested
yet.

Advanced Calibration for Process Simulation User Guide 77

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Figure 10 Orientation-dependent SPER in 2D for an amorphized pocket on a (100) Si wafer:

(upper left) SPER phase field shown in as-implanted Si, in partially recrystallized
Si, and (lower right) disabled in fully recrystallized Si. In this example, regrowth is
faster in the vertical (100) direction than in the lateral (110) direction.

The physical mechanism behind the anisotropy of SPER velocity is that recrystallization is
slowest in (111) planes, and the recrystallization front propagates along (111) nanofacets in the
(100) direction and along (111) nanoridges in the (110) direction, as explained in [141]. This
mechanism and its consequences on SPER near material boundaries can be modeled accurately
with lattice KMC simulations [140], but not yet with continuum process simulations.

In 2D and 3D continuum process simulations, after switching on AdvancedSPERModel, you

can enforce equal SPER regrowth rate in all directions by inserting the line:
pdbSet Si SPER R.Fac.Aniso { 100 20.0 110 20.0 111 20.0 }

This might be considered the safest option when using AdvancedSPERModel because dopant
redistribution during SPER in the (110) and (111) directions will not exceed the dopant
redistribution in the (100) direction, which has been calibrated against SIMS data.

78 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Doping Redistribution

A general calibration is provided for the doping or impurity redistribution of B, As, P, In, F, and
C. For solutions to be redistributed during SPER, the diffusivity in a-Si (DAmor) and the
segregation energy (SPER.Energy) must be defined. Nonzero diffusivities allow for diffusion
in a-Si, while nonzero segregation energies allow for the snow plow effect by a semipermeable
a/c interface.

Boron

According to Venezia et al. [142], no boron is swept by the recrystallization front. The
diffusivity in a-Si, which was reported to be high at the very beginning of SPER and then to
become smaller during SPER, as an effect of increasing order inside a-Si [143], was set to
[Arr 1.0 2.68].

Clustering of B in a-Si is not taken into account in continuum Sentaurus Process. Note that the
diffusivity of boron in a-Si has the same Arrhenius energy (2.68 eV) as the velocity of the a/c
interface during SPER.

Arsenic

Following Venezia et al. [142] and Suzuki et al. [144], arsenic is swept by the recrystallization
front (see Figure 11 on page 80). Note that the diffusivity of As in a-Si has the same Arrhenius
energy (2.68 eV) as the velocity of the a/c interface during SPER. With this choice, the snow
plow effect for As is independent of the temperature at which SPER occurs.

A very small diffusivity in a-Si has been set for As, following Duffy et al. [145] reporting little
significant As diffusion at 600°C .

Advanced Calibration for Process Simulation User Guide 79

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

annealed (simulation)
1021 annealed (SIMS)
as-implanted (simulation)
as-implanted (SIMS)
Concentration [cm−3]

1020

1019

1018
0 10 20 30
Depth [nm]
Figure 11 Simulation of recrystallization with arsenic redistribution during SPER for Ge
PAI + As 2 keV 1015 cm–2 implantation and anneal at 700oC for 2 hours [142]. The
simulation results have been achieved with Advanced Calibration, with
AdvancedModels and AdvancedSPERModel switched on, and with ifactor for Ge
implantation set to 3.0.

Phosphorus

Phosphorus is swept by the recrystallization front, and a small diffusivity in a-Si has been set
for P, following Duffy et al. [145], reporting little significant P diffusion at 600°C .

Indium

Following Duffy et al. [146], indium is swept by the recrystallization front. The diffusivity in
a-Si has been calibrated to [Arr 1.0 2.57]. The Arrhenius energy of the diffusivity
(2.57 eV) is slightly smaller than the Arrhenius energy of the recrystallization velocity
(2.68 eV). As a consequence, the snow plow effect for indium depends on temperature. The
lower the temperature of SPER, the more indium is pushed towards the surface (see Figure 12
on page 81).

80 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

800°C, 1 s
19 600°C, 2 minutes
10
550°C, 10 minutes
as-implanted
Indium Concentration [cm−3]

SIMS

1018

1017

1016
0 50 100
Depth [nm]

Figure 12 Indium redistribution towards the surface during SPER at 550oC, 600oC, and
800oC. The snow plow effect is strongest if SPER is completed at the lowest
temperature. SIMS data is taken from [146].

Carbon

Carbon is swept by the recrystallization front, and the diffusivity has been calibrated to agree
with SIMS.

Fluorine

Following the literature [139][147], fluorine is swept by the recrystallization front. For the
diffusivity of F in a-Si, an Arrhenius energy of 2.6 eV has been selected, which is at the upper
end of possible values reported by Nash et al. [148].

Germanium

Germanium is assumed to diffuse in amorphized regions. For deep amorphization of Si/SiGe

layers, the diffusion of Ge during SPER can result in significant broadening of the Ge profile.

Advanced Calibration for Process Simulation User Guide 81

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Melting Laser Anneal

See section 4.11 of AdvCal_2019.03.fps.

Melting laser anneal (MLA) uses an extremely short (submicrosecond) annealing time and
achieves very high (almost 100%) levels of dopant activation and a low defect concentration in
recrystallized molten regions. Therefore, MLA is ideal for backside processing without
affecting the already existing front side of a thinned structure as in modern discrete power
IGBTs [149][150] or backside imagers [151].

A phase field model for MLA has been implemented based on [152]. The phase field variable
HeatPhase ( ϕ ) describes whether the material is liquid ( ϕ =0) or solid ( ϕ =1). Since the
melting or solidification process occurs too quickly to observe the dopant diffusion in a solid
region, it is assumed that the dopant atoms diffuse mostly in the liquid phase. The different
chemical potentials of dopants at liquid, solid, and boundary regions induce segregation effects
that are taken into account by the model. During MLA, the temperature varies greatly
depending on location. Therefore, the dopant diffusion equation must be solved by coupling it
to the heat equation and the phase equation. The heat generation rate G in the heat equation is
calculated from the laser intensity profile. Dopant diffusion ∂C ⁄ ∂t is coupled to the local
temperature and depends on the diffusivity D , where D is a function of the phase field variable
ϕ that follows from the phase equation.

In the beginning of a laser anneal, all dopants and point defects are initialized in the standard
way using the diffPreProcess procedure according to the models selected. Then, during
melting, all point defects and clusters are dissolved in the liquid phase. The point defect and
cluster solutions are reset to zero in a molten region during diffusion, which implies that all
dopants in a liquid (and subsequently recrystallized) region are fully activated. The instant
recrystallization of an amorphous region, that is, the initialization of cluster solutions, is
performed before diffusion, but information about the implanted damage is taken into account
for calculating the degree of structural disorder α (degree of amorphization), which then
affects the thermodynamic properties of solid material. In practically all cases, molten silicon
solidifies as crystalline (single or poly, depending on the substrate), regardless of whether it
was amorphous or crystalline before melting. This is taken into account by a special equation
for α , which gradually reduces to zero upon melting.

Most parameters of the MLA model are already calibrated and are set as defaults in Sentaurus
Process. A dedicated procedure AdvancedMLAModel is part of the Advanced Calibration file
to set typical parameters necessary for MLA simulations that are not default. In addition, the
procedure contains some parameters that are already default but can be used as references for
calibration. This procedure is not activated by default and must be called before the diffuse
command corresponding to the MLA. Typically, MLA is applied at the end of a process
simulation; however, when it is followed by other anneal steps, it is necessary to restore some
critical defaults that were changed in the AdvancedMLAModel procedure. In such cases, the
procedure AdvancedMLAModelReset must be run after the MLA diffuse command.

82 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Equipment-specific settings (such as light absorptivity, laser intensity model, and laser pulse
intensity shape) and settings for the surface reflectivity are not included in
AdvancedMLAModel and, therefore, must be specified in the input file explicitly.

NOTE The IGBT Process Simulation With Backside Melt Laser Annealing and
Its Device Simulations for DC, Switching, and Short-Circuit
Characteristics project was created for TCAD Sentaurus to facilitate the
use of the MLA model. The project demonstrates the simulation of
MLA backside processing integrated in a 2D IGBT process
simulation [153].

Settings for Simulation Optimization

The seed term in the HeatPhase equation is switched on when the HeatPhase solution is
larger than the value of SeedOnPhase everywhere in the structure. Since the derivation of the
phase field model does not include material interfaces and the initialization of melting
explicitly, it does not describe the physics precisely when the solid–liquid interface reaches
region boundaries during solidification. This might lead to a large nonphysical increase in
temperature (far above the melting temperature) near the surface of the silicon at the end of the
recrystallization. Setting SeedOffPhase < SeedOnPhase < 1.0 allows you to switch on the
seed term at the end of recrystallization, thereby suppressing the strong nonphysical increase
in temperature at the surface of the silicon. Therefore, SeedOnPhase is set to 0.002 in
AdvancedMLAModel. Its value can be increased if there is poor convergence of the MLA
model.

AdvancedMLAModel switches off interstitial clusters together with interstitial and vacancy
transport equations. This reduces the simulation time by approximately two times for typical
MLA simulations:
pdbSet Si Int ClusterModel None
pdbSetBoolean Defect ForcedTurnOff 1

The following settings ensure steady numeric convergence in the MLA model:
pdbSet Math NegErrCntrl 1
math fullNewton

MaxTimeStep=1.0e-9 (in seconds) provides good resolution for the beginning of the laser
pulse. The following error controls for temperature and dopants allow for the automatic
– 12
reduction of the time step to much smaller values (approximately 10 s ) at critical parts of
the MLA simulation during melting recrystallization. Without such a strong time-step
reduction, simulated profiles would be very noisy:
pdbSet Si Temperature Abs.Error 1e-8
pdbSet Si Temperature Rel.Error 1e-6
pdbSetDouble Si Temperature Transient.Rel.Error 3.0e-6

Advanced Calibration for Process Simulation User Guide 83

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

pdbSetDouble Si Dopant Abs.Error 1e-3

pdbSetDouble Si Dopant Rel.Error 1e-8
pdbSetDouble Si Dopant Transient.Rel.Error 1e-5

Meshing
–7
The default value ( 10 cm ) of the variable Heat.Phase.Width (interface thickness δ ) is
typically good for simulations with melting depths of approximately 100 nm. For larger
–7
melting depths, for example, 1 μm , a value of approximately 2 ×10 cm can reduce the
simulation time and still maintain reasonable accuracy:
pdbSet Si Heat.Phase.Width 1e-7 ;# (cm)

Mesh spacing in the molten region must always be smaller than Heat.Phase.Width to reach
proper convergence. A larger mesh spacing results in a faster simulation and less noise. Larger
values of Heat.Phase.Width can speed up convergence even for a fixed mesh spacing. The
simulated dopant distribution depends on both mesh spacing and Heat.Phase.Width. For an
equidistant 1D mesh, the simulation results are almost the same for all mesh spacing less than
Heat.Phase.Width. For an inhomogeneous 1D mesh, the simulated melting front speed
changes when the solid–liquid interface reaches the region of mesh inhomogeneity, unless the
maximum mesh spacing is smaller than Heat.Phase.Width/8. A mesh finer than
Heat.Phase.Width/8 might result in a larger CPU time. This basically means that mesh
refinements are allowed only if the background mesh is finer than Heat.Phase.Width/8.
Therefore, an equidistant mesh must be used in the melting region whenever possible.

A numeric error appears in the expression for heat generation if mesh spacing is not much
smaller than 1/absorptivity. The total integrated dose of HeatRate is a good indicator of the
presence of such an error. It can be checked in the output file. For example, for a 0.5 nm mesh
and an absorptivity=1.46e6 cm–1 the HeatRate dose is equal to 1.0004e14. The analytic total
–3
integrated HeatRate dose is equal to 1.0e14. A difference of more than 10 might cause a
pronounced increase of the melting depth. Two solutions are possible:
■ A finer mesh at the outer silicon interface. It can be limited by increasing the CPU time
related to the mesh inhomogeneity constraint previously described.
14
■ The laser fluence can be multiplied by the factor 10 /TotalHeatRateDose. This is
performed automatically when pdbSet Heat Correct.Energy.Dose 1 is applied.
This solution works well in most typical 1D cases and allows you to obtain good results at
meshes, when mesh spacing does not satisfy the criterion of mesh spacing << 1/
absorptivity.

84 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Thermodynamics of Silicon

The thermodynamic properties of bulk silicon have been taken from the literature [154][155]
[156] [157][158]:
pdbSet Si Melting.Interface.Mobility 600 ;# uc (cm4/J/sec)
pdbSet Si Amorphous.Melting.Interface.Mobility 600 ;# ua (cm4/J/sec)
pdbSet Si Surface.Tension 3.45e-5 ;# gammac (J/cm2)
pdbSet Si Amorphous.Surface.Tension 3.45e-5 ;# gammaa (J/cm2)
pdbSet Si AmorpDensity 1.15e22 ;# Dmax (cm-3)
pdbSet Si MassDensity 2.33 ;# (g/cm3)
pdbSet Si LatticeDensity 5.0e22 ;# (cm-3)
pdbSet Si Melting.Point 1687 ;# Tmc0 (K)
pdbSet Si Amorphous.Melting.Point 1420 ;# Tma0 (K)
pdbSet Si Latent.Heat 1.78e6 ;# Lc (J/kg)
pdbSet Si Amorphous.Latent.Heat 1.32e6 ;# La (J/kg)
pdbSet Si SpecificHeatCapacity \
{1176.0+Temperature*(1.3e-4*Temperature-0.252)-1.19e5/Temperature}
;# Cpc (J/kg/K)
pdbSet Si Amorphous.SpecificHeatCapacity \
{1176.0+Temperature*(1.3e-4*Temperature-0.252)-1.19e5/Temperature}
;# Cpa (J/kg/K)
pdbSet Si Liquid.SpecificHeatCapacity {968.47} ;# Cpl (J/kg/K)
pdbSet Si ThermalConductivity {0.01* \
(-73.85+Temperature*(-1.36e-5*Temperature+5.72e-2)+6.21e4/Temperature)}
;# ks (W/cm/K)
pdbSet Si Amorphous.ThermalConductivity {0.018} ;# ka (W/cm/K)

Note that, for better simulation robustness, the surface tension at the interface between
amorphous solid and liquid silicon is set to the same value as for the interface between
crystalline and liquid silicon. This is different from [155] and [158].

The temperature-dependent value of the specific heat of amorphous silicon is set to be the same
as in crystalline silicon, because they are close (depending on the exact state of amorphous
silicon) [159], and little experimental data is available.

Thermal conductivity of amorphous silicon is in agreement with the experimental value for
room temperature from [156]. Although you might expect an increase of this thermal
conductivity for larger temperatures, it was set constant due to the lack of additional data.

Thermal conductivity of liquid silicon in the parameter database of Sentaurus Process is

defined according to [160]. The MLA model does not explicitly take into account the density
change during the phase transition.

Advanced Calibration for Process Simulation User Guide 85

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

Therefore, in quasi-1D simulations, Liquid.ThermalConductivity must be corrected for

the density change, that is, the parameter given here (whose value corresponds to [160] and is
close to the Sentaurus Process default) must be multiplied by the factor 2.57/2.33. This
correction has a small influence on the simulation results:
pdbSet Si Liquid.ThermalConductivity \
{(2.57/2.33)*(0.56+3.28e-4*(Temperature-[pdbGet Silicon Melting.Point]))} \
;# kl (W/cm/K)

The Vogel–Fulcher model is recommended for solid–liquid interface velocity simulations. The
values of its parameters were set to obtain the maximum solid–liquid interface velocity of
15 m/s and 25 m/s for crystalline and amorphous silicon, respectively, as reported in [158].

Silicon Absorptivity

The recommended value of the light absorptivity in silicon for a 308 nm laser can be set by:
pdbSet Si Absorptivity 1.46e6 ;# (cm-1)

An absorptivity of 1.46e6 cm–1 corresponds to the liquid silicon and is close to the value for
crystalline silicon at room temperature [161]. It increases for large temperatures, but since the
value 1.46e6 is already sufficiently large, increasing absorptivity further does not greatly
influence the melting dynamics. Therefore, the above constant value is recommended for the
simulation of a 308 nm laser. In principle, any temperature- and phase-dependent expression
can be used for absorptivity, which then becomes time dependent. In that case, the following
flag must be switched on to account for this:
pdbSet Heat UpdateHeatRate 1 ;#default 0

The optical properties of monocrystalline silicon at room temperature for wavelengths from
234 nm to 840 nm can be found in [162].

MLA Calibration

Synopsys calibrated the MLA model in the framework of the ATEMOX project [163] to
experimental results of the Excico UV 308 nm laser with a pulse duration of 150–200 ns. The
2
laser irradiation area was approximately 1 cm (full chip) per shot. Only flat silicon wafers
with native oxide were used; therefore, the use of 1D simulation is fully justified in these cases.
No structures with surface amorphization were considered; therefore, the resulting parameters
are reliable only for MLA of crystalline silicon.

The calibration consisted of two stages:

■ Calibration of melting dynamics to fine-tune the melting depth and the melting duration
using temperature- and phase-dependent reflectivity of the wafer surface. Since reflectivity
for the main pulse wavelength (308 nm) was not measured directly, it was the main

86 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

calibration parameter for the melting depth in this model. It was entered into the parameter
Heat Intensity.Table.Factor, which scales the intensity of laser radiation
depending on the temperature and the phase of the exposed silicon surface. For the silicon
covered with only thin native oxide, it has the following form:
pdbSet Heat Intensity.Table.Factor "\[expr $fluence*(1- \
($SurfPhase*(0.575+4e-5*$SurfTempK)+(1-$SurfPhase)*0.78))\]"
where:
• $fluence is the laser energy density (in the case of Intensity.Table normalized
to 1 J/cm2).
• $SurfPhase is the current phase of the silicon near the surface (it can vary between 0
for the liquid state and 1 for the solid state).
• $SurfTempK is the silicon surface temperature.

Emphasis was placed on keeping the surface reflectivity similar between different
experiments and close to the values suggested in the calibration performed by other
ATEMOX partners earlier.
■ Calibration of dopant diffusion in the liquid phase and its segregation at the solid–liquid
interface was performed for boron and phosphorus. First, values of Melting.Seg.E were
set according to experimental equilibrium partition coefficients of dopants from [164].
Second, the dopant diffusivities and interface chemical potentials were set to obtain the
best visual fit of simulated profiles to SIMS profiles. Interface diffusivities (Dils.0) were
–9 2
set to a small value ( 10 cm /s ) to avoid the formation of an artificial notch at the
maximum melting depth position. No temperature dependency was set for the diffusivities
because available data could be fitted without it.

Figure 13 on page 88 shows the results of the calibration for boron, and Figure 14 on page 88
shows the results of the calibration for phosphorus.

The SIMS profiles were provided to Synopsys within the ATEMOX project (most profiles are
also published). The results and parameters for boron are qualitatively similar to the ones
obtained in [165]. Note that SIMS profiles are usually more smooth (and their peaks are shifted
in depth by several nanometers) than simulated dopant profiles after MLA. This is caused by
inaccuracy of the SIMS data (see the comparison of raw and convolved simulated profiles in
[165]). The calibration for phosphorus was based on data from [166].

Advanced Calibration for Process Simulation User Guide 87

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 4: Comprehensive and Slow Models

B 3 keV, 3e13 cm–2, tilt = 7, rot = 23; 200 ns laser pulse (Excico)
R = SurfPhase * (0.575+4e-5 * SurfTempK) + (1 – SurfPhase) * 0.715
SIMS Elas = 2.35 J/cm2
SIMS Elas = 2.05 J/cm2
1019 SIMS Elas = 1.95 J/cm2
Simulation as-implanted
Simulation Elas = 2.35 J/cm2
Simulation Elas = 2.05 J/cm2
Simulation Elas = 1.95 J/cm2
Boron Concentration [cm−3]

1018

1017

1016
0 100 200
Depth [nm]
Figure 13 Comparison of simulated B profiles with SIMS profiles [165] for different laser
energies (Elas); the same "200 ns" laser pulse shape was used for all simulated
Elas shown

P 200 keV, 1e14 cm−2, tilt = 7, rot = 23; 180 ns laser pulse (Excico)
R = SurfPhase * (0.536 = 4e-5 * SurfTempK) + (1 − SurfPhase) * 0.8
Phosphorus Concentration [cm−3]

1018

SIMS as-implanted
SIMS Elas = 3.55 J/cm2
1017 SIMS Elas = 3.0 J/cm2
SIMS Elas = 2.56 J/cm2
SIMS Elas = 2.11 J/cm2
Simulation Elas = 3.55 J/cm2
Simulation Elas = 3.0 J/cm2
Simulation Elas = 2.56 J/cm2
Simulation Elas = 2.11 J/cm2

0 100 200 300 400

Depth [nm]

Figure 14 Comparison of simulated P profiles with SIMS profiles [166]; the same "180 ns"
laser pulse shape was used for all simulated laser energies (Elas) shown

88 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 5: Accelerating Simulations for Power Technologies

Limitations of MLA Model

The MLA model provides good simulation results for dopant redistribution during the melting
process and dopant activation in the recrystallized region for typical quasi-1D cases. The
following limitations remain in the model:
■ MLA has not been calibrated for structures with amorphous regions.
■ Only simple diffusion equations for dopants are coupled with the heat equation. For the
remainder (such as complex clusters), only a constant temperature is supported. This
impedes the simulation and calibration of dopant activation and defect annealing in the
solid phase during MLA. Therefore, a dedicated calibration of the electrical activation of
dopants immediately below the melted surface-near region has not yet been performed, that
is, the simulated dopant activation and defect annealing in the solid phase during
submicrosecond MLA might be inaccurate.

Part 5: Accelerating Simulations for Power Technologies

Part 5 of AdvCal_2019.03.fps is designed for the needs of process simulations for various
power devices made of silicon such as LDMOS, VDMOS, IGBT, and superjunction MOSFET.

For many power devices, process simulation is time consuming because many mesh points are
needed for a proper discretization of the simulation domain and because a high number of
thermal anneals and oxidations is used in the fabrication process. In addition, for many power
devices, from the simulation perspective, the process simulation can be divided into two parts:
■ In the first part of process simulation, the concentration of dopants is small (far below solid
solubility), and the overall thermal budget for thermal annealing and oxidation is large.
■ In the second part of process simulation, typically associated with the formation of
electrical contacts, the dopant concentration can be very high (above solid solubility), and
the thermal budget for thermal annealing after implantation is small or moderate.

In these devices, usually the first part of the process simulation consists of the majority of time-
consuming process steps (in particular, oxidation steps). However, the CPU time spent for the
corresponding process simulation can be reduced significantly by taking advantage of the low
dopant concentration and high thermal budgets. To speed up the first part of the process
simulation, you can use simpler models and simpler settings than those used by the default
Advanced Calibration.

For the second part of the process simulation (typically, starting with the first high-dose dopant
implantations into silicon), it is recommended to use the full set of standard Advanced
Calibration models to obtain accurate results for dopant transient-enhanced diffusion and
activation.

Advanced Calibration for Process Simulation User Guide 89

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 5: Accelerating Simulations for Power Technologies

Part 5 of AdvCal_2019.03.fps contains the procedures AdvancedPowerDeviceMode and

AdvancedPowerDeviceModeReset, which are designed to reduce CPU time for process
simulation of power devices without sacrificing simulation accuracy. The procedure
AdvancedPowerDeviceMode switches on settings for fast process simulation of the first part
of power processes. The procedure AdvancedPowerDeviceModeReset reverts to the
standard settings of Advanced Calibration for accurate simulation of dopant diffusion and
activation in the second part of power processes.

Using AdvancedPowerDeviceMode in the first part of the process simulation and reverting
to the standard models with AdvancedPowerDeviceModeReset for the second part of the
process simulation typically result in a reduction of total process simulation CPU time by
20–40%, depending on the process flow. At the same time, the simulation results do not change
significantly when compared to simulations where Advanced Calibration standard models are
used for the completed process flow.

Using AdvancedPowerDeviceMode

Variant 1: Explicitly Reverting to Standard Models

The procedure AdvancedPowerDeviceMode is called without arguments immediately after

the AdvancedCalibration command line. The AdvancedPowerDeviceModeReset
procedure is called when the first part of the process simulation (with low doping concentration
in silicon) is completed and the second part of process simulation starts, for example, with
14 –2
high-dose (dose > 10 cm ) implantations for contact regions.

To select the models and parameters, the process simulation input file contains the following
flow of commands:
AdvancedCalibration 2019.03
AdvancedPowerDeviceMode
source ./mycalib.fps
... ;# first part of process
AdvancedPowerDeviceModeReset
... ;# second part of process

You can validate whether the acceleration or the process simulation with
AdvancedPowerDeviceMode is justified by comparing the CPU time and simulation results
to reference simulations, in which the lines AdvancedPowerDeviceMode and
AdvancedPowerDeviceModeReset are commented out.

90 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 5: Accelerating Simulations for Power Technologies

Variant 2: Automatically Reverting to Standard Models

The procedure AdvancedPowerDeviceMode is called with an argument <value>, which is

interpreted as a dose. Sentaurus Process calls the AdvancedPowerDeviceModeReset
procedure automatically at the first implantation with dose > <value> during the implantation
14
preprocessing. A typical value of the argument is 10 .

The process simulation input file contains the following flow of commands:
AdvancedCalibration 2019.03
AdvancedPowerDeviceMode 1e14
source ./mycalib.fps
... ;# first part of process
... ;# second part of process

This variant is convenient for TCAD beginners. Apart from ease of use, automatically reverting
to the standard models has no advantage. In some use cases, it does not give the best results,
and the first variant is preferred. Use cases, where the first variant is preferred, include:
■ Processes in which the first high-dose implantation is not intended to enter the silicon
region of the device. For example, if the entire device is covered by polysilicon or
photoresist, it might not be necessary to revert to the standard Advanced Calibration
models at the first high-dose implantation.
■ Processes in which the first high-dose implantation (dose > <value>) is part of a series of
implantations. Here, the procedure AdvancedPowerDeviceModeReset must be called
before the first implantation of that series, even if it is a low-dose implantation.

Contents of AdvancedPowerDeviceMode
Synopsys has investigated the simplifications contained in AdvancedPowerDeviceMode
based on a large set of various power-device simulation projects. Besides simplified physics,
the simplifications also include appropriate parameters for time-stepping and numeric solvers.

Increased Time Steps, Deposition Steps, and Temperature Steps

AdvancedPowerDeviceMode includes the lines:

# Numerics: Time steps
pdbSetDouble Diffuse InitTimeStep 1.0e-3 ;# default 1.0e-4
pdbSet Diffuse IncreaseRatio 4.0 ;# default 2.0
pdbSet Diffuse ReduceRatio 0.20 ;# default 0.25
pdbSet Math Time.Step.Function Linear ;# default Damped
# Numerics: Deposition steps
pdbSet Diffuse dThickness 0.001 ;# default 0.001;

Advanced Calibration for Process Simulation User Guide 91

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
Part 5: Accelerating Simulations for Power Technologies

;# 0.002 is faster
;# but less robust
pdbSet Diffuse dThicknessEpi 0.50 ;# default 0.1
# Numerics: Temperature steps
pdbSet Diffuse MaxGrowthStep 50.0 ;# default 4.0
pdbSet Diffuse delT 50.0 ;# default 10
pdbSet Diffuse delNT 50.0 ;# default 50
pdbSet Diffuse delTox 50.0 ;# default 50

These lines increase the time steps, the deposition steps, the oxidation time steps
(MaxGrowthStep), and the temperature steps compared to the default values of Advanced
Calibration for Sentaurus Process.

The values are a good trade-off between simulation robustness (where small time steps help to
eliminate noise from numeric errors and, in particular, to keep interfaces between oxide and
silicon smooth during oxidation) and CPU time, where large time steps are favorable. To
further speed up power-device process simulation, it is useful to test at least once for each
technology if the parameters InitTimeStep and dThickness can be increased without loss
of simulation robustness. According to Synopsys’ experience, in some power technologies,
you can further reduce process simulation CPU time by an additional 20–40% by increasing
InitTimeStep from 1.0e-3 to 2.0 and dThickness from 0.001 to 0.002, without
affecting simulation robustness or results.

Simplified Physics

For low dopant concentrations (<< solid solubility), you can save CPU time by switching off
the equations for dopant clusters. In AdvancedPowerDeviceMode, all dopants are made
electrically active by setting ActiveModel to None.

For very high thermal budgets, typical for oxidation steps used in the first part of power-device
process simulations, the TED of dopants caused by interstitials created during ion implantation
can be neglected compared to regular diffusion of dopants. In AdvancedPowerDeviceMode,
TED is suppressed by the line:
pdbSetBoolean AdvCal_IFactorsVFactorsZero 1

When AdvCal_IFactorsVFactorsZero is set to 1, the generation of interstitials or

vacancies by ion implantation is suppressed (as implemented in section 3.3.1 of
AdvCal_2019.03.fps). In the absence of TED, no equations need to be solved for interstitial
clusters. In AdvancedPowerDeviceMode, the ClusterModel for interstitials is set to None.
In addition, the parameters Abs.Error, Rel.Error, and Transient.Rel.Error for point
defects are increased with respect to their default values, to allow for larger time steps in the
diffusion solver.

92 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
References

Speedup Methods Not Included in AdvancedPowerDeviceMode

Synopsys tested other simplifications of physical models, but they were not found to be helpful
in all application cases. None of them is included in AdvancedPowerDeviceMode. Some
comments about most of the simplifications are:
■ Use the simpler ChargedFermi model for dopant diffusion instead of the ChargedPair
model.
With this choice, the simulation of oxidation-enhanced diffusion (OED) becomes very
inaccurate. Using the ChargedFermi model saves CPU time (typically ~20%) and is
suitable for selected technologies, but it is not recommended in general.
■ Use the ChargedFermi model and – in addition – switch off the point-defect equations
completely by setting the parameter ForcedTurnOff to 1.
With this choice, OED is completely suppressed. Diffusion of B and P during oxidation is
severely underestimated. This choice is not recommended.
■ Use local charge neutrality instead of the Poisson equation, by adding the line
pdbSetBoolean Silicon Potential Poisson 0.

This reduces the number of equations to be solved by one. However, it is often detrimental
to convergence. Therefore, in most applications, it increases the CPU time.
■ Switch off equations in polysilicon.
Usually, this has no impact on the CPU time.
■ Use the Segregation model instead of the ThreePhaseSegregation model for dose
loss of As, B, and P.
Usually, this has no impact on the CPU time.

The biggest potential to further accelerate accurate process simulation for power devices is to
define a good meshing strategy and parameters for (adaptive) meshing. Since a good mesh
depends on the details of the technology, meshing strategies and parameters for adaptive
meshing are not included yet in AdvancedPowerDeviceMode. A method for creating good
meshes for power-device process simulation is presented in Meshing on page 155.

References
[1] P. Pichler, Intrinsic Point Defects, Impurities, and Their Diffusion in Silicon,
Computational Microelectronics, Vienna: Springer, 2004.
[2] C. S. Rafferty et al., “Simulation of cluster evaporation and transient enhanced diffusion
in silicon,” Applied Physics Letters, vol. 68, no. 17, pp. 2395–2397, 1996.

Advanced Calibration for Process Simulation User Guide 93

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[3] IST Project 027152 ATOMICS, Advanced Front-End Technology Modeling for
Ultimate Integrated Circuits, for more information, go to https://
www.iisb.fraunhofer.de/content/dam/iisb2014/en/Documents/Research-Areas/
Simulation/final-report-atomics.pdf.
[4] A. Tsibizov, A. Terterian, and C. Zechner, “Influence of Poisson equation boundary
conditions and quantum corrections to carrier concentrations at material interfaces in
TCAD process simulation,” Physica Status Solidi C, vol. 11, no. 1, pp. 101–104, 2014.
[5] S. Strauss et al., “Analytic model for ion channeling in successive implantations in
crystalline silicon,” Materials Science and Engineering B, vol. 124–125, pp. 376–378,
December 2005.
[6] H. Bracht, N. A. Stolwijk, and H. Mehrer, “Equilibrium Concentrations of Intrinsic
Point Defects in Silicon Determined by Zinc Diffusion,” in Proceedings of the Seventh
International Symposium on Silicon Materials Science and Technology (Semiconductor
Silicon), vol. 94-10, San Francisco, CA, USA, pp. 593–602B, May 1994.
[7] H. Bracht, E. E. Haller, and R. Clark-Phelps, “Silicon Self-Diffusion in Isotope
Heterostructures,” Physical Review Letters, vol. 81, no. 2, pp. 393–396, 1998.
[8] A. Ural, P. B. Griffin, and J. D. Plummer, “Self-Diffusion in Silicon: Similarity between
the Properties of Native Point Defects,” Physical Review Letters, vol. 83, no. 17,
pp. 3454–3457, 1999.
[9] N. E. B. Cowern et al., “Energetics of Self-Interstitial Clusters in Si,” Physical Review
Letters, vol. 82, no. 22, pp. 4460–4463, 1999.
[10] B. Colombeau and N. E. B. Cowern, “Modelling of the chemical-pump effect and C
clustering,” Semiconductor Science and Technology, vol. 19, no. 12, pp. 1339–1342,
2004.
[11] P. M. Fahey, P. B. Griffin, and J. D. Plummer, “Point defects and dopant diffusion in
silicon,” Reviews of Modern Physics, vol. 61, no. 2, pp. 289–388, 1989.
[12] M. D. Giles, “Defect-Coupled Diffusion at High Concentrations,” IEEE Transactions on
Computer-Aided Design, vol. 8, no. 5, pp. 460–467, 1989.
[13] I. Bork and H. Matsumoto, “On the Determination of Boron Diffusivities and Boron
Interstitial Pair Binding Energies in Silicon,” in International Conference on Simulation
of Semiconductor Processes and Devices (SISPAD), Tokyo, Japan, pp. 91–92,
September 1996.
[14] J. Vanhellemont, P. Spiewak, and K. Sueoka, “On the solubility and diffusivity of the
intrinsic point defects in germanium,” Journal of Applied Physics, vol. 101, no. 3,
p. 036103, 2007.
[15] J. Vanhellemont et al., “Intrinsic point defect properties and engineering in silicon and
germanium Czochralski crystal growth,” in 5th International Symposium on Advanced
Science and Technology of Silicon Materials, Kona, HI, USA, November 2008.

94 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[16] J. Vanhellemont and E. Simoen, “Brother Silicon, Sister Germanium,” Journal of The
Electrochemical Society, vol. 154, no. 7, pp. H572–H583, 2007.
[17] P. A. Packan and J. D. Plummer, “Temperature and time dependence of B and P
diffusion in Si during surface oxidation,” Journal of Applied Physics, vol. 68, no. 8,
pp. 4327–4329, 1990.
[18] P. B. Griffin and J. D. Plummer, “Process Physics Determining 2-D Impurity Profiles in
VLSI Devices,” in IEDM Technical Digest, Los Angeles, CA, USA, pp. 522–525,
December 1986.
[19] E. Napolitani et al., “Silicon interstitial injection during dry oxidation of SiGe/Si
layers,” Journal of Applied Physics, vol. 97, no. 3, p. 036106, 2005.
[20] P. A. Stolk et al., “Physical mechanisms of transient enhanced dopant diffusion in ion-
implanted silicon,” Journal of Applied Physics, vol. 81, no. 9, pp. 6031–6050, 1997.
[21] H. Saleh et al., “Energy dependence of transient enhanced diffusion and defect kinetics,”
Applied Physics Letters, vol. 77, no. 1, pp. 112–114, 2000.
[22] N. Zographos, C. Zechner, and I. Avci, “Efficient TCAD Model for the Evolution of
Interstitial Clusters, {311} Defects, and Dislocation Loops in Silicon,” in MRS
Symposium Proceedings, Semiconductor Defect Engineering—Materials, Synthetic
Structures and Devices II, vol. 994, San Francisco, CA, USA, p. 0994-F10-01, April
2007.
[23] N. Zographos and A. Erlebach, “Process simulation of dopant diffusion and activation
in germanium,” Physica Status Solidi A, vol. 211, no. 1, pp. 143–146, 2014.
[24] E. Napolitani et al., “Transient enhanced diffusion of B mediated by self-interstitials in
preamorphized Ge,” Applied Physics Letters, vol. 96, no. 20, p. 201906, 2010.
[25] L. F. Giles et al., “Transient enhanced diffusion of B at low temperatures under extrinsic
conditions,” Solid-State Electronics, vol. 49, no. 4, pp. 618–627, 2005.
[26] S. Uppal et al., “Diffusion of boron in germanium at 800–900oC,” Journal of Applied
Physics, vol. 96, no. 3, pp. 1376–1380, 2004.
[27] G. Impellizzeri et al., “Role of fluorine in suppressing boron transient enhanced
diffusion in preamorphized Si,” Applied Physics Letters, vol. 84, no. 11, pp. 1862–1864,
2004.
[28] B. Colombeau et al., “Current Understanding and Modeling of B Diffusion and
Activation Anomalies in Preamorphized Ultra-Shallow Junctions,” in MRS Symposium
Proceedings, Silicon Front-End Junction Formation—Physics and Technology, vol. 810,
San Francisco, CA, USA, pp. 91–102, April 2004.
[29] A. Satta et al., “Diffusion, activation, and recrystallization of boron implanted in
preamorphized and crystalline germanium,” Applied Physics Letters, vol. 87, no.17,
p. 172109, 2005.

Advanced Calibration for Process Simulation User Guide 95

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[30] B. R. Yates et al., “Anomalous activation of shallow B+ implants in Ge,” Materials

Letters, vol. 65, no. 23–24, pp. 3540–3543, 2011.
[31] Y.-S. Oh and D. E. Ward, “A Calibrated Model for Trapping of Implanted Dopants at
Material Interface During Thermal Annealing,” in IEDM Technical Digest, San
Francisco, CA, USA, pp. 509–512, December 1998.
[32] A. N. Larsen et al., “Heavy doping effects in the diffusion of group IV and V impurities
in silicon,” Journal of Applied Physics, vol. 73, no. 2, pp. 691–698, 1993.
[33] S. T. Dunham and C. D. Wu, “Atomistic models of vacancy-mediated diffusion in
silicon,” Journal of Applied Physics, vol. 78, no. 4, pp. 2362–2366, 1995.
[34] S. Brotzmann and H. Bracht, “Intrinsic and extrinsic diffusion of phosphorus, arsenic,
and antimony in germanium,” Journal of Applied Physics, vol. 103, no. 3, p. 033508,
2008.
[35] M. Koike et al., “Diffusion and activation of n-type dopants in germanium,” Journal of
Applied Physics, vol. 104, no. 2, p. 023523, 2008.
[36] C. O. Chui et al., “Germanium n-type shallow junction activation dependences,”
Applied Physics Letters, vol. 87, no. 9, p. 091909, 2005.
[37] S. Koffel et al., “Experiments and simulation of the diffusion and activation of the n-
type dopants P, As, and Sb implanted into germanium,” Microelectronic Engineering,
vol. 88, no. 4, pp. 458–461, 2011.
[38] P. Tsouroutas, D. Tsoukalas, and H. Bracht, “Experiments and simulation on diffusion
and activation of codoped with arsenic and phosphorous germanium,” Journal of
Applied Physics, vol. 108, no. 2, p. 024903, 2010.
[39] B. Sahli et al., “Ab initio calculations of phosphorus and arsenic clustering parameters
for the improvement of process simulation models,” Materials Science and Engineering
B, vol. 154-155, pp. 193–197, December 2008.
[40] A. Chroneos et al., “Vacancy-arsenic clusters in germanium,” Applied Physics Letters,
vol. 91, no. 19, p. 192106, 2007.
[41] R. Kasnavi et al., “Characterization of arsenic dose loss at the Si/SiO2 interface,”
Journal of Applied Physics, vol. 87, no. 5, pp. 2255–2260, 2000.
[42] A. Satta et al., “Diffusion, activation, and regrowth behavior of high dose P implants in
Ge,” Applied Physics Letters, vol. 88, no. 16, p. 162118, 2006.
[43] D. P. Brunco et al., “Germanium: The Past and Possibly A Future Material for
Microelectronics,” ECS Transactions, vol. 11, no. 4, pp. 479-493, 2007.
[44] V. Mazzocchi et al., “Experimental Investigation of the Impact of Implanted
Phosphorus Dose and Anneal on Dopant Diffusion and Activation in Germanium,” in
MRS Symposium Proceedings, Doping Engineering for Front-End Processing,
vol. 1070, San Francisco, CA, USA, p. 1070-E01-08, March 2008.

96 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[45] V. Mazzocchi et al., “Diffusion and activation of Boron and Phosphorus in

preamorphized and crystalline Germanium using ultra fast spike anneal,” in 17th IEEE
International Conference on Advanced Thermal Processing of Semiconductors – RTP,
Albany, NY, USA, September 2009.
[46] P. Tsouroutas et al., “Modeling and experiments on diffusion and activation of
phosphorus in germanium,” Journal of Applied Physics, vol. 105, no. 9, p. 094910,
2009.
[47] A. Chroneos et al., “Engineering the free vacancy and active donor concentrations in
phosphorus and arsenic double donor-doped germanium,” Journal of Applied Physics,
vol. 104, no. 11, p. 113724, 2008.
[48] J. Dabrowski et al., “Mechanism of dopant segregation to SiO2/Si(001) interfaces,”
Physical Review B, vol. 65, p. 245305, May 2002.
[49] N. Ioannou et al., “Germanium substrate loss during low temperature annealing and its
influence on ion-implanted phosphorous dose loss,” Applied Physics Letters, vol. 93,
no. 10. p. 101910, 2008.
[50] V. C. Venezia et al., “Dopant redistribution effects in preamorphized silicon during low
temperature annealing,” in IEDM Technical Digest, Washington, DC, USA,
pp. 489–492, December 2003.
[51] C. Zechner, D. Matveev, and A. Erlebach, “Phase-field model for the dopant
redistribution during solid phase epitaxial regrowth of amorphized silicon,” Materials
Science and Engineering B, vol. 114–115, pp. 162–165, 2004.
[52] T. Aoyama, H. Tashiro, and K. Suzuki, “Diffusion of Boron, Phosphorus, Arsenic, and
Antimony in Thermally Grown Silicon Dioxide,” Journal of the Electrochemical
Society, vol. 146, no. 5, pp. 1879–1883, 1999.
[53] H. Rücker et al., “Suppressed diffusion of boron and carbon in carbon-rich silicon,”
Applied Physics Letters, vol. 73, no. 12, pp. 1682–1684, 1998.
[54] H. Rücker et al., “Erratum: Suppressed diffusion of boron and carbon in carbon-rich
silicon,” Applied Physics Letters, vol. 75, no. 1, p. 147, 1999.
[55] P. Lavéant et al., “Engineering the diffusion behavior of dopants (B, Sb) in silicon by
incorporation of carbon,” Nuclear Instruments and Methods in Physics Research B,
vol. 186, no. 1–4, pp. 292–297, 2002.
[56] V. Moroz et al., “Optimizing boron junctions through point defect and stress
engineering using carbon and germanium co-implants,” Applied Physics Letters,
vol. 87, p. 051908, August 2005.
[57] B. J. Pawlak et al., “Suppression of phosphorus diffusion by carbon co-implantation,”
Applied Physics Letters, vol. 89, p. 062102, August 2006.

Advanced Calibration for Process Simulation User Guide 97

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[58] C. Zechner et al., “Modeling Ultra Shallow Junctions Formed by Phosphorus-Carbon

and Boron-Carbon Co-implantation,” in MRS Symposium Proceedings, Semiconductor
Defect Engineering—Materials, Synthetic Structures and Devices II, vol. 994, San
Francisco, CA, USA, p. 0994-F11-17, April 2007.
[59] S. Brotzmann et al., “Diffusion and defect reactions between donors, C, and vacancies
in Ge. I. Experimental results,” Physical Review B, vol. 77, no. 23, p. 235207, 2008.
[60] A. Chroneos, B. P. Uberuaga, and R. W. Grimes, “Carbon, dopant, and vacancy
interactions in germanium,” Journal of Applied Physics, vol. 102, no. 8, p. 083707,
2007.
[61] A. Chroneos et al., “Diffusion and defect reactions between donors, C, and vacancies in
Ge. II. Atomistic calculations of related complexes,” Physical Review B, vol. 77, no. 23,
p. 235208, 2008.
[62] F. J. Morin and J. P. Maita, “Electrical Properties of Silicon Containing Arsenic and
Boron,” Physical Review, vol. 96, no. 1, pp. 28–35, 1954.
[63] B. E. Deal and A. S. Grove, “General Relationship for the Thermal Oxidation of
Silicon,” Journal of Applied Physics, vol. 36, no. 12, pp. 3770–3778, 1965.
[64] A. Ohta et al., “Dry Oxidation of Germanium (100) and (111) Surfaces - Impact of
Oxidation Temperature on Ge Oxide Growth,” in International Conference on Solid
State Devices and Materials (SSDM), Kyoto, Japan, pp. 743–744, September 2012.
[65] N. Zographos et al., “Process Modeling of Chemical and Stress Effects in SiGe,” in 19th
International Conference on Ion Implantation Technology (IIT), Valladolid, Spain,
pp. 212–216, June 2012.
[66] M. Ogino, Y. Oana, and M. Watanabe, “The Diffusion Coefficient of Germanium in
Silicon,” Physica Status Solidi A, vol. 72, no. 2, pp. 535–541, 1982.
[67] C. Ahn and S. T. Dunham, “Calculation of dopant segregation ratios at semiconductor
interfaces,” Physical Review B, vol. 78, no. 19, p. 195303, 2008.
[68] N. E. Cowern, “Diffusion in a Single Crystal within a Stressed Environment,” Physical
Review Letters, vol. 99, p. 155903, October 2007.
[69] N. E. B. Cowern et al., “Overlayer stress effects on defect formation in Si and Ge,” Thin
Solid Films, vol. 518, no. 9, pp. 2442–2447, 2010.
[70] Y. Zaitsu et al., “Boron Diffusion in Compressively Stressed Float Zone-Silicon
Induced by Si3N4 Films,” Journal of the Electrochemical Society, vol. 145, no. 1,
pp. 258–264, 1998.
[71] R. Braunstein, A. R. Moore, and F. Herman, “Intrinsic Optical Absorption in
Germanium-Silicon Alloys,” Physical Review, vol. 109, no. 3, pp. 695–710, 1958.
[72] Ioffe Physical Technical Institute, parameter database, for more information, go to
http://www.ioffe.ru/SVA/NSM/Semicond/SiGe/bandstr.html.

98 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[73] G. M. Dalpian et al., “Ab initio calculations of vacancies in SixGe1–x,” Applied Physics
Letters, vol. 81, no. 18, pp. 3383–3385, 2002.
[74] P. Venezuela et al., “Vacancy-mediated diffusion in disordered alloys: Ge self-diffusion
in Si1–xGex,” Physical Review B, vol. 65, no. 19, p. 193306, 2002.
[75] P. Ramanarayanan, K. Cho, and B. M. Clemens, “Effect of composition on vacancy
mediated diffusion in random binary alloys: First principles study of the Si1–xGex
system,” Journal of Applied Physics, vol. 94, no. 1, pp. 174–185, 2003.
[76] N. Zographos et al., “Multiscale modeling of doping processes in advanced
semiconductor devices,” Materials Science in Semiconductor Processing, vol. 62,
pp. 49–61, May 2017.
[77] R. T. Crosby, Evolution of Self-Interstitials Induced by Ion-Implantation in SiGe Alloys,
Ph.D. thesis, University of Florida, USA, 2005.
[78] V. Moroz et al., “Dissolution of extended defects in strained silicon,” Journal of
Vacuum Science & Technology B, vol. 26, no. 1, pp. 439–442, 2008.
[79] P. Castrillo et al., “Physical modeling and implementation scheme of native defect
diffusion and interdiffusion in SiGe heterostructures for atomistic process simulation,”
Journal of Applied Physics, vol. 109, p. 103502, May 2011.
[80] C. Zechner and N. Zographos, “Silicon germanium interdiffusion in SiGe device
fabrication: A calibrated TCAD model,” Materials Science in Semiconductor
Processing, vol. 42, pp. 230–234, February 2016.
[81] R. Kube et al., “Composition dependence of Si and Ge diffusion in relaxed Si1–xGex
alloys,” Journal of Applied Physics, vol. 107, p. 073520, April 2010.
[82] Y. Dong et al., “A unified interdiffusivity model and model verification for tensile and
relaxed SiGe interdiffusion over the full germanium content range,” Journal of Applied
Physics, vol. 111, p. 044909, February 2012.
[83] G. Xia, J. L. Hoyt, and M. Canonico, “Si–Ge interdiffusion in strained Si/strained SiGe
heterostructures and implications for enhanced mobility metal-oxide-semiconductor
field-effect transistors,” Journal of Applied Physics, vol. 101, p. 044901, February 2007.
[84] N. E. B. Cowern et al., “Diffusion in Strained Si(Ge),” Physical Review Letters, vol. 72,
no. 16, pp. 2585–2588, 1994.
[85] H. H. Silvestri et al., “Diffusion of silicon in crystalline germanium,” Semiconductor
Science and Technology, vol. 21, no. 6, p. 758, 2006.
[86] G. Xia and J. L. Hoyt, “Si–Ge interdiffusion under oxidizing conditions in epitaxial
SiGe heterostructures with high compressive stress,” Applied Physics Letters, vol. 96,
p. 122107, March 2010.
[87] A. N. Larsen and P. Kringhoj, “Diffusion of Sb in relaxed Si1–xGex,” Applied Physics
Letters, vol. 68, no. 19, pp. 2684–2686, 1996.

Advanced Calibration for Process Simulation User Guide 99

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[88] P. Laitinen, Self- and Impurity Diffusion in Intrinsic Relaxed Silicon - Germanium,
Ph.D. thesis, University of Jyväskylä, Finland, 2004.
[89] N. Zangenberg, Defect and Diffusion Studies in Si and SiGe, Ph.D. thesis, University of
Aarhus, Denmark, 2003.
[90] P. Kuo et al., “Boron Diffusion in Si and Si1–xGex,” in MRS Symposium Proceedings,
Strained Layer Epitaxy - Materials, Processing, and Device Applications, vol. 379,
pp. 373–378, 1995.
[91] N. Moriya et al., “Boron Diffusion in Strained Si1–xGex Epitaxial Layers,” Physical
Review Letters, vol. 71, no. 6, pp. 883–886, 1993.
[92] K. Rajendran and W. Schoenmaker, “Studies of boron diffusivity in strained Si1–xGex
epitaxial layers,” Journal of Applied Physics, vol. 89, no. 2, pp. 980–987, 2001.
[93] S. Uppal et al., “Diffusion of Boron in Germanium and Si1–xGex (x>50%) alloys,” in
MRS Symposium Proceedings, CMOS Front-End Materials and Process Technology,
vol. 765, p. D6.16.1, January 2003.
[94] C. Ahn, Atomic scale modeling of stress and pairing effects on dopant behavior in
silicon, Ph.D. thesis, University of Washington, USA, 2007.
[95] S. Eguchi et al., “Comparison of arsenic and phosphorus diffusion behavior in
silicon–germanium alloys,” Applied Physics Letters, vol. 80, no. 10, pp. 1743–1745,
2002.
[96] S. Eguchi et al., “Germanium-concentration dependence of arsenic diffusion in silicon
germanium alloys,” Applied Physics Letters, vol. 84, no. 3, pp. 368–370, 2004.
[97] Y.-M. Sheu et al., “Modeling Dopant Diffusion in Strained and Strain-Relaxed Epi-
SiGe,” in International Conference on Simulation of Semiconductor Process and
Devices (SISPAD), Tokyo, Japan, pp. 75–78, September 2005.
[98] C. C. Wang et al., “Boron diffusion in strained and strain-relaxed SiGe,” Materials
Science and Engineering B, vol. 124–125, pp. 39–44, December 2005.
[99] F. K. LeGoues et al., “Oxidation studies of SiGe,” Journal of Applied Physics, vol. 65,
no. 4, pp. 1724–1728, 1989.
[100] T. Tezuka, N. Sugiyama, and S. Takagi, “Fabrication of strained Si on an ultrathin SiGe-
on-insulator virtual substrate with a high-Ge fraction,” Applied Physics Letters, vol. 79,
no. 12, pp. 1798–1800, 2001.
[101] Z. Y. Xue et al., “Study of Ge loss during Ge condensation process,” Thin Solid Films,
vol. 557, pp. 120–124, April 2014.
[102] M. Spadafora et al., “Oxidation rate enhancement of SiGe epitaxial films oxidized in dry
ambient,” Applied Physics Letters, vol. 83, no. 18, pp. 3713–3715, 2003.
[103] N. Sugiyama et al., “Temperature effects on Ge condensation by thermal oxidation of
SiGe-on-insulator structures,” Journal of Applied Physics, vol. 95, no. 8,
pp. 4007–4011, 2004.

100 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[104] B. Vincent et al., “Fabrication of SiGe-on-insulator substrates by a condensation

technique: an experimental and modelling study,” Semiconductor Science and
Technology, vol. 22, no. 3, pp. 237–244, 2007.
[105] E. Long et al., “Nano-structuring in SiGe by oxidation induced anisotropic Ge self-
organization,” Journal of Applied Physics, vol. 113, no. 10, p. 104310, 2013.
[106] O. Gourhant et al., “Ge Condensation using Rapid Thermal Oxidation for SGOI
Substrate Preparation,” ECS Transactions, vol. 64, no. 6, pp. 469–478, 2014.
[107] V. Boureau et al., “Strain/composition interplay in thin SiGe layers on insulator
processed by Ge condensation,” Materials Science in Semiconductor Processing,
vol. 42, pp. 251–254, February 2016.
[108] T. Shimura et al., “Self-limiting oxidation of SiGe alloy on silicon-on-insulator wafers,”
Applied Physics Letters, vol. 89, no. 11, p. 111923, 2006.
[109] Y. Zhang et al., “Experimental evidence of oxidant-diffusion-limited oxidation of SiGe
alloys,” Journal of Applied Physics, vol. 106, no. 6, p. 063508, 2009.
[110] W. S. Liu et al., “Wet oxidation of GeSi at 700 °C,” Journal of Applied Physics, vol. 71,
no. 8, pp. 4015–4018, 1992.
[111] M. Tanaka et al., “Abnormal oxidation characteristics of SiGe/Si-on-insulator structures
depending on piled-up Ge fraction at SiO2/SiGe interface,” Journal of Applied Physics,
vol. 103, no. 5, p. 054909, 2008.
[112] G. L. Bir and G. E. Pikus, Symmetry and Strain-Induced Effects in Semiconductors, New
York: John Wiley & Sons, 1974.
[113] C. G. Van de Walle, “Strain effects on the valence-band structure of SiGe,” in Properties
of Silicon Germanium and SiGe:Carbon, EMIS Datareviews Series, pp. 135–139,
London: INSPEC, The Institute of Electrical Engineers, 2000.
[114] M. Diebel, Application of Ab-initio Calculations to Modeling of Nanoscale Diffusion
and Activation in Silicon, Ph.D. thesis, University of Washington, Seattle, WA, USA,
2004.
[115] Y. Kawamura et al., “Self-diffusion in compressively strained Ge,” Journal of Applied
Physics, vol. 110, p. 034906, August 2011.
[116] C. Ahn et al., “Stress effects on impurity solubility in crystalline materials: A general
model and density-functional calculations for dopants in silicon,” Physical Review B,
vol. 79, no. 7, p. 073201, 2009.
[117] R. F. Lever, J. M. Bonar, and A. F. W. Willoughby, “Boron diffusion across
silicon–silicon germanium boundaries,” Journal of Applied Physics, vol. 83, no. 4,
pp. 1988–1994, 1998.
[118] J. S. Christensen, Dopant diffusion in Si and SiGe, Ph.D. thesis, KTH, Royal Institute of
Technology, Stockholm, Sweden, 2004.

Advanced Calibration for Process Simulation User Guide 101

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[119] G. Hobler and S. Selberherr, “Two-Dimensional Modeling of Ion Implantation Induced

Point Defects,” IEEE Transactions on Computer-Aided Design, vol. 7, no. 2,
pp. 174–180, 1988.
[120] M. D. Giles, “Transient Phosphorus Diffusion Below the Amorphization Threshold,”
Journal of the Electrochemical Society, vol. 138, no. 4, pp. 1160–1165, 1991.
[121] A. Renau, “Device performance and yield - A new focus for ion implantation,” in
Extended Abstracts of International Workshop on Junction Technology (IWJT),
Shanghai, China, May 2010.
[122] F. Khaja et al., “Benefits of Damage Engineering for PMOS Junction Stability,” in 18th
International Conference on Ion Implantation Technology (IIT), Kyoto, Japan, June
2010.
[123] F. A. Khaja et al., “Physical understanding of cryogenic implant benefits for electrical
junction stability,” Applied Physics Letters, vol. 100, no. 11, p. 112102, 2012.
[124] M. Posselt, L. Bischoff, and J. Teichert, “Influence of dose rate and temperature on ion-
beam-induced defect evolution in Si investigated by channeling implantation at different
doses,” Applied Physics Letters, vol. 79, no. 10, pp. 1444–1446, 2001.
[125] R. D. Goldberg, J. S. Williams, and R. G. Elliman, “Amorphization of silicon by
elevated temperature ion irradiation,” Nuclear Instruments and Methods in Physics
Research B, vol. 106, no. 1–4, pp. 242–247, 1995.
[126] H.-J. Gossmann et al., “Predictive Process Simulation of Cryogenic Implants for
Leading Edge Transistor Design,” in 19th International Conference on Ion Implantation
Technology (IIT), Valladolid, Spain, pp. 225–228, June 2012.
[127] N. Zographos, A. Tsibizov, and C. Zechner, “Continuum Modeling of Implantation and
Thermal Processes for Advanced Devices Formation,” in 20th International Conference
on Ion Implantation Technology (IIT), Portland, OR, USA, pp. 189–194, June 2014.
[128] J. Li and K. S. Jones, “{311} defects in silicon: The source of the loops,” Applied
Physics Letters, vol. 73, no. 25, pp. 3748–3750, 1998.
[129] C. Zechner et al., “Simulation of dopant diffusion and activation during flash lamp
annealing,” Materials Science and Engineering B, vol. 154-155, pp. 20–23, December
2008.
[130] S. Solmi, F. Baruffaldi, and R. Canteri, “Diffusion of boron in silicon during post-
implantation annealing,” Journal of Applied Physics, vol. 69, no. 4, pp. 2135–2142,
1991.
[131] M. Diebel and S. T. Dunham, “Ab Initio Calculations to Model Anomalous Fluorine
Behavior,” Physical Review Letters, vol. 93, no. 24, p. 245901, 2004.
[132] M. Diebel and S. T. Dunham, “Reply to Ab Initio Calculations to Model Anomalous
Fluorine Behavior,” Physical Review Letters, vol. 96, p. 039602, January 2006.

102 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[133] K. Vollenweider et al., “Fluorine clustering and diffusion in silicon: Ab initio

calculations and kinetic Monte Carlo model,” Journal of Vacuum Science & Technology
B, vol. 28, no. 1, pp. C1G1–C1G6, 2010.
[134] P. A. Schultz and J. S. Nelson, “Fast through-bond diffusion of nitrogen in silicon,”
Applied Physics Letters, vol. 78, no. 6, pp 736–738, 2001.
[135] N. Stoddard et al., “Ab Initio Identification of the Nitrogen Diffusion Mechanism in
Silicon,” Physical Review Letters, vol. 95, no. 2, p. 025901, 2005.
[136] J. S. Nelson, P. A. Schultz, and A. F. Wright, “Valence and atomic size dependent
exchange barriers in vacancy-mediated dopant diffusion,” Applied Physics Letters,
vol. 73, no. 2, pp 247–249, 1998.
[137] J. P. Goss et al., “Vibrational modes and electronic properties of nitrogen defects in
silicon,” Physical Review B, vol. 67, no. 4, p. 045206, 2003.
[138] G. L. Olson and J. A. Roth, “Kinetics of Solid Phase Crystallization in Amorphous
Silicon,” Materials Science Reports, vol. 3, no. 1, pp. 1–78, 1988.
[139] S. Mirabella et al., “Fluorine segregation and incorporation during solid-phase epitaxy
of Si,” Applied Physics Letters, vol. 86, no. 12, p. 121905, 2005.
[140] I. Martin-Bragado and V. Moroz, “Facet formation during solid phase epitaxy regrowth:
A lattice kinetic Monte Carlo model,” Applied Physics Letters, vol. 95, p. 123123,
September 2009.
[141] K. L. Saenger et al., “A study of trench-edge defect formation in (001) and (011) silicon
recrystallized by solid phase epitaxy,” Journal of Applied Physics, vol. 101, no. 2,
p. 024908, 2007.
[142] V. C. Venezia et al., “Dopant redistribution effects in preamorphized silicon during low
temperature annealing,” in IEDM Technical Digest, Washington, DC, USA,
pp. 489–492, December 2003.
[143] S. Mirabella et al., “Mechanism of Boron Diffusion in Amorphous Silicon,” Physical
Review Letters, vol. 100, p. 155901, April 2008.
[144] K. Suzuki et al., “Analytical Model for Redistribution Profile of Ion-Implanted
Impurities During Solid-Phase Epitaxy,” IEEE Transactions on Electron Devices,
vol. 54, no. 2, pp. 262–271, 2007.
[145] R. Duffy et al., “Dopant diffusion in amorphous silicon,” in MRS Symposium
Proceedings, Silicon Front-End Junction Formation—Physics and Technology,
vol. 810, San Francisco, CA, USA, p. C10.2.1, April 2004.
[146] R. Duffy et al., “Influence of preamorphization and recrystallization on indium doping
profiles in silicon,” Journal of Vacuum Science & Technology B, vol. 22, no. 3,
pp. 865–868, 2004.
[147] G. Impellizzeri et al., “Fluorine in preamorphized Si: Point defect engineering and
control of dopant diffusion,” Journal of Applied Physics, vol. 99, p. 103510, May 2006.

Advanced Calibration for Process Simulation User Guide 103

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[148] G. R. Nash et al., “Activation energy for fluorine transport in amorphous silicon,”
Applied Physics Letters, vol. 75, no. 23, pp. 3671–3673, 1999.
[149] M. Rahimo et al., “Thin-Wafer Silicon IGBT With Advanced Laser Annealing and
Sintering Process,” IEEE Electron Device Letters, vol. 33, no. 11, pp. 1601–1603, 2012.
[150] T. Gutt and H. Schulze, “Deep melt activation using laser thermal annealing for IGBT
thin wafer technology,” in Proceedings of the 22nd International Symposium on Power
Semiconductor Devices and ICs (ISPSD), Hiroshima, Japan, pp. 29–32, June 2010.
[151] K. Huet et al., “High Performance and High Yield Junction Formation with Full Device
Exposure Laser Thermal Annealing,” in International Image Sensor Workshop (IISW),
Hokkaido, Japan, p. R12, June 2011.
[152] A. Karma and W.-J. Rappel, “Quantitative phase-field modeling of dendritic growth in
two and three dimensions,” Physical Review E, vol. 57, no. 4, pp. 4323–4349, 1998.
[153] IGBT Process Simulation With Backside Melt Laser Annealing and Its Device
Simulations for DC, Switching, and Short-Circuit Characteristics, available from
TCAD Sentaurus Version P-2019.03 installation, go to Applications_Library/
Power/IGBT_MeltLaserAnneal.
[154] P. D. Desai, “Thermodynamic Properties of Iron and Silicon,” Journal of Physical and
Chemical Reference Data, vol. 15, no. 3, pp. 967–983, 1986.
[155] A. La Magna et al., “A phase-field approach to the simulation of the excimer laser
annealing process in Si,” Journal of Applied Physics, vol. 95, no. 9, pp. 4806–4814,
2004.
[156] B. L. Zink, R. Pietri, and F. Hellman, “Thermal Conductivity and Specific Heat of Thin-
Film Amorphous Silicon,” Physical Review Letters, vol. 96, no. 5, p. 055902, 2006.
[157] Y. He, D. Donadio, and G. Galli, “Heat transport in amorphous silicon: Interplay
between morphology and disorder,” Applied Physics Letters, vol. 98, p. 144101, April
2011.
[158] A. Mittiga, L. Fornarini, and R. Carluccio, “Numerical modeling of laser induced phase
transitions in silicon,” Applied Surface Science, vol. 154–155, pp. 112–117, February
2000.
[159] P. Roura et al., “Measurement of the specific heat and determination of the
thermodynamic functions of relaxed amorphous silicon,” Journal of Applied Physics,
vol. 113, p. 173515, May 2013.
[160] E. Yamasue et al., “Thermal conductivities of silicon and germanium in solid and liquid
states measured by non-stationary hot wire method with silica coated probe,” Journal of
Crystal Growth, vol. 234, no. 1, pp. 121–131, 2002.
[161] S. de Unamuno and E. Fogarassy, “A Thermal Description of the Melting of c- and a-
Silicon Under Pulsed Excimer Lasers,” Applied Surface Science, vol. 36, no. 1–4,
pp. 1–11, 1989.

104 Advanced Calibration for Process Simulation User Guide

P-2019.03
 2: Advanced Calibration for Silicon, SiGe, and Germanium
References

[162] G. E. Jellison Jr., “Optical functions of silicon determined by two-channel polarization

modulation ellipsometry,” Optical Materials, vol. 1, no. 1, pp. 41–47, 1992.
[163] ICT Project 258547 ATEMOX, Advanced Technology Modeling for Extra-
Functionality Devices, for more information, go to https://www.atemox.eu/index.htm.
[164] J. P. Garandet, “New Determinations of Diffusion Coefficients for Various Dopants in
Liquid Silicon,” International Journal of Thermophysics, vol. 28, no. 4, pp. 1285–1303,
2007.
[165] M. Hackenberg et al., “Modeling Boron Profiles in Silicon after Pulsed Excimer Laser
Annealing,” in 19th International Conference on Ion Implantation Technology (IIT),
Valladolid, Spain, pp. 241–244, June 2012.
[166] K. Huet et al., “Experimental and Theoretical analysis of Dopant Activation in Double
Implanted Silicon by Pulsed Laser Thermal Annealing,” in 17th International
Conference on Advanced Thermal Processing of Semiconductors (RTP), Albany, NY,
USA, September 2009.

Advanced Calibration for Process Simulation User Guide 105

P-2019.03
2: Advanced Calibration for Silicon, SiGe, and Germanium
References

106 Advanced Calibration for Process Simulation User Guide

P-2019.03
CHAPTER 3 Guidelines for Additional
Calibration

This chapter provides guidelines for additional calibration.

The Advanced Calibration file is the recommended starting point for accurate process
simulation with Sentaurus Process. However, the Advanced Calibration cannot fully replace an
additional calibration by the user for 2D or 3D applications. With a customized process
calibration, the accuracy can always be further increased for any technology of interest. A
customized calibration of process and device simulation models needs to be performed by the
user or can be requested from Synopsys in the context of a customer service project.

To further improve the Advanced Calibration, Synopsys appreciates feedback from customers
regarding the accuracy obtained with the parameter files for different process conditions, and
suggestions for improved models or parameter values.

Accuracy and Limitations of Advanced Calibration of

Sentaurus Process
The Advanced Calibration is based on scientific literature on process simulation models and
on a continual calibration effort based on the Synopsys collection of SIMS profiles from state-
of-the-art device manufacturing technology. A good agreement is obtained for a large portion
of the SIMS data for silicon. However, in many cases, there is a significant mismatch between
simulation results obtained with Advanced Calibration and the experimental data for several
reasons:
■ Many models are simplifications of real physics.
■ Only a few parameters of diffusion and reaction physics in silicon can be determined by
direct measurements. Therefore, the calibration is difficult.
■ A very large range of possible experiment data needs to be reproduced with a single,
consistent set of models. For example, in standard CMOS technology, dopant
16 –3 22 –3
concentrations range from 10 cm to 10 cm ; temperatures range from 500°C to
1350°C .
■ In extreme conditions, the models used are often overburdened. For example, the dopant
21 –3
clustering models, which work well at dopant concentrations up to 10 cm , are less
reliable at higher dopant concentrations, which might occur after high-dose ion
implantation at very low energies.

Advanced Calibration for Process Simulation User Guide 107

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

■ The experimental data is not perfectly accurate. Errors arise from insufficient equipment
calibration (implanter dose, furnace temperature), from SIMS measurements, and from
missing details in the process description such as thickness of the oxide layer on top of
silicon. As a result, occasionally, small discrepancies are observed between data obtained
by different groups for the same nominal process conditions.
■ For some physical phenomena, no adequate calibration is available, either due to the lack
of a physical model in Sentaurus Process or to the lack of data for performing a reliable
calibration.

In this section, the accuracy of the Advanced Calibration is discussed in detail. In particular, it
will be explained for which process conditions the accuracy is limited and which parameters
can be tuned by users to increase the accuracy in a process window of interest. Unless
mentioned otherwise, the discussion focuses on the default model switches of Advanced
Calibration. For the option AdvancedModels (part 4 of AdvCal_2019.03.fps), see Part 4:
Comprehensive and Slow Models on page 68.

Error Control
By default, the Math option NegErrCntrl (as well as FTS.NegErrCntrl and
AMS.NegErrCntrl) is switched off. When switched on, NegErrCntrl enforces stricter error
control at Newton iterations and, typically, leads to an increased CPU time for the simulation
of thermal annealing.

With NegErrCntrl switched off, the simulation results can include the corresponding
numeric errors. For most applications, these numeric errors are irrelevant. However, it is
worthwhile checking the numeric errors at least once for a process simulation setup. If
simulations with and without the option NegErrCntrl produce different results, it is
recommended to switch on NegErrCntrl (or FTS.NegErrCntrl and AMS.NegErrCntrl)
either globally or for selected diffusion steps and data fields, to avoid numeric errors.

Point Defects

Bulk Parameters

The parameter values for the equilibrium concentration, diffusivity, and charge-state
distribution of point defects in silicon have been chosen by Synopsys, as a careful compromise
between various suggestions in recent publications [1][2][3][4][5][6][7].

The bulk recombination is based on the assumption that there is no energy barrier for I–V
recombination, and that interstitials and vacancies with the same charge state do not
recombine.

108 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

Changing any of these parameters might affect the diffusion and activation behavior of several
dopants. Therefore, for the purpose of improving the accuracy of diffusion of one dopant, it is
not recommended to change point-defect parameters for both silicon and germanium, due to
the possible unwanted effects on other dopants.

Surface Boundary Conditions

At the Si–SiO2 interface, fast recombination is assumed for interstitials and vacancies. The
recombination length is assumed to be 1 nm. For interfaces to other materials (for example,
nitride), the point-defect surface boundary conditions have not yet been calibrated, due to a
lack of suitable data.

Oxidation-Enhanced Diffusion

Oxidation-enhanced diffusion (OED) is simulated by interstitial injection and a reduced

concentration of vacancies at the moving Si–SiO2 interface. The interstitial injection rate
depends on the local oxide growth rate at the interface. The calibration for the intrinsic
condition is based on literature data for the growth of thick oxide in O2 atmosphere. For
additional fine-tuning of OED in lowly doped regions, it is recommended to adjust the
parameter theta. For a separate fine-tuning of OED in highly doped regions, it is
recommended to adjust the parameters m, mm for n-type doping and the parameters p, pp for p-
type doping.

In the process simulation of NMOS devices, increasing the value of theta during poly
reoxidation typically increases the reverse short-channel effect by enhancing the B pileup
towards the Si–SiO2 interface under the gate, which is driven by the injection of interstitials at
the oxidizing surfaces.

Clusters of Interstitials
The one-moment model suggested by Rafferty et al. [8] is used in the Advanced Calibration.
It gives a reasonably accurate description of Si self-interstitial supersaturation during anneals,
in which the transient-enhanced diffusion (TED) is dominated by the release and capture of
interstitials by {311} defects. This is the case for annealing at medium or high temperatures
(T > 800°C ) after nonamorphizing ion implantation.

The model is not suitable for investigating the initial stage of TED for low-temperature
annealing (< 800°C ). For processes where the initial phase of TED is crucial, the Rafferty
model is not the most accurate choice. Instead, consider using the procedure
AdvancedModels to benefit from complete modeling of small interstitial clusters, {311}
defects, and dislocation loops.

Advanced Calibration for Process Simulation User Guide 109

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

After amorphizing implantation, dislocation loops might form at the amorphous–crystalline

interface. Loops are known to be much more stable interstitial clusters than {311} defects. As
an effect, the rate of interstitial release is smaller. The formation and dissolution of dislocation
loops is not included in the 1Moment model for TED. However, you can imitate the presence
of dislocation loops by reducing the evaporation rate Ikr in the Rafferty model, after
amorphizing implantations:
pdbSetDouble Silicon ICluster Ikr {[Arr value1 value2]}

With the default value of Ikr in AdvCal_2019.03.fps, the stability of the interstitial
clusters is adjusted to the stability of {311} defects and is significantly lower than the expected
stability of dislocation loops.

Figure 15 shows the TEM data on the amount of interstitials in {311} clusters after a 40 keV,
13 –2
5 × 10 cm silicon implantation and annealing at different temperatures. The experimental
data (symbols) from the literature [9] is compared to the simulation results with
AdvCal_2019.03.fps (solid lines).

1014
Interstitials in {311} Defects

1013

670ºC
1012 738°C
815°C

1011
100 101 102 103 104 105
Time [s]

Figure 15 TEM data points on the amount of interstitials in {311} clusters after 40 keV,
5 x 1013 cm–2 silicon implantation and annealing at different temperatures [9]
compared to simulation results (lines)

If the Full cluster model is switched on for interstitial clusters by AdvancedModels, the
initial conditions after implantation are specified by:
pdbSet Si I2 InitPercent 1.0

This means excess interstitials are placed into I2 clusters at the beginning of post-implantation
anneals. In most simulations, results are not sensitive to this particular choice of initial
conditions. For very low (< 600°C) temperature anneals of B and P, or for thermal anneals that

110 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

start with a very slow ramp from very low temperatures, the TED of B and P can be sensitive
to these initial conditions. In such cases, you can adjust the diffusion tail from TED at very low
temperatures, by initially placing interstitials into either more stable clusters or a combination
of clusters, for example:
pdbSet Si I2 InitPercent 0.5
pdbSet Si I3 InitPercent 0.5

Vacancy Clusters

In the Advanced Calibration, vacancy clusters are not taken into account. This is justified for
most processes, but not for the so-called defect-engineering [10], where vacancy-rich silicon
regions are created near the surface by high-energy ion implantation, in order to form highly
activated and steep boron profiles.

Boron Diffusion and Clustering

Diffusion and Pairing in Silicon

The diffusivity of boron has been measured by many groups, with similar but not fully identical
results [1]. Omitting the most extreme published values, the spread between the lowest and
highest diffusivities for B is approximately a factor of two for high temperatures
( 900°C – 1100°C ). You should not change the B diffusivity by more than 30% in your own
calibrations. At temperatures less than 800°C , the B diffusivity is much less reliable. In
AdvCal_2019.03.fps, boron diffuses only using B–I pairs, with the diffusion coefficient
given by:
pdbSetDoubleArray Si B Int D { 0 {[Arr 0.123 3.57]}
1 {[expr [Arr 3.71 3.67]+[Arr 2.5e-6 2.5]]}
2 {[Arr 39.8 4.37]} }

Here, the first component represents B–I pair diffusion using B–I0 pairs, the second line
represents diffusion using B–I+ pairs, and the third line represents diffusion using B–I++ pairs.

The largest contribution comes from B–I+ pairs. For high temperatures (T > 800°C ), the term
[Arr 3.71 3.67] dominates. At lower temperatures, a significant contribution comes also
from the term [Arr 2.5e-6 2.5], which has a much smaller Arrhenius prefactor and a
smaller Arrhenius energy. A fine-tuning of B diffusivity can be performed as follows:
■ To increase or decrease B diffusivity in general, all contributions to B diffusivity can be
increased by the same factor. This factor should not be far from 1.0 in order not to
contradict literature data.
■ To increase or decrease B diffusivity at low temperatures only, change the prefactor of the
term [Arr 2.5e-6 2.5]. For example, in NMOS simulations, this can be useful to adjust

Advanced Calibration for Process Simulation User Guide 111

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

the B pileup towards the gate oxide during a low-temperature spacer deposition step and to
modify, thereby, the reverse short-channel effect. Sometimes, for modeling of B diffusion
at very low temperatures, it is beneficial to remove the [Arr 2.5e-6 2.5] term
completely.
■ To influence the shape of the B profile after thermal anneal, alter the relative contribution
of the diffusivity using B–I0, B–I+, and B–I++ pairs. For PMOS p-n junctions, increasing the
contribution of B–I0 will decrease the steepness of the tail, while increasing the contribution
of B–I++ will lead to a more box-like shape, as shown in Figure 16. For NMOS p-n
junctions, where B is used as a pocket dopant, increasing the contribution of B–I0 will
change the shape of the B profile near the p-n junction, as shown in Figure 17 on page 113.
For B and P, these changes must be undertaken with care in order not to conflict with direct
experimental data summarized in [1]. At temperatures higher than 900°C , to preserve the
agreement with literature investigation on B diffusivity, B–I+ should contribute at least 75%
of the B diffusivity in intrinsic silicon. For other species such as In, you might consider
more significant adjustments in the charge distribution of diffusing dopant–defect pairs.

B diffusivity: increased weight of B–I++

21
10 B diffusivty: AdvCal default
B diffusivity: increased weight of B–I0
As-implanted
Boron Concentration [cm–3]

20
10

1019

1018

Ge 10 keV 1e15 + B 500 eV 1e15 + spike anneal 1050°C

1017
0 0.01 0.02 0.03
Depth [µm]
Figure 16 Impact of B diffusivity on the shape of a B USJ profile. Increasing the relative
contribution of diffusion using B–I++ pairs to the diffusivity of B, while reducing the
relative contribution of B–I+ pairs, results in a more abrupt profile. Increasing the
relative weight of diffusion using B–I0 pairs results in a less steep profile.

112 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

B as-implanted
1019 B diffusivity: increased weight of B–I++
B diffusivity: AdvCal default
B diffusivity: increased weight of B–I0
Arsenic, annealed
Concentration [cm–3]

1018

As 2 keV 1e15 + BF2 22 keV 1e13 tilt 30 + spike anneal 1050°C

0 0.02 0.04
Depth [µm]

Figure 17 Impact of B diffusivity on the shape of a B pocket profile. Increasing the relative
contribution of diffusion using B–I++ pairs to the diffusivity of B, while reducing the
relative contribution of B–I+ pairs, results in a flatter B profile at the B side of the p-
n junction. Increasing the relative weight of diffusion using B–I0 pairs results in a
pronounced minimum of the B concentration close to the p-n junction, due to the
high electric field at the p-n junction.

A fine-tuning of the relative contributions of different dopant–defect charge states also can be
considered as a means of calibration for all other dopants. The effect is similar to that illustrated
in Figure 16 on page 112 and Figure 17. The pairing coefficient between B and interstitials is
less accurately known. However, it does not have a huge effect on the simulation results.

Boron diffusion during oxidation is enhanced because of interstitial injection at the oxidizing
silicon surfaces. To fine-tune boron diffusion during oxidation, first consider adjusting the
parameter theta, which scales the injection of interstitials at surfaces during oxidation.

Hopping Length (for ChargedReact Model)

The hopping length for B–I pairs has been extracted by Giles et al. [11]. It is only relevant if
the five-stream model is switched on by:
pdbSet Silicon Boron DiffModel ChargedReact

By default, the B diffusion model is ChargedPair. This is a simplified case of the

ChargedReact model and allows significantly faster simulations due to better convergence of
the equations and, therefore, larger time steps.

Advanced Calibration for Process Simulation User Guide 113

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

The hopping length lambdaK decreases with increasing temperature. For boron, at 600°C , it
is 11 nm; at 1000°C , it is 1.4 nm. The ChargedReact model gives significantly different
results from the ChargedPair model for the annealing of very steep B profiles at low
temperatures. This is illustrated in Figure 18 where a boron marker layer is annealed at 700°C .

Figure 18 shows that the ChargedPair model gives a Gaussian shape to the profile and the
ChargedReact model gives exponential-like tails, as observed in experiment. The slope of the
exponential tail is a measure for the hopping length of B–I pairs at 700°C .

Initial marker layer

1018 ChargedReact (five-stream)
ChargedPair (three-stream)
Boron [cm-3]

1017

1016

1015
0 50 100
Depth [nm]
Figure 18 Boron marker layer at depth 100 nm, before and after Si implantation and anneal
at 700oC

Furthermore, for spike annealing of shallow B implantations, there is sometimes a small

difference in the diffusion tail, which is formed at the very beginning of the ramp-up, where
the temperature is still low and the interstitial supersaturation is very high. In most situations,
it is well justified to use the faster ChargedPair model.

Effect of Fluorine

The presence of fluorine can affect the diffusion and activation of boron. It has been shown [12]
that F does not form clusters with B at low concentrations and, therefore, does not reduce the
diffusivity of B directly. Instead, fluorine–vacancy complexes, which are present in silicon
after solid phase epitaxial regrowth (SPER), were found to be able to capture silicon self-
interstitials. After capturing interstitials, F diffuses very fast towards the surface or deep into
the silicon bulk. In addition, it was observed [12][13] that F is redistributed towards the surface
during SPER of amorphized regions, due to a strong segregation effect at the
amorphous–crystalline interface. Furthermore, the speed of SPER is reduced by the presence
of F, which might cause an increased dopant redistribution during the regrowth.

114 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

In Advanced Calibration, you have two options to simulate the influence of fluorine on boron
diffusion. As the standard model of Advanced Calibration, the boron diffusivity is reduced in
regions with a high fluorine concentration. Fluorine is assumed to be immobile. The boron
diffusivity is performed by the function:
term name=BoronDiffFactor add Si \
eqn="(1.6e20+0.025*Fluorine)/(1.6e20+Fluorine)"

For many cases, the above formula for an effective reduction of B diffusivity by F is useful.
The formula has been calibrated by a comparison of SIMS data for boron diffusion after boron
and BF2 implantation. It often works well for the simulation of BF2 annealing but, sometimes,
the numbers used in the above formula need to be adjusted. The formula is not expected to be
predictive for the simulation of ultrashallow junction formation after separate implantation of
B and F atoms at different implantation energies.

To be more predictive in the case of fluorine coimplantation, the physics-based but slower
model can be used by calling the procedure:
AdvancedFluorineModel

This is called automatically if AdvancedModels is used, but it is also very useful in

combination with the standard models of Advanced Calibration. For fine-tuning of the
AdvancedFluorineModel, see Fluorine Diffusion and Clustering on page 136.

The diffusivity of B in oxide is increased in the presence of F. This effect has also been
calibrated for the simple model based on boron SIMS data measured after BF2 implantation
and annealing. As a result, the boron dose loss increases, as observed in the SIMS data.

Boron Clustering and Activation

As the default, the relatively simple Transient model is chosen for the clustering and
electrical activation of boron. Four substitutional B atoms can form a B cluster. Only the
substitutional B atoms will be assumed to be electrically active for the device simulation.

After ion implantation, only a limited concentration of B atoms is assumed to be substitutional.

19
This concentration depends on the implantation dose in crystalline silicon and is 7.0 × 10 in
amorphized silicon. The values have been calibrated using annealed boron SIMS profiles. It
should be mentioned that Colombeau et al. [14] reported higher values for the initial activation
in preamorphized silicon. The highly activated B after preamorphization and SPER was found
to be deactivated by the formation of boron–interstitial clusters BmIn (BICs) and reactivated
during the dissolution of these BICs. However, with the B clustering model chosen in the
Advanced Calibration, an initial activation well below solid solubility is necessary to reproduce
SIMS data of boron ultrashallow junctions.

Advanced Calibration for Process Simulation User Guide 115

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

The predictive power of the Transient boron clustering model is limited. The deactivation
and subsequent activation of boron USJ formed after preamorphization and low-energy
implantation cannot be simulated with the present model. For the fitting of SIMS data, a
satisfactory agreement has been obtained with a high percentage of B profiles from the
Synopsys database, but a good fit to all SIMS data cannot be achieved with the clustering
model used. The parameters in the Advanced Calibration file are an optimized compromise for
the complete collection of SIMS data of Synopsys. To optimize the accuracy in a process
window, additional user calibration might be necessary.

The solid solubility is fit to data collected by Pichler [1]. A double Arrhenius function is used
with a strong dependency of solubility on temperature at low temperatures and a smaller
dependency at high temperatures. Advanced Calibration defaults are:
pdbSet Si B Solubility {[ArrBreak 3.1614e24 1.062 7.964e21 0.4055 1000]}
pdbSet Si B TotSolubility {[ArrBreak 9.4842e24 1.062 23.892e21 0.4055 1000]}

You can sometimes achieve better results by reducing the difference in the Arrhenius energy of
boron solubility between low-temperature and high-temperature regions (especially if B
activation is underestimated after long-time anneals at very low temperatures). When doing so,
you should change parameters carefully in such a way that B solubility is steady at the
ArrBreak break temperature. For example, you can use:
pdbSet Si B Solubility {[ArrBreak 6.7155e23 0.9 2.7645e+22 0.55 1000]}
pdbSet Si B TotSolubility {[ArrBreak 2.0146e24 0.9 8.2935e+22 0.55 1000]}

In this parameter change, you keep the relation TotSolubility = 3.0*Solubility.

Alternatively, you can consider using a single Arrhenius function for Solubility with an
Arrhenius energy close to 0.65 eV. In addition, the following parameters are used frequently to
fine-tune B activation during thermal annealing:
term Si name=BoronTClusterForwardFac store add eqn= "(Int/EqInt)^2.2"
pdbSet Si B AmInit 7.0e19
pdbSet Si B CluRate {[Arr 1.65e13 5.6]}

The term BoronTClusterForwardFac takes into account that B clusters form faster in
regions of interstitial supersaturation (Int/EqInt > 1). It works well in many applications, but
not always. If B deactivation is too fast in the presence of excess interstitials, it is recommended
to reduce the exponent of (Int/EqInt) in the definition of BoronTClusterForwardFac to
a value smaller than 2.2, or even to set the entire term to 1.0. This might be necessary in SiGe
process modeling or when modeling boron deactivation during low-temperature backend
processes. In addition, as EqInt can become very small at low temperatures, you can consider
replacing (Int/EqInt) by the expression ((Int+1e10)/(EqInt+1e10)) or similar.
Adding 1e10 to the nominator and denominator is a measure to avoid unrealistically high
values of the ratio at very low temperatures.

116 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

The value of AmInit (7e19) is smaller than the B activation in recrystallized regions measured
in experiments (> 2e20). Using a ‘too low’ value for AmInit compensates for the deficiency
of the Transient model to accurately describe B deactivation in the first moment after SPER,
when a high supersaturation of interstitials from implantation leads to the fast formation of
BICs. For the simulation of processes with a very small thermal budget after implantation, it
might be necessary to increase AmInit to a value close to 2.3e20 (the value used in
AdvancedModels).

Finally, the clustering rate CluRate has been calibrated to reproduce boron SIMS profiles for
B or BF2 implantation followed by a single rapid thermal anneal (RTA) or spike anneal. The
effect of tuning the clustering rate is illustrated in Figure 19.
1021 {[Arr 4.00e13 5.6]}
{[Arr 1.65e13 5.6]} (default)
{[Arr 5.00e12 5.6]}
1020
{[Arr 1.00e12 5.6]}
Boron [cm-3]

1019

1018

1017
0 20 40 60 80
Depth [nm]
Figure 19 Effect of boron clustering rate CluRate on final shape of a boron profile
after 1 keV 1015 cm–2 boron implantation and spike annealing at 1050oC
20 –3
If boron clusters are located in regions of high carbon concentration (> 10 cm ), it might be
necessary to reduce the clustering rate greatly. The profile tail at boron concentrations less than
18 –3
10 cm is formed at the beginning of annealing, where most of the boron is still clustered.
Changing CluRate is a powerful method to fine-tune the junction depth and sheet resistance.

PMOS Extension Formation

In PMOS device fabrication, the B or BF2 implantation is sometimes followed by the following
temperature cycle:
■ RTA
■ Spacer deposition (typically, 20–120 minutes at 600°C – 800°C )
■ Spike anneal

In this case, the Advanced Calibration parameters often overestimate the release of boron
atoms from boron clusters, which are located close to the surface, into silicon. It seems that, in

Advanced Calibration for Process Simulation User Guide 117

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

the wafer processing, the B clusters become more stable during the low-temperature spacer
deposition step. However, since the model has only one type of B cluster, this type of
‘stabilization’ cannot be described. To obtain a good fit of the final B profile, it might be
necessary to reduce the parameter CluRate by a factor of 5–30.

NOTE For the process simulation of PMOS extension formation, it might be

necessary to reduce the parameter CluRate by a factor of 5–30.

As shown in Figure 19 on page 117, CluRate mainly influences the high-concentration

regions of the B profiles. The diffusion tail in the low-concentration regions can be adjusted by
fine-tuning the amount of interstitials created by ion implantation.

The most convenient way to make such an adjustment is to redefine the procedure
ifactor_Boron after sourcing the Advanced Calibration file. This is described in more detail
in Performing Additional Calibration on page 147.

The diffusion tail of a boron profile formed by a spike anneal can also be adjusted by varying
the stability of interstitial clusters. The higher the dissolution rate Ikr for interstitial clusters,
the earlier are the interstitials released during the ramp-up of the spike anneal. Consequently,
interstitials from ion implantation are available for B diffusion at a lower temperature, where
they are more effective for enhancement of B diffusion.

Figure 20 illustrates the effect of increasing the interstitial dissolution rate. The diffusion in the
tail region is widely driven by B atoms, which are substitutional at the beginning of the anneal.
These diffuse more if Ikr is increased. Another effect of increasing Ikr is that most of the
interstitials have already recombined before a substantial part of the B clusters that are near the
surface are dissolved.
1021
High Ikr, high B-CluRate
High Ikr
Default
1020
Boron [cm-3]

1019

1018

1017
0 20 40 60 80
Depth [nm]
Figure 20 Effect of interstitial cluster dissolution rate Ikr on the final shape of a boron profile
after 1 keV 1015 cm–2 B implantation and spike annealing at 1050°C

118 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

As a consequence, less boron can diffuse out of the B clusters, as can be seen in the dashed line
of Figure 20 on page 118. By changing the rate for interstitial cluster dissolution and B cluster
dissolution, both by a factor of ten (symbols in Figure 20), you can modify the B tail almost
independently from the highly doped region.

In Figure 20, the solid line shows the simulation result, and the dashed line shows the result
with Ikr increased by a factor of 10, and the symbols show the simulation result with Ikr
increased by a factor of 10 and boron CluRate increased by a factor of 2.

Preamorphization Implantation

If boron ultrashallow junctions are formed using a preamorphization implantation (Ge or Si),
the accuracy can be reduced by an inaccurate calculation of the initial amount of interstitials in
the structure. In this situation, you can consider adjusting the ifactor for Ge or Si
implantation, for example, by defining a procedure ifactor_Germanium. By default,
ifactor=1 is used for Ge implantation. Since Ge has a similar mass as arsenic, a much higher
value (2–10) can be justified, as illustrated for an amorphizing As implantation in Figure 5 on
page 61.

BIC Model

By executing the procedure AdvancedModels (see Part 4: Comprehensive and Slow Models
on page 68), you can switch on the ChargedCluster BIC model for boron. Its disadvantage
– six equations need to be solved instead of a single equation – can be compensated by an
increased accuracy of results. In particular, for the simulation of B diffusion and (de)activation
during low-temperature annealing, and of B (de)activation during millisecond annealing, the
ChargedCluster model is sometimes better suited than the Transient cluster model. For
the modeling of boron USJ formation by spike annealing, with high or extremely high
21 –3
( > 10 cm ) B surface concentration, the ChargedCluster model provides similar
accuracy to the Transient cluster model.

With Advanced Calibration default parameters, all clustered boron in recrystallized regions is
initially placed into B2 clusters. Possibly, as BICs can be energetically more favorable than B
clusters without interstitials bound inside, it can be beneficial to put some of the B into B3I
clusters after implantation. This will affect B activation kinetics and TED (by the release of
interstitials during dissolution of BICs), for example:
pdbSetDouble Si B2 FractionAmor 0.8 ;# default 1.0
pdbSetDouble Si B3I FractionAmor 0.2 ;# default 0.0

Boron Dose Loss

The parameters for the three-phase segregation model have been calibrated by Synopsys using
SIMS data from B implantation and annealing experiments. In most situations, the dose loss of
B into thermal oxide is accurately described. After BF2 implantation, the dose loss of B is

Advanced Calibration for Process Simulation User Guide 119

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

increased. This is taken into account by increasing the diffusivity of B in the oxide as a function
of F concentration.

In device manufacturing, thermal oxide is not the only dielectric material used. In oxynitride,
which is frequently used in gate dielectrics, the boron diffusivity is typically reduced,
depending on nitrogen concentration. Deposited oxides can contain hydrogen atoms, which
drastically increase the diffusivity of B in oxide. This might be important in devices with oxide
spacers that are capped by a silicon-nitride layer, which can act as a barrier for hydrogen out-
diffusion.

Arsenic Diffusion and Activation

The diffusion of arsenic in silicon is well calibrated for a wide range of process conditions,
including the formation of ultrashallow junctions. Only a few accuracy problems are known to
Synopsys:
■ The model for arsenic diffusion and clustering is not very accurate for extremely high
arsenic surface concentrations, which might occur after high-dose implantation with very
15 –2
low implantation energy (for example, 2 × 10 cm at an energy of 1 keV and less).
■ In some situations, it is useful to perform additional fine-tuning of arsenic dose loss. In
particular, the parameter CMax, which specifies the concentration of arsenic atoms that can
be built into the interface layer, can be considered for the fine-tuning of As ultrashallow
junctions (see Figure 21).

1021
Arsenic Concentration [cm−3]

1020

1019

CMax = {[Arr 1.00e+15 0.2]}

1018 CMax = {[Arr 2.64e+15 0.2]} (def.)

CMax = {[Arr 5.00e+15 0.2]}
As-implanted
1017
0 10 20
Depth [nm]

Figure 21 Effect of CMax on simulation result for arsenic USJ formation; graphs were
calculated for 1.5 keV 1015 cm–2 As implantation, followed by a spike anneal
at 1050oC (simulations performed with Advanced Calibration 2013.12)

120 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

For very shallow arsenic profiles, the steepness of the slope can be overestimated with the
ChargedPair diffusion model. A flatter slope can be obtained with the ChargedReact
model. Within the ChargedReact model, you can even adjust the steepness by tuning the
arsenic–defect hopping length lambda.

Figure 22 shows a comparison for an As ultrashallow junction, calculated with the

ChargedPair model and the ChargedReact model with default parameters. Note that SIMS
data tends to underestimate the slope of very steep profiles because, during the SIMS
measurement itself, the profile is flattened due to the ‘knock-on’ effect.

ChargedReact (five-stream)
1022
ChargedPair (three-stream)
As-implanted
Arsenic Concentration [cm−3]

1021

1020

1019

1018

1017

1016
0 10 20 30
Depth [nm]

Figure 22 Comparison of simulation results with the ChargedPair and ChargedReact

models for As USJ formation; graphs were calculated for 1.5 keV 1015 cm–2
As implantation, followed by a spike anneal at 1050oC (simulations performed
with Advanced Calibration 2013.12)

Phosphorus Diffusion and Activation

The calibration of diffusion of phosphorus in silicon covers a wide range of process conditions,
including USJ formation and phosphorus well anneals. Phosphorus has a higher solid solubility
than arsenic.

It should be mentioned though that for very high-dose and very low-energy ion implantation
15 –2 22 –3
(for example, 5 × 10 cm , 2 keV), where phosphorus concentrations close to 10 cm are
reached, the simulation underestimates the stability of P clusters. At such high concentrations,
Si cannot be expected to show a typical behavior.

Furthermore, if P implantation at low implantation energies is followed by low-temperature

thermal annealing, so-called uphill diffusion of P towards the surface has been observed in
experiments [15].

Advanced Calibration for Process Simulation User Guide 121

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

Uphill diffusion is not reproduced with the model for P in Advanced Calibration.

Phosphorus Diffusion in Silicon

In Advanced Calibration, at high donor concentrations, the diffusivities are modified by the
PhosphorusIntDiffFactor and PhosphorusVacDiffFactor correction factors, as
explained in Arsenic–Phosphorus Co-Diffusion on page 31.

By default, the phosphorus diffusion model is ChargedPair. For advanced CMOS technology
(gate length < 65 nm), if phosphorus is used for source/drain, LDD, or pocket implantation, it
is recommended to use the ChargedReact model. It can be switch on using:
pdbSet Si P DiffModel ChargedReact

In the ChargedReact model, to fine-tune the steepness of an almost perfectly exponential

diffusion tail, typical for post-implantation TED at low temperatures, you can consider
adjusting the parameters lambdaK, which govern the average distance between the formation
and dissolution of P–I and P–V pairs, for example:
pdbSet Si P Int lambdaK {[Arr 4.0e-9 -0.50]} ;# reduced
pdbSet Si P Vac lambdaK {[Arr 4.5e-9 -0.33]} ;# reduced

Phosphorus diffusion during oxidation is enhanced because of interstitial injection at the

oxidizing silicon surfaces. To fine-tune phosphorus diffusion during oxidation, first consider
adjusting the parameter theta, which scales the injection of interstitials at surfaces during
oxidation.

Phosphorus Activation in Silicon

In Advanced Calibration, activation and deactivation of P are modeled by the Transient
model, in which the only P clusters formed are P3. By default, in the Transient model,
interstitial supersaturation has no affect on the formation and dissolution of P3 clusters. In real
silicon, a high supersaturation of Si self-interstitials can influence the rate of P cluster
formation.

To fine-tune P clustering in post-implantation anneals with low thermal budget and high
interstitial supersaturation, the Transient model for P can be tuned in a similar way as for B,
by modifying forward (or backward) reaction rates with terms. For example, the P cluster
formation rate will increase because of interstitial supersaturation, when defining:
term Si name=PhosphorusTClusterForwardFac store add \
eqn= "((Int+1e10)/(EqInt+1e10))^0.2"

122 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

Here, Int is the interstitial concentration, EqInt is the equilibrium concentration of

interstitials, and 1e10 (or a similar number) is used to avoid too extreme values at very low
temperatures, where EqInt is very small.

Phosphorus Dose Loss at Oxide–Silicon Interfaces

As in the case of arsenic, a useful parameter for the fine-tuning of P segregation is CMax.

Co-Diffusion of Arsenic and Phosphorus

The physics of the co-diffusion of As and P in high concentration in silicon is very complex for
several reasons:
■ The concentration of point defects is affected by the diffusion of As–I, As–V, P–I, and P–V
pairs.
■ The recombination of point defects in the presence of As and P is not precisely known.
■ Arsenic and phosphorus can form mixed clusters.
■ The Fermi-level dependence of P diffusivity has a large influence, but it is not known with
high precision from the literature.
■ Arsenic diffusivity sharply increases at concentrations of n-type dopants (As or P) higher
20 –3
than 2 ×10 cm . Possible reasons for the increased As diffusion are:
(a) Diffusion is through small mobile clusters.
(b) Percolation-type diffusion of As–V pairs, where vacancies can hop from one donor
atom to neighboring donor atoms.

The Synopsys calibration of the co-diffusion of As and P is considered a good starting point
for technology calibration, but it is not predictive for all possible process conditions.
Depending on the process window of interest, different methods of parameter fine-tuning will
result in a good overall agreement.

The following approaches have been identified by Synopsys to be suitable for such fine-tuning.
They are listed in order of importance:
■ Modify the number of As or P trapping sites at the Si–SiO2 interface (CMax). This will
increase or decrease the dose loss. Typically for As–P coimplants, you need to reduce CMax
for P.
■ Modify the terms PhosphorusVacDiffFactorDopant and
PhosphorusIntDiffFactorDopant.
■ Modify the number of interstitials created by ion implantation (by changing ifactor or
dfactor for P or As implantation).

Advanced Calibration for Process Simulation User Guide 123

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

■ Switch on the ChargedReact model for phosphorus. When the ChargedReact model is
switched on for P, the hopping length of P–I pairs (pdbSet Si P Int lambdaK) can be
fine-tuned to adjust the steepness of the tail of the P profile outside the region with high As
doping.
■ Modify the ArsenicVacDiffFactorDopant and ArsenicTClusterBackwardFac
terms.
■ Create a term PhosphorusTClusterBackwardFac to enforce an influence of As on P
cluster dissolution. For example:
term Si name=PhosphorusTClusterBackwardFac store add \
eqn= "2.e19/(2e19+AsActive)"

The impact of these parameters is illustrated in Figure 23 to Figure 28 on page 127. All of these
have been calculated for the following process recipe for As–P co-diffusion:
15 –2
■ P implantation: Dose 2 ×10 cm , energy 8 keV
15 –2
■ As implantation: Dose 3 ×10 cm , energy 25 keV
■ Spike RTA with peak temperature 1050°C

In each figure, the straightforward simulation result, obtained with AdvCal_2013.12.fps, is

compared to the result of a simulation in which one parameter for arsenic or phosphorus has
been changed from the default.

Reduced CMax (dose loss) for P o

1021
AdvancedCalibration: P
AdvancedCalibration: As

1020
Concentration [cm−3]

1019

1018

P 8 keV, 2e15 + As 25 keV, 3e15 + spike anneal 1050°C

1017
0 50 100
Depth [nm]

Figure 23 As–P co-diffusion: The lines with triangles illustrate the impact of reducing CMax
for phosphorus from {[Arr 6.0e16 0.50]} to {[Arr 1.0e16 0.5]}. The dose loss of P
at the Si–SiO2 interface is reduced, but the impact on the profile tails is relatively
small.

124 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

NOTE While simulation results will differ slightly with the latest Advanced
Calibration file, the illustrated trends with respect to parameter changes
remain the same.

PhosphorusIntDiffFactor reduced in As region A

1021 AdvancedCalibration: P
AdvancedCalibration: As

1020
Concentration [cm−3]

1019

1018

P 8 keV, 2e15 + As 25 keV, 3e15 + spike anneal 1050°C

1017
0 50 100
Depth [nm]
Figure 24 As–P co-diffusion: The lines with triangles illustrate the impact of changing the
term PhosphorusIntDiffFactor as a function of As concentration. For this plot, the
factor PhosphorusIntDiffFactorDopant has been reduced from “(\[Arr 3.846e21
0.5\]/(\[Arr 3.846e21 0.5\]+Arsenic))” (Advanced Calibration default) to “(\[Arr
3.846e21 0.5\]/(\[Arr 3.846e21 0.5\]+3.0*Arsenic))”. As a consequence, the
diffusion tail for P shifts towards a lower concentration, while the As profile and
the surface-near region of the P profile have hardly changed.

Advanced Calibration for Process Simulation User Guide 125

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

PhosphorusVacDiffFactor increased in As region

1021 AdvancedCalibration: P
AdvancedCalibration: As

1020
Concentration [cm−3]

1019

1018

P 8 keV, 2e15 + As 25 keV, 3e15 + spike anneal 1050°C

1017
0 50 100
Depth [nm]

Figure 25 As–P co-diffusion: The lines with triangles illustrate the impact of changing the
term PhosphorusVacDiffFactor as a function of As concentration. For this plot, the
factor PhosphorusVacDiffFactorDopant has been increased from “(\[Arr 1.667e32
3.0\]/(\[Arr 1.667e32 3.0\]+Arsenic))” to “1.0”. There is more P diffusion, but not in
the P tail, which is dominated by P–I pair diffusion.

Reduced ifactor for As implantation

1021
AdvancedCalibration: P
AdvancedCalibration: As

1020
Concentration [cm−3]

1019

1018

P 8 keV, 2e15 + As 25 keV, 3e15 + spike anneal 1050°C

1017
0 50 100
Depth [nm]
Figure 26 As–P co-diffusion: The lines with triangles illustrate the impact of decreasing the
amount of interstitials created by the As implantation. The ifactor for As
implantation was reduced from 8.7 (Advanced Calibration default for this
condition) to 3.0. Phosphorus diffusion is reduced (in particular in the tail region).

126 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

ChargedReact model, lambda of P−I pairs increased

1021 ChargedReact model for As and P
AdvancedCalibration: P
AdvancedCalibration: As
1020
Concentration [cm−3]

1019

1018

P 8 keV, 2e15 + As 25 keV, 3e15 + spike anneal 1050°C

1017
0 50 100 150
Depth [nm]
Figure 27 As–P co-diffusion: The lines with symbols illustrate the impact of using the
ChargedReact diffusion model for As and P. The tail of the P profiles can be made
flatter by increasing the hopping length of P–I pairs. In the line with circles, the
corresponding parameter lambda has been increased by a factor of 2. (Using the
ChargedReact model has a larger impact on thermal anneals at lower
temperature.)

As cluster stability not reduced by P

1021
AdvancedCalibration: P
AdvancedCalibration: As

1020
Concentration [cm−3]

1019

1018

P 8 keV, 2e15 + As 25 keV, 3e15 + spike anneal 1050°C

1017
0 50 100
Depth [nm]

Figure 28 As–P co-diffusion: The lines with symbols illustrate the impact of reducing the
term ArsenicTClusterBackwardFac from “1.0+PActive/1e20” to “1.0”. The As
clusters become more stable in the region with high P co-doping.

Advanced Calibration for Process Simulation User Guide 127

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

Indium Diffusion and Activation

Nonamorphizing Condition

The calibration of indium diffusion in silicon and dose loss is very accurate for
nonamorphizing conditions and usually does not require additional fine-tuning.

Amorphizing Ion Implantation

If amorphous layers are created by ion implantation, the behavior of indium is more
complicated and cannot be modeled correctly with the chosen set of models in the Advanced
Calibration. Several things might happen:
■ Indium is swept out of silicon during solid phase epitaxial regrowth (SPER) of the
amorphous layer [16].
■ Indium might segregate to end-of-range defects [17].
■ If a high concentration of indium is left in silicon after SPER, clusters might form due to
the low solid solubility of indium.

The first effect is responsible for a pronounced increase of indium dose loss with increasing
indium dose [18] for NMOS channel doping, when the indium dose is above the amorphization
threshold. The procedure AdvancedSPERModel can be called to switch on a physics-based
description of this effect (see Solid Phase Epitaxial Regrowth on page 76). An alternative,
simple possibility for mimicking the sweep-out effect is to cut the indium concentration
manually in amorphous regions, immediately after the ion implantation, using the select
command of Sentaurus Process.

SPER Model Usage

The SPER phase field model can be used to simulate indium redistribution in the amorphized
silicon region. In the procedure AdvancedSPERModel, a general calibration is provided for
the doping or impurity redistribution, but additional calibration might be required in certain
cases. For solutions to be redistributed during SPER, the diffusivity in a-Si (DAmor) and the
segregation energy (SPER.Energy) must be defined. Nonzero diffusivities allow for diffusion
in a-Si, while nonzero segregation energies allow for the snow plow effect by a semipermeable
a/c interface. Figure 29 on page 129 shows the sensitivities of these parameters.

128 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

annealed (SIMS) annealed (SIMS)

1021 1021
annealed (SPER.Energy=2.8) annealed (SPER.Energy=2.8)
annealed (SPER.Energy=5.6) annealed (SPER.Energy=5.6)
as-implanted (SIMS) 20
as-implanted (SIMS)
1020 10
as-implanted as-implanted

Concentration [cm−3]
Concentration [cm−3]

1019 1019

1018 1018

1017 1017

1016 1016

0 50 100 150 0 50 100 150

Depth [nm] Depth [nm]

Figure 29 Simulation of (left) incomplete and (right) complete recrystallization with indium
redistribution during SPER for PAI + In 90 keV 4e13 cm–2 implantation and anneal
at 600oC. The influence of SPER.Energy on the indium redistribution is shown for
the calibrated value (green) and the doubled value (red). A similar influence has
DAmor giving almost the same simulation results for the calibrated and the
doubled value (not shown here).

As stated in Solid Phase Epitaxial Regrowth on page 76, a decent fine mesh (1–2 nm spacing)
in the amorphized region is required by the phase field model to be robust and accurate in one
and two dimensions. The mesh spacing in amorphized regions should not exceed the phase
transition width of 3 nm to allow for accurate and stable simulation results. Figure 30 on
page 130 shows the mesh spacing sensitivity.

Advanced Calibration for Process Simulation User Guide 129

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

SIMS SIMS
10 nm 10 nm
5 nm 19
5 nm
1019 10
3 nm 3 nm
2 nm 2 nm
1 nm 1 nm

Concentration [cm−3]
Concentration [cm−3]

1018 1018

1017 1017

1016 1016

0 50 100 0 50 100
Depth [nm] Depth [nm]

Figure 30 The influence of the mesh spacing on the indium redistribution is shown for the
simulation of (left) incomplete and (right) complete recrystallization with indium
redistribution during SPER for PAI + In 90 keV 4e13 cm–2 implantation and anneal
at 600oC. For incomplete SPER after 60 s (left), the recrystallization speed is
approximately the same for spacing < 3 nm, while it slows down significantly for
coarser meshes. The total indium redistribution after complete SPER after 120 s
(right) is comparable to spacing <= 3 nm. Note that the SPER for 10 nm is not
completed after 120 s, while it is completed for spacing < 10 nm.

Depending on the activation energy of the diffusivity in a-Si, the doping redistribution during
SPER has a specific temperature dependency. If the activation energy of the diffusivity in a-Si
is similar to the SPER one (2.68 eV), the redistributed dopant shows little temperature
dependency. However, if the activation energy of the diffusivity in a-Si is different from the
SPER one, the dopant redistribution is temperature dependent (see Figure 31 on page 131).
Therefore, you must take care about the correct anneal temperature profile during ramp-up,
even at low temperatures, to ensure accurate simulation.

130 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

SPER 700°C SPER 700°C

1020 1020
SPER 600°C SPER 600°C
SPER 500°C SPER 500°C
as-implanted as-implanted
1019 1019
Concentration [cm−3]

Concentration [cm−3]
1018 1018

1017 1017

1016 1016

0 50 100 0 50 100
Depth [nm] Depth [nm]

Figure 31 The influence of the anneal temperature and a-Si diffusivity on the indium
redistribution is shown for the simulation of complete recrystallization with indium
redistribution during SPER for PAI + In 90 keV 4e13 cm–2 implantation and
anneal. If the activation energy of the diffusivity in a-Si is higher (for example,
3.3 eV) than the SPER one (2.68 eV), more In is redistributed for higher
temperatures (left). If the activation energy of the diffusivity in a-Si is lower (for
example, 2.2 eV) than the SPER one (2.68 eV), more In is redistributed for lower
temperatures (right).

Antimony Diffusion and Activation

Antimony diffuses only with vacancies and, therefore, does not show transient enhanced
diffusion with excess interstitials created by ion implantation. In addition, the diffusivity of Sb
in silicon is low. The calibration of Sb diffusion and activation in silicon is based on the
literature [1] and Sb data for ion implantation and subsequent annealing, with implantation
energies of 50 keV and higher, and annealing temperatures ranging from 700°C to 1100°C .
For these conditions, the calibration is very reliable.

Antimony ultrashallow junction (USJ) formation in silicon has not been taken into account yet
in the Advanced Calibration.

For users who are interested in Sb USJ formation, an additional calibration of the Sb
parameters is required. In particular, it will probably be necessary to define more sophisticated
initial conditions for point defects and Sb activation after low-energy ion implantation, using
the procedure ifactor_Antimony, and the parameters AcInit and AmInit.

Advanced Calibration for Process Simulation User Guide 131

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

Carbon

Carbon–Interstitial Clusters

The NeutralCluster model has been used for C–I clusters in silicon to reproduce
experimental data for USJ formation based on Ge+C+B or C+P cocktail implantations and
subsequent spike anneals [19]. The model has not been calibrated yet for other process
conditions where C implantation can be used, such as low-temperature annealing or As+C
coimplantations.

The main effect of the C coimplantation in silicon is a reduction of interstitial supersaturation

due to interstitial trapping by C–I clusters, leading to a reduction of dopant TED and
deactivation. Carbon is mainly effective in combination with (pre)amorphization and
recrystallization. In recrystallized regions, C is incorporated without an extra I and, therefore,
is very efficient in trapping I. The simulation results are relatively sensitive to the
amorphization layer thickness generated during implantation and, therefore, the dfactor of
preamorphizing implantations, or self-amorphizing C implantations, are recommended fitting
factors.

For additional calibration of the C–I cluster model, you can consider modifying the initial
conditions for C after implantation. By default, it is assumed that, in regions amorphized by ion
implantation and recrystallized by solid phase epitaxy, carbon is in the substitutional state up
20 –3
to concentrations of 3.0 ×10 cm immediately after recrystallization. Higher concentrations
are clustered as C2. In nonamorphized regions, carbon is assumed to be mostly in C3I2 clusters
at the beginning of thermal annealing.

This assumption can be changed by the parameters AcInit, AmInit, FractionCryst, and
FractionAmor. For example:
pdbSetDouble Si Carbon AcInit 1e18
pdbSetDouble Si C2I1 FractionCryst 0.5
pdbSetDouble Si C3I2 FractionCryst 0.5
pdbSetDouble Si Carbon AmInit 1e20
pdbSetDouble Si C2 FractionAmor 1.0

With these parameter values, C would be initialized as follows: In crystalline regions, only a
18 –3
maximum of 1 ×10 cm (AcInit) C atoms is substitutional initially. The additional C atoms
are distributed to C2I clusters (50%) and to C3I2 clusters (50%). In amorphized regions, the first
20 –3
1 ×10 cm C atoms are put into substitutional sites. If the total concentration of C is higher,
the rest is put into C2 clusters initially. It should be mentioned that the interstitials that are
contained in the C–I clusters represent an independent additional contribution to the total
number of interstitials present after implantation.

132 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

For low C concentrations, the C–I model can be simplified by disabling the C2-cluster type,
since it is only formed at high concentrations. The following command disables the C2 cluster:
pdbSetString Si C Int ClusterSizes { {1 0} {1 1} {2 1} {3 2} {3 3} }

In addition, the initial conditions should not include C2.

Carbon–Boron Clusters

According to [20], B and C can form relatively stable mixed clusters in silicon. The C–B
clusters might not have a significant influence on B at high concentrations such as for USJ
formation based on Ge+C+B coimplants, but B at low concentrations such as in B-doped
NMOS channels can be deactivated by the presence of high C concentrations [21].

The following statements enable mixed C–B clusters in the form of the cluster type CBI, in
addition to the C–I clusters and B4 clusters of Advanced Calibration:
pdbSet Si B More.Active.Model.List ComplexCluster
pdbSet Si C More.Active.Model.List ComplexCluster

Arbitrary calibrated reaction rates with diffusion-limited formation and a binding energy of
2.0 eV are defined by:
pdbSet Si BCI KF {[expr 1.0e8*[DiffLimit Si Int 0.0]]}
pdbSet Si BCI KR {[Arr 1.0e5 2.0]}

To increase the agreement with SIMS and spreading resistance profiles of [21], cluster types
with higher C content are beneficial.

To enable the cluster type C2BI with arbitrary calibrated reaction rates, use the following
statements:
pdbSet Si BCI Component.List { Boron 1 Carbon 2 Int 1 }
pdbSet Si BCI KF {[expr 2.0e8*[DiffLimit Si Int 0.0]]}
pdbSet Si BCI KR {[Arr 1.0e3 2.0]}

Impact of Carbon on Hole Mobility

Zschätzsch [22] reported that the presence of C in silicon reduces the mobility of holes in boron
USJ causing an increase of the sheet resistance. To include this effect in device simulations, the
following workaround can be used:
■ At the end of the process simulation, 50% (or a lower percentage) of the C atoms is added
to the concentration of arsenic substitutional atoms, and the same percentage of the C
atoms is added to the B substitutional atoms, using the select command of Sentaurus
Process. This changes the sum of donors and acceptors, but does not change the net active

Advanced Calibration for Process Simulation User Guide 133

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

concentration of dopants. If As and B are present in the device, and if the Transient
model (default) is used for B and As activation, this can be performed by the lines:
select Si name=BActive z="BActive+CTotal/2.0" store
select Si name=Boron z="Boron+CTotal/2.0" store
select Si name=AsActive z="AsActive+CTotal/2.0" store
select Si name=Arsenic z="Arsenic+CTotal/2.0" store
diffuse temp=460 time=1e-30
■ In device simulations with Sentaurus Device, use the DopingDep mobility model for the
mobility in highly doped regions. In the DopingDep model, the total concentration of
donors and acceptors is used to calculate the carrier mobility.

An alternative method to take into account the impact of C on carrier mobility is to use a
physical model interface (PMI) for the calculation of mobility in high-concentration regions.
The PMI can be implemented in such a way that the presence of carbon reduces the mobility
of holes.

It is not known yet if carbon co-doping also impacts the mobility of electrons.

Molecular Implantation
Molecules such as borane (B10H14, B18H22) and carborane (C2B10H12) can be used to implant
C, B, and H at the same time [23][24]. In contrast to a subsequent implantation of C and B
atoms, large molecules are implanted. For TCAD modeling, the main difference is that a
molecule has a significantly higher capability to amorphize silicon. The many atoms of the
molecule, which all enter the wafer at essentially the same position, are more likely to produce
amorphous pockets in silicon than the implantation of individual atoms, which are not spatially
correlated. This difference is most significant for low-energy implantations where the path
length of the implanted atoms is relatively short. For high-energy implantations, the atoms of
the molecule eventually separate so far that the damage that they produce no longer overlaps.
Hydrogen is believed to out-diffuse rapidly without impact on B diffusion.

Carborane implantation is performed by implanting the species C2B10H12, which is defined

by default in Sentaurus Process:
implant C2B10H12 dose=1.0e14 energy=6.7

While the models for B diffusion and activation, and C–I cluster formation have been found to
work well after carborane implantation, care needs to be taken for the implantation and for
setting the initial conditions after implantation. The file AdvCal_2019.03.fps includes a
calibration for implantation of BF2 and carborane, but not for other molecules. This section
summarizes the current recommendations of Synopsys for molecular implantation (other than
BF2).

134 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

Analytic Implantation

Sentaurus Process does not include implantation tables for molecules such as borane and
carborane. Therefore, if analytic implantation is used for borane or carborane, Sentaurus
Process converts the borane or carborane implantation into a B implantation with adjusted
energy and dose. This underestimates the amorphization by borane or carborane implantation
and ignores the implantation of C. Within analytic implantation, a better alternative is to
replace the borane or carborane implantation directly by a combination of B and C
implantation, with increased values of dfactor.

Sentaurus MC

It is recommended to use Sentaurus MC for molecular implantation. The file

AdvCal_2019.03.fps includes a calibration for carborane implantation, with parameters
integrated in callback procedures to obtain accurate results for the amorphization and good
initial conditions (see Implantation Preprocessing and Postprocessing on page 65). However,
for all other molecular implantations, the following settings must be defined by users in the
case of a calibration:
1. Switch off the hydrogen solution to save CPU time in subsequent anneals:
solution name=Hydrogen nosolve store
2. Adjust amorphization by molecular implantation. If $Energy is the energy of the
molecule, the following parameters can be set as a function of $Energy before the
molecular implantation:
pdbSet Si B surv.rat <f($Energy)>
fproc mcdfactor_<molecule> { energy dose } { return f($energy) }
where surv.rat (‘survival rate’) scales the crystal damage, which is generated by nuclear
collisions during the implantation. In particular for small energies, it must be set to very
high values; otherwise, ion channeling during implantation is overestimated.
mcdfactor_<molecule> will be called after the implantation for a final scaling of the
crystal damage. With AdvCal_2019.03.fps, the value is applied once for each atom
species of the molecule.
After the molecular implantation, you should reset surv.rat for B (and other species in
the molecule) to the default values, as there might be other B ion implantations later in the
process simulation:
pdbSet Si B surv.rat 0.225 ; # Sentaurus Process default
3. Set the number of interstitials to be generated by molecular implantation. When using
AdvCal_2019.03.fps, Sentaurus Process adds automatically one interstitial for each
atom of the molecule. However, this is not reasonable for H, which probably comes to rest
in an interstitial position and is unlikely to kick out Si atoms from the lattice site.

Advanced Calibration for Process Simulation User Guide 135

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

A correction of the ‘automatic’ calculation of excess interstitials can be implemented with

the callback procedure UserMCPostProcess:
proc ifactor_<molecule> { e d } { return 1.0 }
proc UserMCPostProcess { Species Name Energy Dose Tilt Rotation Slice Mode
MCIFac MCVFac } {
if { $Species == "<molecule>" && $Name == "Hydrogen" } {
set ifactor [ifactor_<molecule> $Energy $Dose]
sel z = "Int_Implant - $ifactor * Hydrogen_LastImp"
Silicon name=Int_Implant store }
}

The amorphization of silicon by molecular implantation other than carborane and BF2
implantation has not been calibrated yet.

Fluorine Diffusion and Clustering

The procedure AdvancedFluorineModel defines a physics-based model for fluorine
diffusion and clustering in silicon (see Fluorine Diffusion and Clustering on page 74). The
model has been calibrated to reproduce experimental data for USJ formation based on F, BF2,
and cocktail implantations (Ge+F+B, Ge+C+BF2), and subsequent spike anneals [25].

The main effect of the F coimplantation is a reduction of interstitial supersaturation due to

interstitial trapping by F–I clusters, leading to a reduction of dopant TED and deactivation.
Fluorine is mainly effective in combination with (pre)amorphization and recrystallization. In
recrystallized regions, F is incorporated as Fi or F–I clusters. The dissolution of F–I clusters
leads to free F–I, by reactions that consume free I. The simulation results are relatively
sensitive to the thickness of the amorphization layer generated during implantation. Therefore,
the dfactor of preamorphizing implantations, or self-amorphizing F or BF2 implantations,
are recommended fitting factors.

For additional calibration of the F–I cluster model, you can consider modifying the initial
conditions for F after implantation. By default, it is assumed that, in regions amorphized by ion
implantation and recrystallized by solid phase epitaxy, fluorine is in the interstitial state up to
18 –3
concentrations of 10 cm immediately after recrystallization. Higher concentrations are
clustered as F2I and F3I2. In nonamorphized regions, fluorine is assumed to be mostly in an
interstitial position at the beginning of thermal annealing. This assumption can be changed by
the parameters AcInit, AmInit, FractionCryst, and FractionAmor, for example:
pdbSetDouble Si F AmInit 1.0e20
pdbSetDouble Si F2I1 FractionAmor 0.0
pdbSetDouble Si F3I2 FractionAmor 1.0
pdbSetDouble Si F2I0 FractionAmor 0.0

136 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

With these parameter values, F would be initialized as follows: In amorphized regions, the first
20
F atoms are put into interstitial sites. If the total concentration of F is higher than 10 , the rest
is put into only F3I2 clusters initially. Compared to the default initial conditions, more
interstitials would be present after solid phase epitaxy and, therefore, also more dopant TED.

The TED of dopants can be adjusted further by changing the stability of the F–I clusters, for
example, by lowering the cluster binding energies:
pdbSetDoubleArray Si F Int ClusterFormE {Fluorine 0 F2 -2.3 F2I -3.3 F3I2 -6.2}

This will lead to an earlier decay of F–I clusters and, therefore, a stronger reduction of TED.

Finally, the fluorine-dependent boron dose loss can be disabled by:

solution name=BF nosolve store

Nitrogen Diffusion and Clustering

The procedure AdvancedNitrogenModel defines a physics-based model for nitrogen
diffusion and clustering in silicon (see Nitrogen Diffusion and Clustering on page 75). The
model has been calibrated to reproduce experimental data for USJ formation based on N
[26][27] and cocktail implantations (Ge+N+B, N+As), and subsequent RTA or spike anneals.

The main effect of the nitrogen coimplantation is a reduction of interstitial supersaturation due
to interstitial trapping by N–I clusters, leading to a reduction of dopant TED and deactivation.
Nitrogen is mainly effective in combination with (pre)amorphization and recrystallization. In
recrystallized regions, N is assumed to be incorporated as (Ni)2V clusters that can capture free
interstitials, thereby transforming into mobile (Ni)2 or immobile (Ni)2I clusters.

The calibration included in AdvancedNitrogenModel provides a good starting point, but

often does not provide satisfactory immediate accuracy. Fine-tuning is often needed. In the
following, the most important parameters for the fine-tuning of AdvancedNitrogenModel
are explained.

To increase the capability of nitrogen coimplantation to decrease dopant TED, the most
important method is:
■ Adjust the thickness of the amorphization layer generated during implantation. This can be
modified by tuning the parameter dfactor (for analytic implantation) or the return value
of mcdfactor_Nitrogen (for MC implantation). The thicker the amorphous layer, the
more N is initially incorporated in (Ni)2V clusters, which contribute most to interstitial
trapping.

Advanced Calibration for Process Simulation User Guide 137

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

To decrease the capability of nitrogen coimplantation to reduce TED, the most important
methods are:
■ Define ifactor greater than 0 for nitrogen implantation. (Typical values for nitrogen are
in the range 0–1).
■ Change the initialization of nitrogen in recrystallized regions. Instead of putting 100% of
the nitrogen into (Ni)2V clusters, a significant percentage of nitrogen can be initialized as
(Ni)2 or (Ni)2I by changing the corresponding values of FractionAmor. Apart from
increasing TED, this also will decrease the stability of nitrogen in recrystallized regions.

To fine-tune the nitrogen profile itself, the most important methods are:
■ To achieve faster dissolution of nitrogen clusters, reduce the fraction of N that is initially
put into (Ni)2V or (Ni)2I clusters, using the parameters FractionCryst (for
nonamorphized regions) and FractionAmor (for amorphized regions).
■ To change the stability of (Ni)2I clusters, reduce the binding energy for (Ni)2 + I <=> (Ni)2I.
The default value (parameter NDimer Int BindCluster) is 2.65 eV. Good values are
typically between 2.40 eV and 2.75 eV.

Diffusion in Strained Silicon and Silicon Germanium

Most of the calibration of the effects of Ge concentration and stress on dopant diffusivity have
been performed by using published data based on marker layer experiments [28]. In this type
of experiment, the diffusion of dopants is investigated by annealing experiments and
subsequent SIMS measurements in wafers, which contain (biaxially) strained or relaxed Si and
SiGe layers that are formed by molecular beam epitaxy (MBE). Often, the dopant atoms are
built into the wafers during MBE.

The calibration for SiGe is not expected to be as accurate as the one for silicon, since it has not
been tested to the same extent against data from device manufacturing processes, which
combine ion implantation and rapid thermal annealing, and in which the concentration of point
defects, defect clusters, and dopant–defect clusters are often far from thermal equilibrium. In
particular, for elements other than B, and for high Ge mole fractions in general, little data has
been published that allows for testing the accuracy of models for dopant diffusion and
activation in strained Si and SiGe.

Advanced Calibration does not include any SiGe or stress effects for the implantation damage
calibration. Since implantation into SiGe shows more amorphization than into pure Si [29], it
is recommended to increase slightly the dfactors in the case of SiGe implantation.

In the case of Monte Carlo implantation into SiGe, more accurate results are often achieved for
as-implanted profiles if you assume implantation into pure Si instead of the compound
material.

138 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

To treat SiGe as pure Si, the minimal concentration threshold can be increased to an arbitrary
high value by the following command:
pdbSet Si SiliconGermanium.MCmin 1e30

As-Implanted Dopant Profiles

In general, the dopant distribution after ion implantation is calculated with high accuracy. For
1D structures, similar accuracy is obtained for Monte Carlo ion implantation and for analytic
tables. This is because the tables have been generated by extraction of dual Pearson parameters
from Crystal-TRIM simulations [19].

NOTE Sentaurus Process checks the implantation tables and sets the
implantation mode to beam dose or wafer dose, both for analytic and
MC implantation, depending on the implantation tables format. As a
consequence, beam dose is selected for Ge, C, F, and N, while wafer
dose is selected for the other species. You can use a global switch for the
same dose control for all implantations:
pdbSet ImplantData DoseControl <WaferDose | BeamDose>

The entries of the Default implantation tables, which are used by default, depend on the
species, energy, dose, tilt, and capping-layer thickness. For P, they also depend on the rotation;
whereas, for all other species, the Default tables have been generated with the assumption
rotation=0. For thick capping layers, the majority of the implanted ions come to rest in the
capping layer, and the profile in silicon does not allow an unambiguous extraction of the nine
parameters of double-Pearson functions. This has resulted in some entries in the Default
tables having Pearson parameters with bad values, which in turn lead to bad interpolation
results between neighboring entries.

To overcome this situation, Sentaurus Process has new implantation tables for B, As, P, and Sb,
which are widely identical with the Default tables, but in which the Pearson parameters
extracted for thick capping layers have been replaced by the Pearson parameters extracted for
the next thinner through oxides. This modification eliminates the bad entries from the tables
and ensures a substantially improved interpolation quality for implantation through thick
capping layers, such as tilted pocket implantations through the polysilicon gate.

These new tables are not selected in AdvCal_2019.03.fps to keep results for as-implanted
profiles backward compatible with previous versions of Advanced Calibration. To select these
tables, use the commands:
pdbSetString ImplantData As TableVersion 2010.12
pdbSetString ImplantData B TableVersion 2010.12
pdbSetString ImplantData P TableVersion 2010.12
pdbSetString ImplantData Sb TableVersion 2010.12

Advanced Calibration for Process Simulation User Guide 139

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

The amount of ion channeling must be fine-tuned frequently because it depends strongly on the
dose rate and the wafer temperature during the implantation, as shown by experiments of [30].
Both parameters are usually not specified in the implant command. To fine-tune ion
channeling in MC implantation, use the parameter surv.rat (Sentaurus MC) or dacc
(Crystal-TRIM). To fine-tune ion channeling in analytic implantation, use the parameter
ratio, which specifies the fraction of ions described by the first Pearson function.

For high-energy implantation into (100) wafers with zero tilt, the dopant profile is very
sensitive to a small deviation of the tilt angle from tilt=0. For such conditions, a mismatch
between SIMS and simulation might be due to insufficient experimental precision in the
tilt=0 ion implantation.

If several implantations are performed subsequently without an intermediate diffusion step, the
ion channeling is reduced due to the increasing implantation damage. In Monte Carlo ion
implantation, this is taken into account automatically; in analytic ion implantation, the
CoImplant model is used for calculating the reduction of ion channeling. This model is
described in the Sentaurus™ Process User Guide.

In 2D or 3D applications, the simulation results for Monte Carlo implantation and analytic
implantation are often different, due to differences in the lateral distribution of dopants in
structured geometries. Typically, the Monte Carlo simulation gives a more accurate lateral
distribution of dopants. For CMOS technology, the difference in as-implanted profiles between
Monte Carlo and analytic implantations is most important for the tilted halo implantations. The
advantage of an analytic implantation is that it is much faster than a Monte Carlo implantation
and that the result does not show any statistical noise. The latter is very important if you want
to investigate the effect of small variations of implantation conditions on the device
performance.

You can decide to use Monte Carlo implantation or analytic implantation outside of the
calibration file by using one of the keywords sentaurus.mc and crystaltrim, or not.

Coimplantation Model

The coimplantation model is switched on by default. This model will automatically reduce the
ion channeling in successive ion implantations.

However, the location of the crystal damage is not taken into account by the coimplantation
model. Occasionally, in 2D and 3D process simulations, successive implantations are
performed into different regions of the wafer, because implantation masks are removed and
other masks are deposited between the implantation steps. In this case, the earlier implantation
does not reduce the ion channeling of the latter implantation, and it is recommended to switch
off the coimplantation model for these ion implantations.

140 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

Preamorphization Implantation Model

For analytic simulation of low-energy implantations that follow a preamorphization

implantation (PAI), the PAI model is a superior alternative to the coimplantation model. For
example, this might be the case for Ge+B source–drain implantations. The PAI model is also
described in the Sentaurus™ Process User Guide.

To switch it on for a single boron implantation step, insert the keyword pai in the implant
command.

Cold and Hot Implantation

Cold or cryogenic implantations are implantations with substrates kept below room
temperature. Under these conditions, there is less dynamic annealing and the implantation
amorphizes more rapidly, leading to deeper amorphous layers and less implantation
channeling [31][32][33]. Hot or heated implantations are implantations with substrates kept
above room temperature. Under these conditions, there is more dynamic annealing and the
implantation amorphizes less rapidly, leading to shallower amorphous layers and more
implantation channeling [30][34].

Dynamic annealing is lattice self-repair during the implantation. Implantation creates much
damage, basically I and V, which can cluster into so-called amorphous pockets. Similar to
single I and V, the I and V in amorphous pockets can recombine and, therefore, reduce the
damage. This recombination is hindered by an energy barrier, but it still happens at relatively
low temperatures during the implantation. Different implantation dose rates and temperatures
lead to different balances of damage generation and lattice self-repair during the implantation
and, therefore, different amorphization and channeling. In addition, the increase of channeling
due to the lower vibration amplitude of the Si lattice at lower implantation temperatures is a
much weaker effect than the reduction of channeling due to stronger amorphization.

These effects can all be simulated with Sentaurus Process Kinetic Monte Carlo [35] (see
Amorphization on page 240). However, neither the implantation temperature nor the dose rate
is taken into account in Advanced Calibration for continuum Sentaurus Process. To account for
the temperature dependency of the implantation, the following measures can be taken to
increase the accuracy of amorphization and channeling.

Amorphization

To account for the changed amorphization of cold and hot implantations, you can increase or
decrease the damage scaling factor dfactor. The damage can be calibrated based either on
experimental data (TEM) [35] or on kinetic Monte Carlo [36] reference simulations.

Advanced Calibration for Process Simulation User Guide 141

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

It is possible to define an expression for the dfactor that depends on the implantation
temperature. The implantation temperature specified in the implant command:
implant Carbon dose=1.0e+15 energy=6.0 temperature=-100

can be accessed in the implantation callback procedures and, therefore, in the

<x>factor_$Species procedures by the simGetDouble Diffuse temp statement, for
example by:
proc dfactor_Carbon { Energy Dose } {
set T [simGetDouble Diffuse temp] ;# in degree Celsius
return [expr 2.0*(1.0+tanh(-($T+20.)/85.))]
}

Similarly, expressions for ifactor and vfactor can be introduced.

For Sentaurus MC implantation, it is recommended to adjust the parameter surv.rat as

described in the following section.

Channeling
The analytic implantation profiles are for room temperature implantations in general and might
need to be adjusted for cold and hot implantations. For adjustments of the implantation tables,
you can change the ‘ratio’ of the two Pearsons to fit the channeling:
implant species=Phosphorus Silicon ratio=0.5
implant Phosphorus dose=1e+14 energy=30.0 tilt=0.0 rot=0.0
ResetImplantParams Silicon Phosphorus

For Sentaurus MC implantations, the implantation channeling can be adjusted by changing the
dynamic annealing factor, that is, by increasing for cold implantations [35] or decreasing for
hot implantations [37] the parameter for the survival rate of the Frenkel pairs for the
noncascade damage model:
pdbSet Si <species> surv.rat <value>

This not only increases (or decreases) the damage during implantation to influence the
channeling, but also leads to more (or less) amorphization at the end of the implantation.

142 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

CTotal (−100°C)
1021 ITotal (–100°C)
CTotal (RT)
ITotal (RT)

1020
Concentration [cm–3]

1019

1018

1017
0 50 100 150
Depth [nm]

Figure 32 Total chemical carbon profiles (CTotal) for room temperature (RT) and cryogenic
(–100oC) carbon implantation (1015 cm–2 at 6 keV) by Sentaurus MC in good
agreement with SIMS provided by AMAT-VSE1 (not depicted). The RT
implantation shows more channeling than the –100oC one due to an increased
surv.rat parameter value for the cold implantation. At the same time, the cold
implantation generates an amorphous layer up to a depth of 27 nm indicated by
the initial condition of the total interstitial profile (ITotal), while the RT implantation
amorphizes only marginally.

Dose Loss at Silicon–Oxide Interfaces

For As, B, and P, the ThreePhaseSegregation model is used to describe dose loss at
Si–SiO2 interfaces. It takes into account that dopants can be trapped at the interface layer. In
thermal equilibrium, the concentration of boron at both sides of the interface, which can be
named as the segregation coefficient, can be expressed as:
BOx / BSi = p/ni * TrappingRateSi / TrappingRateOx * EmissionRateOx /
EmissionRateSi * 5e22 / SolubilitySi

Here, p/ni is the normalized hole concentration at the silicon side of the interface. SolubilitySi
is the solid solubility of B in silicon, and the other parameters are specified in the sections on
dose loss in AdvCal_2019.03.fps. In comparison, if the simple Segregation model is
used, the same relation can be expressed as:
BOx / BSi = p/ni / Segregation

1. Applied Materials - Varian Semiconductor Equipment.

Advanced Calibration for Process Simulation User Guide 143

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

Here, Segregation is the parameter of the Segregation model. In some limiting cases in
which the amount of dopants trapped at the interface can be neglected, you can achieve similar
results with the Segregation model and the ThreePhaseSegregation model. For this, the
parameter Segregation must be defined as:
Segregation = TrappingRateOx / TrappingRateSi
x EmissionRateSi / EmissionRateOx
x SolubilitySi / 5e22 (13)

For As, B, and P, the Advanced Calibration parameters used in the

ThreePhaseSegregation model correspond to the following values of Segregation in
the Segregation model:
pdbSet Ox_Si As Segregation {[Arr 97777 0.6]}
pdbSet Ox_Si B Segregation {[ArrBreak 30349.63 1.302 76.455 0.6455 1000]}
pdbSet Ox_Si P Segregation {[Arr 46984 0.32]}

When switching from the ThreePhaseSegregation model to the simpler Segregation

model, consider defining the above values for Segregation, instead of the default values.

Note that, in both the ThreePhaseSegregation model and the Segregation model,
segregation into oxide is enhanced for high dopant concentrations by the factor p/ni (Poni) for
acceptors and n/ni (Noni) for donors, in agreement with experimental evidence.

Calibration of ThreePhaseSegregation Model

To increase dose loss, the easiest way is to increase the value of the parameter CMax or
TrappingRate_Silicon.

During calibration, it might be necessary to change dose loss in a different way for high and
low surface concentrations of a dopant. This can be achieved in two ways:
■ Decreasing CMax and at the same time increasing TrappingRate_Silicon can have an
opposite effect for high and low surface concentration. For high surface concentration,
most interface traps can be filled. Reducing CMax will, therefore, reduce dose loss even for
very high values of TrappingRate_Silicon. For low surface concentration, reducing
CMax might be overcompensated by an increase of TrappingRate_Silicon.
■ To increase dose loss especially for high concentrations of a dopant, you can consider to
implement a pairing reaction for dopants in the interface layer, as it is done for P in
AdvCal_2019.03.fps.

Sometimes, it might be necessary to adjust dose loss in a different way for inert and oxidizing
atmospheres. During oxidation, the interface moves. The dopants do not have much time to
cross the interface to establish segregation equilibrium between dopants in oxide and dopants
in silicon. For this reason, for oxidation, the dose loss is more sensitive to the rate at which

144 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

atoms change places. Changing all trapping rates and emission rates for a dopant by the same
factor will have a much greater effect on dose loss for oxidizing atmospheres than for inert
atmospheres.

A flexible instrument to modify dose loss is defining the terms CMaxFactor and
Side.SS.Factor. CMax is multiplied by the first term, if defined, and the emission rate from
the interface layer into silicon is multiplied by the second term, if defined. For example:
term name=PhosphorusCMaxFactor add Oxide /Silicon \
eqn = "2.0e14/(Arsenic+2.0e14)" store
term name=BoronCMaxFactor add Oxide /Silicon \
eqn = "(y<@<lgate/2.0>@) ? 0.2 : 1"

Side.SS.Factor is used in SiGe_and_Stress_Effect to take into account the

modification of emission rates into strained silicon due to the impact of pressure on solid
solubility. In CMOS process simulation, an important use case of CMaxFactor is to define a
different dose loss into the gate oxide (which is typically oxynitride) and into the oxide spacer.

Oxidation
Parameters for the silicon oxidation rates have not changed in AdvCal_2019.03.fps, except
for the Massoud model parameters for wet oxidation. Some comments might be useful for
additional fine-tuning.

The growth of thick oxides (thickness > 30 nm) is governed by the linear rate constant (B/A)
in the Deal–Grove model. For very thick oxides (thickness > 100 nm), the parabolic rate
constant (A) of the Deal–Grove model also becomes important. The default values of these
parameters can be considered to be well calibrated in Sentaurus Process for dry, wet, and mixed
atmospheres.

For thin oxides (thickness < 5 nm), the growth rate is dominated by the Massoud correction
term of the Deal–Grove model. This correction term is taken from Sentaurus Process defaults
for dry oxidation and has been calibrated for wet oxidation. Recommendations for the fine-
tuning of the Massoud correction term for wet oxidation are presented in Massoud Model
Parameters for Wet Oxidation of Silicon on page 36.

For very thin oxides and oxynitrides (thickness < 2 nm), which are often grown at low
temperatures, the models of Sentaurus Process are not sufficiently well calibrated to expect
precise results. For CMOS gate oxides (oxynitrides), it is recommended to extract the oxide
thickness from a comparison of simulated and measured C–V characteristics in accumulation,
and to adjust the oxide thickness in process simulations by defining the thickness of the initial
oxide layer.

Advanced Calibration for Process Simulation User Guide 145

P-2019.03
3: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process

Bird’s Beak in CMOS Devices

CMOS devices sometimes exhibit a so-called bird’s beak. During poly reoxidation, the gate
oxide grows slightly thicker at the gate edge than in the middle of the transistor. For the
calibration of the bird’s beak, in principle, you can use high-resolution TEM pictures of the
gate oxide, but often the bird’s beak is too small to be visible in TEM. For thin gate oxides, the
growth of the bird’s beak is dominated by the Massoud correction term of the Deal–Grove
model. This correction term has been calibrated for the growth of planar oxides, and it possibly
needs to be modified for the oxide geometry at the gate edge.

The most radical modification is to completely suppress oxidant diffusion in the gate oxide
during poly reoxidation. This will completely eliminate the formation of a bird’s beak shape.
An advantage of this approach is that the Si–SiO2 interface remains perfectly parallel to the
horizontal mesh lines, which is beneficial for the convergence of device simulation.

To completely suppress oxidant diffusion in the gate oxide, define the term O2DiffFactor
(H2ODiffFactor):
term name=O2DiffFactor add Oxide store \
eqn="((y<($lgate/2.0)) * (x>-0.01)) ? 0 : 1"

In this definition, it is assumed that the gate oxide is located in the region y<($lgate/2.0)
and x>-0.01; whereas, other oxide is located outside this region.

Besides its impact on the final shape of the gate oxide, altering the bird’s beak has an impact
on the OED during poly reoxidation and, thereby, on the NMOS reverse short-channel effect.

Diffusion in Polysilicon and Out-Diffusion From

Polysilicon
The dopant diffusion in polysilicon and the out-diffusion of dopants from polysilicon have not
been calibrated. In CMOS simulations, it is often sufficient to assume a constant doping profile
in polysilicon, which can be adjusted manually with the help of C–V measurements.

For devices where the out-diffusion from polysilicon into silicon is important (such as in
processes for the fabrication of bipolar devices), you need to select the model and parameters
for the dopant flux at the interface.

146 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Performing Additional Calibration

Dopant Penetration Through Gate Oxide

The dopant penetration through gate oxide has not been calibrated. This is a difficult task
because, instead of pure oxide, oxynitride is commonly used as the gate dielectric material.
Dopant diffusion from polysilicon through the gate oxide into the channel region can be
completely suppressed by the user by setting the transfer coefficient at the polysilicon–oxide
interface to 0.

Diffusion and Activation in Germanium

Sentaurus Process supports diffusion and activation in crystalline germanium with a Ge–GeO2
interface. A basic calibration has been performed, and parameters for material properties,
defect evolution, and dopant diffusion and activation have been introduced to the Sentaurus
Process parameter database and Advanced Calibration.

In general, the calibration for germanium is less accurate than the one for silicon, since it is less
mature and based on a smaller experimental dataset. Therefore, additional calibration might be
required.

Defect recombination and dopant segregation are calibrated only for the GeO2–Ge
(GeOxide_Germanium) interface. GeO2 is thermally unstable and water soluble, and
uncapped GeO2 surfaces lead to germanium substrate loss even in low-temperature processes
[38]. Different capping layers (SiO2, Si3N4) on the germanium surface can prevent substrate
loss and lead to a different phosphorus dose loss [38]. However, the effect of different capping
layers has not been calibrated, since surface conditions are not always specified in detail in the
literature. Therefore, the dose loss might need additional calibration for specific cases. Since it
is assumed that, even in the case of a SiO2 capping on Ge, a small layer of GeO2 is formed in
between, specifying a Ge–GeO2 boundary for Ge substrate simulations is reasonable. However,
for usability and SiGe parameter interpolation, the parameters for the Ge–SiO2 interface are
defined in Advanced Calibration by inheriting the parameter values from Ge–GeO2.

NOTE Sentaurus MC is recommended for implantation into germanium, since

no implantation tables for germanium are available.

Performing Additional Calibration

For the process simulation of any new technology or technology node, it is recommended to
use the Advanced Calibration as a starting point.

Advanced Calibration for Process Simulation User Guide 147

P-2019.03
3: Guidelines for Additional Calibration
Performing Additional Calibration

Calibration Procedure
In many cases, Advanced Calibration gives accurate results without requiring additional
parameter fine-tuning. However, in most cases, especially for the development of innovative
technology, additional fine-tuning is needed for a customized TCAD calibration.

The recommended way to perform the calibration of parameters for 2D (or 3D) process
simulation is:
1. Compare SIMS data from your process flow with straightforward results obtained with
Advanced Calibration. Check whether the agreement is good and whether additional
parameter fine-tuning is needed to improve the agreement.
2. If additional models, or parameter fine-tuning, are needed, put the required additional
calibration together in a single user calibration file, which will be sourced immediately
after loading the default Advanced Calibration. For the remainder of this section, this user
calibration file will be called user_calib.fps.
3. After a good agreement with SIMS data is obtained, make a setup for 2D process and
device simulation. Before continuing with the calibration, ensure that the setup is good in
terms of numerics (small CPU time, small numeric noise, and robustness).
4. Perform 2D (or 3D) process and device simulations, and compare the simulated and
measured electrical device characteristics. Analyze whether there is insufficient accuracy,
which is due to process simulation calibration.
5. Perform additional parameter fine-tuning if required. As in the case of 1D simulations, all
fine-tuning in addition to the default Advanced Calibration should be put together in the
user_calib.fps file, which is sourced immediately after the following command:
AdvancedCalibration 2019.03
6. As a result of the fine-tuning (calibration), the same file user_calib.fps should give
accurate results in all 1D and 2D (or 3D) simulations.

It is reasonable to start with 1D simulations and SIMS data, because 1D simulations are much
faster and the SIMS data provides direct information on the dopant distribution, whereas
electrical device data does not always allow you to separate clearly the many effects that
accumulate in a full 2D or 3D process flow.

Synopsys offers calibration service projects. In such projects, the optimized and calibrated
input files for process and device simulation are created by expert application engineers at
Synopsys, and the calibration steps are explained in detail to customers.

148 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Performing Additional Calibration

Loading a User Calibration File

An input file of a Sentaurus Process simulation with additional user calibration should have the
following structure:
# Title
AdvancedCalibration 2019.03 ; # loads the Advanced Calibration
source ./user_calib.fps ; # loads the additional calibration
line ... ; # rest of process simulation file
init ... ; # without any model parameters
implant ...
diffuse ...
deposit ...
save ...
exit

All project-specific calibration is contained in a separate file user_calib.fps, which is

sourced after loading the Advanced Calibration defaults. This setup has several advantages:
■ The simulation input file is easier to read if it is free of physical models and parameters.
Changing the process conditions does not require a high level of TCAD expertise.
■ All project-specific calibration is contained in a single file. Often, this file is short and
contains only approximately ten changes with respect to the default Advanced Calibration.
Ideally, it can be printed on a single page. Therefore, it is very easy to see what has been
performed in the user calibration. The calibration work becomes clearer and easier.
Furthermore, the user calibration from different technologies or technology nodes can be
compared and exchanged conveniently.

See Example of a User Calibration File on page 150.

Recommendations

Miscellaneous

This section gives some recommendations on how to avoid common errors in 2D simulations:
■ In process recipes that contain tilted implantations, the implantation dose is defined either
per wafer area or per beam area. To ensure that Sentaurus Process uses the required
definition for all species, it is strongly recommended to set the DoseControl parameter
at the start of the command file:
pdbSet ImplantData DoseControl <WaferDose | BeamDose>

Advanced Calibration for Process Simulation User Guide 149

P-2019.03
3: Guidelines for Additional Calibration
Performing Additional Calibration

■ A few nanometers of silicon are sometimes removed in cleaning steps. This might be
important to take into account after low-energy ion implantation.
■ Always assume that bare Si surfaces are covered by a natural oxide of 1–2 nm thickness.
In practice, this can mean that it is necessary to add deposition steps of thin oxide layers in
the process simulation. The dose loss model used in Advanced Calibration works only for
Si–SiO2 interfaces not Si–gas interfaces.
■ In CMOS technology, the gate is not always rectangular. If possible, adjust the gate shape
to a TEM picture, to obtain the correct as-implanted extension and halo dopant profiles.
■ For spike annealing and, more importantly, for laser and flash lamp annealing, the heat
cycle to be used in the diffuse statement ideally must be taken from a corresponding
temperature profile measurement.

Lateral Diffusion Along Interface

The three-phase segregation model allows you to specify a dopant diffusivity in the interface
layer. This model is used for B, As, and P in the Advanced Calibration. The interface
diffusivities are set to 0 by using:
pdbSetDouble Ox_Si B D { 0 0 }

By setting positive values of the interface diffusivity, the lateral diffusion of dopants near the
Si–SiO2 interface can be increased without affecting the vertical diffusion far from the
interface. This can be used, for example, to adjust the short-channel effect in deep submicron
MOSFETs. More importantly, setting a positive value for the lateral diffusivity will also change
the shape of the p-n junctions close to the interface [39].

Example of a User Calibration File

The following example is a typical user calibration file, which can be the result of a CMOS
calibration project, for example. The file user_calib.fps is sourced after loading the
default Advanced Calibration file:
# Calibration file for 90nm CMOS technology.
# Six changes after loading AdvCal_2019.03.fps.
# 1) Use 2010.12 implant tables
pdbSetString ImplantData As TableVersion 2010.12
pdbSetString ImplantData B TableVersion 2010.12
pdbSetString ImplantData P TableVersion 2010.12
# 2) Adjust ion channeling and rp for high energy P implant
proc UserImpPreProcess { Species Energy Dose Tilt Rotation Slice Mode } {
ResetImplantParams Silicon Phosphorus ; # reset to default
if { $Species == "Phosphorus" && $Energy > 300 } {

150 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Performing Additional Calibration

implant species=P Si ratio=0.9

implant species=P Si rp=[expr 1.15e-3*exp(log($Energy)*0.99)]
}
}
# 3) For nMOS extension (energy and dose dependence of TED)
proc ifactor_Arsenic { Energy Dose } {
if { $Energy < 5.0 } {
return [expr 2.0e14/($Dose+5e13) + (5-$Energy)/4.0]
}
return [expr 2.0e14/($Dose+5e13)]
}
# 4) For nMOS extension (dose loss)
pdbSetDouble Ox_Si Arsenic CMax {[Arr 5.0e17 0.8]}
# 5) For pMOS extension (energy and dose dependence of TED)
proc ifactor_Boron { Energy Dose } {
if { $Energy < 2.5 } { return 0.5 } ; # pMOS extension implant
return 1.0 ; # nMOS well implant
}
# 6) For pMOS extension (reduced diffusion after spacer deposition)
pdbSetDouble Si B CluRate {[Arr 5e11 5.42]}

This example calibration file, which is a hypothetical result of a CMOS calibration, can be
embedded in a CMOS process simulation input file, as described in Loading a User Calibration
File on page 149. Its contents are discussed briefly.

This file changes six settings of the default calibration file. The first two changes improve the
accuracy of as-implanted profiles. The next two changes target As extension profiles after
annealing, and the last two changes relate to the B extension diffusion in the PMOS.

The implementation of UserImpPreProcess (second change) illustrates how to adjust the

Pearson parameters. You must pay attention to the following aspects:
■ If the modified values of the Pearson parameters should be applied only for some process
conditions (for example, only for high-energy P implantation, but not for low-energy P
implantation), you must ensure that the parameter changes are not permanent for all
subsequent implantations. In this example, this is performed by including the command
ResetImplantParams.
■ When adjusting one of the Pearson parameters (rp, stdev, rp2, stdev2), you should
avoid defining a fixed value. Instead, you can define them as functions of implantation
parameters. In this example, rp is defined as a function of implantation energy in order not
to destroy the energy dependency of rp by the calibration.
With a fixed value, the calibration is valid only for a single condition, but it is not predictive
for a larger process window of interest. Only for the parameters ratio, gamma, beta,
gamma2, and beta2 can you use a fixed value in a larger window of interest.

Advanced Calibration for Process Simulation User Guide 151

P-2019.03
3: Guidelines for Additional Calibration
Recommended Numeric Settings for Monte Carlo Implantation

A permanent change of diffusion parameters is performed with the pdbSet, pdbSetDouble,

or pbdSetDoubleArray command. The initial conditions after ion implantation can also be
calibrated in the user calibration file. For example, as demonstrated in the file above, the
interstitial plus-factor (ifactor), which scales the amount of interstitials generated per
implanted ion, can be adjusted as a function of implantation energy and dose. The default
procedures of the Advanced Calibration for setting ifactor are implemented in section 3.2
of AdvCal_2019.03.fps.

By redefining the procedures ifactor_Arsenic and ifactor_Boron, you can overwrite

the original dependency on implantation energy and dose. Using initial conditions that depend
on implantation energy and dose is a very powerful method to calibrate dopant diffusion in the
energy and dose window of interest for a given technology.

Recommended Numeric Settings for Monte Carlo

Implantation
This section presents recommendations that allow for the reduction of numeric noise and CPU
time for Monte Carlo implantations.

Monte Carlo Pocket Implantation

In process simulation of CMOS devices for 65 nm or smaller gate lengths, Monte Carlo (MC)
simulation is often preferred to analytic implantation models for the pocket implantation,
because it allows for the calculation of the 2D (or 3D) dopant distribution under the gate corner
with higher accuracy.

Recommendations are given to minimize the numeric noise and the CPU time needed for MC
pocket implantations. These recommendations are given for 2D process simulation, but can be
generalized for the 3D case:
■ In 2D CMOS process simulation with a single symmetric gate, MC implantations can be
performed on a half-structure, in which the half-gate is typically at the left side of the
simulation domain. The following choices can be selected for the boundary conditions at
the left-side and right-side simulation domain boundaries:
• pdbSet MCImplant LeftBoundary Reflect
This leads to the fastest simulations. Atoms that reach the left-side boundary are
reflected. A small disadvantage is that ions, which travel in crystal channels before
reflection, are scattered out of their crystal channeling, which might lead to a small
maximum of concentration near the symmetry plane. This disadvantage can be avoided
by using TrueReflect instead of Reflect. For details, refer to the Sentaurus™
Process User Guide.

152 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Recommended Numeric Settings for Monte Carlo Implantation

• pdbSet MCImplant RightBoundary Reflect

This leads to the fastest simulations. It is not perfectly precise. In particular, at the right
side of the device, for implantations with rotation=90, it will result in a decay of the
profile. The alternative is:
pdbSet MCImplant RightBoundary Extend
pdbSet MCImplant MinExtend 0.03 ;# example
pdbSet MCImplant ExtensionLength 0.00
With this choice, the simulation domain is extended artificially by 30 nm (the best
value depends on the implantation condition) at the right side to avoid the decay of the
profile.
■ In 2D CMOS process simulation with multiple symmetric gates in periodic arrangement,
the best choice is to use TrueReflect boundary conditions on both sides of the half-
structure:
pdbSet MCImplant LeftBoundary TrueReflect
pdbSet MCImplant RightBoundary TrueReflect
With these settings, the simulation is fastest, and the shadowing of tilted implantations by
the neighboring gate is taken into account correctly.
In comparison, using Reflect boundaries at both sides does not take into account the
shadowing effect of the neighboring gate. (This is because, with Reflect boundaries, ions
are not reflected as long as they travel in the gas phase.)
■ The following switches should be used:
pdbSet MCImplant TrajectoryReplication 0
pdbSet MCImplant TrajectorySplitting 0
They switch off trajectory replication and the trajectory-splitting algorithm. Trajectory
replication reduces CPU time, but it carries the risk that a family of trajectories is not found
in the initial search of possible particle pathways. For CMOS pocket implantations, it is
better to switch it off, as the benefit in CPU time does not compensate for the numeric noise
corresponding to that risk.
Trajectory splitting is an algorithm developed to improve the statistics of as-implanted
profiles in regions that are reached only by a few pseudoparticles. However, in CMOS
pocket implantation, the focus of interest is the region of maximum as-implanted dopant
concentration, and not so much the profile tails. So improving the statistics in the profile
tail (at the cost of using less CPU time for the profile maxima) is not beneficial for CMOS
pocket implantations.
■ Reduce the interval width from 50 nm to a smaller value, for example, 2 nm:
pdbSet MCImplant Intervals dy 0.002
This increases the CPU time because for each interval a number of pseudoparticles
(particles) will be followed. However, it is better to use a larger number of small
intervals with a small particle number each, than to use a small number of very large

Advanced Calibration for Process Simulation User Guide 153

P-2019.03
3: Guidelines for Additional Calibration
Recommended Numeric Settings for Monte Carlo Implantation

intervals with a large particle number, because as a matter of statistics, in the first case, the
incident pseudoparticles are more evenly distributed over the whole simulation domain.
■ Instead of using the same particle number for all rotation angles, use a higher particle
number for implantations towards the gate corner (usually, rotation=90) and a reduced
particle number for implantations away from the gate corner (usually, rotation=270).
For example, a good choice for a four-rotation pocket implantation, when simulating the
right half-structure of a MOSFET, is:
• rotation=0: particles=1000

• rotation=90: particles=4000

• rotation=180: particles=1000

• rotation=270: particles=300

NOTE If the TrueReflect boundary condition is used at the left-side

boundary, you should use the same particle number for rotation=90
and rotation=270 to ensure good statistics in the middle of the
device.

MC Source/Drain Implantation
In CMOS technology, source/drain implantations are typically performed at tilt angles close to
zero. By colliding with Si lattice atoms, a small percentage of implanted ions is deflected
towards the middle of the channel, which can lead to a significant reduction of the threshold
voltage. The accurate modeling of this requires good statistics of the ions deflected towards the
middle of the channel. This can be achieved by the following recommended settings for MC
source/drain implantation:
pdbSet MCImplant TrajectorySplitting 1
pdbSetSwitch MCImplant SplitModel traj.density

With SplitModel traj.density, a particle is replaced by two child particles, with half the
statistical weight each, when it reaches an element with a small trajectory density. This is much
more suitable for obtaining good lateral statistics in the channel region under the gate than the
default SplitModel depth.sp, where particle splitting depends on the depth (distance from
the impact point) only.

Using SplitModel traj.density increases the CPU time, but this is more than
compensated by the improved statistics of the ions implanted in the channel region. Compared
to SplitModel depth.sp, often a much smaller number of implanted particles suffices to
achieve an equally small statistical noise in device characteristics due to source/drain MC
implantation.

154 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Additional Calibration for Power Technologies

Additional Calibration for Power Technologies

The process simulation of power technologies and smart-power technologies can require too
much CPU time because of the large number of thermal anneals. Part 5: Accelerating
Simulations for Power Technologies on page 89 presents a recommended choice of model
switches and numeric parameters to accelerate process simulations.

Calibration
Starting from the Advanced Calibration parameter set, a typical comprehensive calibration of
process simulation for power and BCD technologies consists of:
■ The calibration of oxide thicknesses and shapes for strong oxidation steps.
For LOCOS simulation, the model for stress-dependent oxidation is recommended to be
switched on:
pdbSet Oxide O2 SDO 1
pdbSet Oxide H2O SDO 1
pdbSet Oxide Oxidant SDO 1
The shape of the LOCOS is usually fine-tuned against TEM reference data by calibrating
the viscosity and stiffness (for example, BulkModulus) parameters for oxide and nitride,
and calibrating the activation volume (parameters Vd and Vk) for stress-dependent
oxidation rates.
■ The adjustment of dopant segregation at silicon–oxide interfaces to obtain accurate
(overlap) capacitance versus gate voltage characteristics and threshold voltages. This is
especially true for laterally diffused channel profiles where segregation, lateral, and
vertical diffusion in oxide and silicon are strongly coupled:
• Dopant and defect profiles after ion implantation: The parameter ratio, one of the
double-Pearson parameters for analytic implantation, is used to scale the fraction of
dopants in the first Pearson distribution.
• The point-defect scaling factor ifactor is used to tune transient-enhanced diffusion.

Meshing
A well-designed mesh is important for 2D or 3D process simulations. The mesh should be fine
enough in the critical regions to keep discretization errors small, but it should have a small total
number of mesh points. Mesh refinement criteria are described in the Sentaurus™ Process
User Guide.

Advanced Calibration for Process Simulation User Guide 155

P-2019.03
3: Guidelines for Additional Calibration
Additional Calibration for Power Technologies

For many applications, a good mesh can be obtained using adaptive meshing. Many adaptive
refinement criteria are available to deal with different fields and situations. All functions
involve some comparison between values on neighboring nodes and possible values between
neighboring nodes.

Recommendations for Bulk Refinement With Adaptive Meshing

In this section, two frequently used refinement criteria for adaptive meshing are presented.
These are the relative difference error and the local dose error that correspond to the
rel.error and max.dose.error arguments of the refinebox command of Sentaurus
Process (see Sentaurus™ Process User Guide, Adaptive Refinement Criteria on page 756). It
is advisable to define one refinement box per diffusion species. By doing so, you can choose
refinement box parameters for a certain species, independently from the other species based on
the implantation conditions pertaining to that particular species.

It is possible to demonstrate that, in the case of a simple Gaussian doping distribution, a simple
yet effective algorithm can be formulated to facilitate the choice of parameters for adaptive
meshing of refinement boxes. To bypass complex technical issues, only the algorithm itself is
discussed here, not the full theoretical foundation of the algorithm.

To apply the algorithm to a process flow, you review the process flow and copy all the
implantation steps pertaining to a particular implantation species into a list. After the list has
been obtained, you then find the implantation with the lowest dose (Dmin) and compute the
value of max.dose.error of the refinebox command as follows:
max.dose.error = Dmin / 100.0 (14)

As a next step, rel.error must be assigned a value that is typically between 0.5 and 1.0. The
density of the mesh is sensitive to rel.error because it represents the target relative change
of the field across an edge. For many typical situations, a value of 1.0 gives a coarse mesh, and
a value of approximately 0.5 gives a fine mesh. The exact value of rel.error depends on the
sensitivity of the simulated device characteristics to the process mesh variations and, therefore,
the value must be found experimentally.

Then, the implantation with the smallest energy (Emin) is found, and the corresponding
standard deviation (stdev.min) is obtained as:
set moments [implant material=Silicon $Species \
energy= $Emin dose= $Dose get.moments]
set stdev.min [lindex $moments 5]

156 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Additional Calibration for Power Technologies

The smallest implantation standard deviation value determines the smallest allowed mesh
spacing that can be set in the refinebox command. The value of refine.min.edge is
calculated as follows:
refine.min.edge = stdev.min * rel.error / 4.0 (15)

Consider a process flow containing the following boron implantations:

implant boron dose= 1e13 energy= 400
implant boron dose= 5e12 energy= 800
implant boron dose= 1e14 energy= 40
implant boron dose= 2e15 energy= 10

Now, the parameters of an adaptive refinement box are obtained according to the proposed
algorithm. The second implantation has the smallest dose, so max.dose.error can be set to
5e10. The last implantation has the smallest standard deviation of 24 nm and, if rel.error
equals 1.0, that gives the smallest allowed mesh spacing of 6 nm. Finally, the corresponding
refinement box for boron would be:
refinebox name= Global_Boron adaptive Silicon \
refine.min.edge= "0.006 0.006 0.006" refine.max.edge= "10 10 10" \
max.dose.error= "Boron= 5e10" rel.error= "Boron= 1.0" \
abs.error= "Boron= 1e15"

where abs.error is a cut-off concentration level for the relative difference error refinement
criterion.

When adaptive meshing is switched on, the default adaptive meshing parameters are applied.
However, there are two issues with keeping the default refinement box. First, it applies one and
the same parameter set for all species and this contradicts best practice. Second, it allows mesh
refinement on implantation damage even when no damage sensitive model is used.

To disable the default adaptive refinement box, the following settings must be used along with
user-defined refinement boxes:
pdbSet Grid AdaptiveField Refine.Abs.Error 1e30
pdbSet Grid AdaptiveField Refine.Rel.Error 1e30
pdbSet Grid AdaptiveField Refine.Target.Length 1e30

To accurately capture the amorphous–crystalline interface position, a damage refinement box

must be placed. Note that this is not necessary in the first part of process simulations with
AdvancedPowerDeviceMode, where models sensitive to implantation damage are switched
off. A typical damage refinement box definition is:
set targetLength 2e-4
set minValue [pdbGet Silicon AmorpDensity]
set maxValue [expr 1.001 * $minValue]

Advanced Calibration for Process Simulation User Guide 157

P-2019.03
3: Guidelines for Additional Calibration
Calibration of Wet Etching Rate Modification by Ion Implantation

refinebox add name= PowerDeviceMesh_Damage adaptive Silicon \

refine.fields= {Damage} refine.type= interval \
target.length= "$targetLength" target.length.scaling= "1.0" \
min.value= "Damage= $minValue" \
max.value= "Damage= $maxValue"

In the above examples the refinement boxes are applied globally to the entire simulation
domain. However, it might be necessary to place some additional refinement boxes in the
critical device areas. If that is the case, you should follow the layout (mask)-driven refinement
approach (see Sentaurus™ Process User Guide, Mask-Driven Meshing on page 890).

Refinement at Material Interfaces

For accurate modeling of dose loss at oxide–silicon interfaces, it is important to use a good
mesh refinement perpendicular to the interface. At the oxide side of the interface, the mesh-
point distance must not exceed the diffusion length of boron inside the oxide, which
corresponds to the thermal budget used. Otherwise, the simulation result for boron dose loss
(this is the dose of boron segregation from silicon into oxide) might strongly depend on the
mesh.

Calibration of Wet Etching Rate Modification by Ion

Implantation
Sentaurus Process allows you to scale the etching rate with a user-defined field. You specify
the field using the etch.rate.modifier argument of the etch command. This feature is
available only for 3D process simulations using the level set method.

In device fabrication, ion implantation can be used to modify the wet etching rate in surface
regions affected by implantation. Modifying the wet etching rate for SiO2 and SiN in diluted
hydrogen fluoride (HF) has been studied using data provided by Applied Materials and
published data [40][41].

This section summarizes key findings and recommendations from this calibration effort. The
corresponding parameter settings are not yet included in the Advanced Calibration file.

158 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
Calibration of Wet Etching Rate Modification by Ion Implantation

Calibration
The wet etching rate ratio (werr) is the ratio of the wet etching rate in implanted regions versus
not implanted regions. Ion implantations affect the werr of SiN and SiO2 by the following
mechanisms:
■ The etching rate increases with implantation damage as follows:
• For low doses, the etching rate increases linearly with the concentration of vacancies
generated by implantation. Using Sentaurus Process, you can calculate this vacancy
density with a full-cascade Monte Carlo implantation simulation. Alternatively, in
good approximation, the etching rate also increases linearly with the implantation
damage (Damage), which you can calculate with a regular Monte Carlo implantation
simulation.
• For medium and high doses, the impact of damage on the etching rate saturates towards
a maximum damage-driven increase, which is a factor of approximately 2.5–6.0.
■ For very high implantation doses, there is an additional chemical effect. The etching rate
changes because implanted ions change the chemical composition of oxide or nitride. After
thermal annealing, typically, there is no more damage effect on werr, but the chemical
effect might be stable.

For the datasets studied, good agreement with measured werr has been achieved with the
following formula. For the damage effect:
Damage 1
werr Damage = 1.0 + ( F – 1.0 ) ⋅ ------------------------------------- ⋅ -------------------------------------------------------------------------
- (16)
Damage reference  α 1⁄α
1.0 +  ------------------------------------- 
Damage
  Damage reference 

where:
■ F is the saturation value for very high implantation Damage.
■ Damage is the implantation damage field as calculated with Sentaurus MC.
■ α is a parameter that determines the sharpness of the transition between the low damage
range at which werr increases linearly and the high damage range at which werr is close to
its saturation value F .

A good fit to measured werr data for many implanted species, provided by Applied Materials
has been obtained with the following calibrated coefficients:
21 –3
■ Oxide: F = 2.70 , Damage reference = 5.2 ⋅ 10 cm , α = 1.3
21 –3
■ Nitride: F = 2.45 , Damage reference = 8.0 ⋅ 10 cm , α = 1.5

For data from other sources, different values might be needed to obtain an excellent match. In
particular, [40] and [41] report about two times higher saturation values F compared to these
values.

Advanced Calibration for Process Simulation User Guide 159

P-2019.03
3: Guidelines for Additional Calibration
Calibration of Wet Etching Rate Modification by Ion Implantation

The reason for the difference between experimental findings from different sources is not
known.

In addition to the damage effect, which dominates at low and medium implantation doses, the
following formula has been applied to selected cases, to model the chemical effect on the wet
etching rate:
C reference + β ⋅ C species
factor chemical = ----------------------------------------------------- (17)
C reference + γ ⋅ C species

Here, β and γ are coefficients. With Eq. 17, for small concentrations ( C species « C reference ),
factor chemical equals 1.0. For huge concentrations ( C species » C reference ), factor chemical equals β / γ .
Some implantation species (for example, F and P) increase the etching rate; whereas, other
species (such as B, C, and Si) decrease it at high implantation doses ( factor chemical < 1.0). This
chemical effect is specific for each species and has not been calibrated yet for all implantation
species and concentrations. Therefore, it is the responsibility of users to find good parameters
for Eq. 17 or a more suitable formula for modeling the chemical effect on werr for a particular
use case.

Application
To calculate implantation damage in oxide and nitride, use the following statements:
pdbSetBoolean Oxide Damage 1
pdbSetBoolean Nitride Damage 1

Ion implantation must be simulated with Sentaurus MC. In addition, you must use the latest
version of Advanced Calibration, that is:
AdvancedCalibration 2019.03

NOTE Older versions of Advanced Calibration do not support the calculation

of Damage in amorphous materials such as oxide and nitride.

It is recommended to set the parameter MCDFactor to 1.0 for all implantation species in oxide
and nitride:
pdbSetDouble Oxide Germanium MCDFactor 1.0
pdbSetDouble Nitride Germanium MCDFactor 1.0
pdbSetDouble Oxide Boron MCDFactor 1.0
pdbSetDouble Nitride Boron MCDFactor 1.0
...

160 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
References

Without these statements, the Advanced Calibration functions mcdfactor_<species>,

which are calibrated for implantation damage in silicon, are applied automatically to oxide and
nitride, but it is better if you use 1.0.

Similarly, for analytic implantation, it is recommended to set the damage scaling factors in
oxide and nitride to 1.0. For example:
fproc dfactor_Oxide_Boron { energy dose } { return 1.0 }

After setting MCDFactor or dfactor for oxide and nitride, the following example shows how
to define the scaling of the etching rate for oxide as a function of implantation damage, using
the calibration previously presented:
sel z="1.0+1.45*(Damage/8e21) / \
(exp((1.0/1.5)*log(1.0+exp(1.5*log(Damage/8e21)))))" \
Oxide name= EtchScale store

A subsequent etch command could be as follows:

etch time= 0.01 rate= 1 material= Oxide isotropic \
etch.rate.modifier = "EtchScale" steps=1 force.full.levelset

To ensure robust results for the etching, you might need to decrease the values of the mgoals
arguments dx, dy, and dz with respect to their defaults. The best choice depends on the use
case. For example:
mgoals dx=0.002 dy=0.002 dz=0.002

References
[1] P. Pichler, Intrinsic Point Defects, Impurities, and Their Diffusion in Silicon,
Computational Microelectronics, Vienna: Springer, 2004.
[2] H. Bracht, N. A. Stolwijk, and H. Mehrer, “Equilibrium Concentrations of Intrinsic
Point Defects in Silicon Determined by Zinc Diffusion,” in Proceedings of the Seventh
International Symposium on Silicon Materials Science and Technology (Semiconductor
Silicon), vol. 94-10, San Francisco, CA, USA, pp. 593–602B, May 1994.
[3] H. Bracht, E. E. Haller, and R. Clark-Phelps, “Silicon Self-Diffusion in Isotope
Heterostructures,” Physical Review Letters, vol. 81, no. 2, pp. 393–396, 1998.
[4] A. Ural, P. B. Griffin, and J. D. Plummer, “Self-Diffusion in Silicon: Similarity between
the Properties of Native Point Defects,” Physical Review Letters, vol. 83, no. 17,
pp. 3454–3457, 1999.
[5] N. E. B. Cowern et al., “Energetics of Self-Interstitial Clusters in Si,” Physical Review
Letters, vol. 82, no. 22, pp. 4460–4463, 1999.

Advanced Calibration for Process Simulation User Guide 161

P-2019.03
3: Guidelines for Additional Calibration
References

[6] P. M. Fahey, P. B. Griffin, and J. D. Plummer, “Point defects and dopant diffusion in
silicon,” Reviews of Modern Physics, vol. 61, no. 2, pp. 289–388, 1989.
[7] I. Bork and H. Matsumoto, “On the Determination of Boron Diffusivities and Boron
Interstitial Pair Binding Energies in Silicon,” in International Conference on Simulation
of Semiconductor Processes and Devices (SISPAD), Tokyo, Japan, pp. 91–92,
September 1996.
[8] C. S. Rafferty et al., “Simulation of cluster evaporation and transient enhanced diffusion
in silicon,” Applied Physics Letters, vol. 68, no. 17, pp. 2395–2397, 1996.
[9] P. A. Stolk et al., “Physical mechanisms of transient enhanced dopant diffusion in ion-
implanted silicon,” Journal of Applied Physics, vol. 81, no. 9, pp. 6031–6050, 1997.
[10] L. Shao et al., “Retardation of boron diffusion in silicon by defect engineering,” Applied
Physics Letters, vol. 78, no. 16, pp. 2321–2323, 2001.
[11] L. F. Giles et al., “Transient enhanced diffusion of B at low temperatures under extrinsic
conditions,” Solid-State Electronics, vol. 49, no. 4, pp. 618–627, 2005.
[12] G. Impellizzeri et al., “Role of fluorine in suppressing boron transient enhanced
diffusion in preamorphized Si,” Applied Physics Letters, vol. 84, no. 11, pp. 1862–1864,
2004.
[13] J. M. Jacques et al., “Fluorine-enhanced boron diffusion in amorphous silicon,” Applied
Physics Letters, vol. 82, no. 20, pp. 3469–3471, 2003.
[14] B. Colombeau et al., “Current Understanding and Modeling of B Diffusion and
Activation Anomalies in Preamorphized Ultra-Shallow Junctions,” in MRS Symposium
Proceedings, Silicon Front-End Junction Formation—Physics and Technology, vol. 810,
San Francisco, CA, USA, pp. 91–102, April 2004.
[15] R. Duffy et al., “Low-temperature diffusion of high-concentration phosphorus in
silicon, a preferential movement toward the surface,” Applied Physics Letters, vol. 86,
no. 8, p. 081917, 2005.
[16] V. C. Venezia et al., “Dopant redistribution effects in preamorphized silicon during low
temperature annealing,” in IEDM Technical Digest, Washington, DC, USA,
pp. 489–492, December 2003.
[17] T. Noda, “Modeling of End-of-Range (EOR) Defects for Indium Channel Engineering,”
in IEDM Technical Digest, Washington, DC, USA, pp. 839–842, December 2001.
[18] H. Graoui et al., “TCAD Modeling and Experimental Investigation of Indium for
Advanced CMOS Technology,” in 14th International Conference on Ion Implantation
Technology (IIT), Taos, NM, USA, pp. 126–130, September 2002.
[19] C. Zechner et al., “Modeling Ultra Shallow Junctions Formed by Phosphorus-Carbon
and Boron-Carbon Co-implantation,” in MRS Symposium Proceedings, Semiconductor
Defect Engineering–Materials, Synthetic Structures and Devices II, vol. 994, San
Francisco, CA, USA, p. 0994–F11–17, April 2007.

162 Advanced Calibration for Process Simulation User Guide

P-2019.03
 3: Guidelines for Additional Calibration
References

[20] C.-L. Liu et al., “Ab initio modeling and experimental study of C–B interactions in Si,”
Applied Physics Letters, vol. 80, no. 1, pp. 52–54, 2002.
[21] I. Ban, M. C. Öztürk, and E. K. Demirlioglu, “Suppression of Oxidation-Enhanced
Boron Diffusion in Silicon by Carbon Implantation and Characterization of MOSFET’s
with Carbon-Implanted Channels,” IEEE Transactions on Electron Devices, vol. 44,
no. 9, pp. 1544–1551, 1997.
[22] G. Zschätzsch et al., “Fundamental Study on the Impact of C Co-Implantation on Ultra
Shallow B Junctions,” in Extended Abstracts of the Ninth International Workshop on
Junction Technology (IWJT), Kyoto, Japan, pp. 123–126, June 2009.
[23] B. Colombeau et al., “Ultra-Shallow Carborane Molecular Implant for 22-nm node p-
MOSFET Performance Boost,” in Extended Abstracts of the Ninth International
Workshop on Junction Technology (IWJT), Kyoto, Japan, pp. 27–30, June 2009.
[24] C. I. Li et al., “Enabling Solutions for 28nm CMOS Advanced Junction Formation,” in
18th International Conference on Ion Implantation Technology (IIT), Kyoto, Japan,
June 2010.
[25] IST Project 027152 ATOMICS, Advanced Front-End Technology Modeling for
Ultimate Integrated Circuits, for more information, go to https://
www.iisb.fraunhofer.de/content/dam/iisb2014/en/Documents/Research-Areas/
Simulation/final-report-atomics.pdf.
[26] L. S. Adam et al., “Diffusion of implanted nitrogen in silicon,” Journal of Applied
Physics, vol. 87, no. 5, pp. 2282–2284, 2000.
[27] L. S. Adam et al., “Diffusion of Implanted Nitrogen in Silicon at High Doses,” in MRS
Symposium Proceedings, Si Front-End Processing–Physics and Technology of Dopant-
Defect Interactions III, vol. 669, p. J3.10, January 2001.
[28] N. Zographos et al., “Process Modeling of Chemical and Stress Effects in SiGe,” in 19th
International Conference on Ion Implantation Technology (IIT), Valladolid, Spain,
pp. 212–216, June 2012.
[29] R. Kögler et al., “Excess vacancies in high energy ion implanted SiGe,” Journal of
Applied Physics, vol. 101, no. 3, p. 033508, 2007.
[30] M. Posselt, L. Bischoff, and J. Teichert, “Influence of dose rate and temperature on ion-
beam-induced defect evolution in Si investigated by channeling implantation at different
doses,” Applied Physics Letters, vol. 79, no. 10, pp. 1444–1446, 2001.
[31] A. Renau, “Device performance and yield - A new focus for ion implantation,” in
Extended Abstracts of International Workshop on Junction Technology (IWJT),
Shanghai, China, 2010.
[32] F. Khaja et al., “Benefits of Damage Engineering for PMOS Junction Stability,” in 18th
International Conference on Ion Implantation Technology (IIT), Kyoto, Japan, June
2010.

Advanced Calibration for Process Simulation User Guide 163

P-2019.03
3: Guidelines for Additional Calibration
References

[33] F. A. Khaja et al., “Physical understanding of cryogenic implant benefits for electrical
junction stability,” Applied Physics Letters, vol. 100, no. 11, p. 112102, 2012.
[34] R. D. Goldberg, J. S. Williams, and R. G. Elliman, “Amorphization of silicon by
elevated temperature ion irradiation,” Nuclear Instruments and Methods in Physics
Research B, vol. 106, no. 1–4, pp. 242–247, 1995.
[35] H.-J. Gossmann et al., “Predictive Process Simulation of Cryogenic Implants for
Leading Edge Transistor Design,” in 19th International Conference on Ion Implantation
Technology (IIT), Valladolid, Spain, pp. 225–228, June 2012.
[36] A. Schmidt et al., “Compact Process Model of Temperature Dependent Amorphization
Induced by Ion Implantation,” in International Conference on Simulation of
Semiconductor Processes and Devices (SISPAD), Bologna, Italy, pp. 197–200,
September 2010.
[37] N. Zographos, A. Tsibizov, and C. Zechner, “Continuum Modeling of Implantation and
Thermal Processes for Advanced Devices Formation,” in 20th International Conference
on Ion Implantation Technology (IIT), Portland, OR, USA, pp. 189–194, June 2014.
[38] N. Ioannou et al., “Germanium substrate loss during low temperature annealing and its
influence on ion-implanted phosphorous dose loss,” Applied Physics Letters, vol. 93,
no. 10, p. 101910, 2008.
[39] Y.-S. Oh and D. E. Ward, “A Calibrated Model for Trapping of Implanted Dopants at
Material Interface During Thermal Annealing,” in IEDM Technical Digest, San
Francisco, CA, USA, pp. 509–512, December 1998.
[40] E. Bellandi and V. Soncini, “SiO2 etch rate modification by ion implantation,” Thin
Solid Films, vol. 524, pp. 75–80, December 2012.
[41] R. Charavel and J.-P. Raskin, “Etch Rate Modification of SiO2 by Ion Damage,”
Electrochemical and Solid-State Letters, vol. 9, no. 7, pp. G245–G247, 2006.

164 Advanced Calibration for Process Simulation User Guide

P-2019.03
CHAPTER 4 Advanced Calibration for 4H-SiC
Process Simulation

This chapter gives an introduction to the use of Advanced

Calibration for silicon carbide (SiC).

Content of Advanced Calibration File for 4H-SiC

Simulation
Sentaurus Process has one material corresponding to silicon carbide – SiliconCarbide –
which also has an alias SiC. The default parameters for material SiliconCarbide
correspond to the 4H-SiC polytype. The Advanced Calibration file for SiC contains the
recommended settings for 4H-SiC simulations. Advanced Calibration mainly supports process
simulation of the 4H-SiC polytype, because it is the most widely used SiC polytype in
electronics production. Only some of its parameters are valid for other SiC polytypes,
therefore, the parameter file must not be used for the simulation of other SiC polytypes as it is,
but must be adapted appropriately by users. In general, this set of parameters is less mature than
the one for silicon and SiGe, especially with respect to dopant activation and diffusion, and the
treatment of point defects in SiC.

The Advanced Calibration file AdvCal_SiC_2019.03.fps for 4H-SiC is divided into the
following parts, which contain numbered sections and are executed in sequence:
■ Part 1: Basic model switches
■ Part 2: Settings for Monte Carlo implantation
■ Part 3: Initial conditions after ion implantation

Part 1: Basic Model Switches

Part 1 of the Advanced Calibration file offers a choice that is recommended by Synopsys for
most SiC applications.

It is often useful to reduce the number of equations to be solved in order to save CPU time. On
the other hand, it might be sometimes necessary to select models that are more sophisticated
than the default choice, even at the cost of increasing the CPU time. In this chapter, the possible
changes with respect to the default model switches are explained.

Advanced Calibration for Process Simulation User Guide 165

P-2019.03
4: Advanced Calibration for 4H-SiC Process Simulation
Part 1: Basic Model Switches

The most elegant way to change a basic model switch is adding a corresponding line at the
beginning of a project-specific or user-specific calibration file, which is sourced after loading
the Advanced Calibration file. In this way, you can track the differences to the default
suggestions of Synopsys.

First, maxAnnealT is increased to the melting point value in 4H-SiC ( 2730°C ) from the
default 1400°C to allow high-temperature annealings, which are typical for SiC process
simulations:
pdbSet Diffuse maxAnnealT 2730.0

The set4H-SiC procedure sets the number of lattice parameters for the 4H-SiC polytype of
SiliconCarbide. Although these parameters are the default ones, the procedure is placed
into the Advanced Calibration file as a reference for users who want to create their own
parameter files for other SiC polytypes, based on SiliconCarbide material.

Numeric solver parameters, which are necessary to ensure correct simulation results at low-
temperature anneal (NegErrCntrl) and stable convergence with the transient activation
model (fullNewton), are set by the following command:
math NegErrCntrl fullNewton

Poisson Equation
The Poisson equation is switched on in SiC by the following command:
pdbSet SiC Potential Poisson 1

Besides the formally more correct distribution of electrostatic potential and mobile carrier
densities, simulations with the Poisson equation lead to better convergence with Fermi-
dependent models, due to smoother potential and carrier profiles.

The value of permittivity for 4H-SiC is set to 10.03, that is, the one which applies to the {0001}
direction (parallel to the c-axis). The value of permittivity normal to the c-axis equals 9.66 [1]:
pdbSet SiC Potential Permittivity 10.03

The intrinsic carrier concentration (ni) of 4H-SiC is set by the density-of-states (DOS) model:
pdbSet SiC Potential Use.DOS 1

166 Advanced Calibration for Process Simulation User Guide

P-2019.03
 4: Advanced Calibration for 4H-SiC Process Simulation
Part 1: Basic Model Switches

Dopant Cluster Models in 4H-SiC

In Sentaurus Process, the Solid model is the default for dopant activation in SiC. The default
values of the Solubility parameters for Al and B are taken without calibration from [2];
whereas, the values of the Solubility parameters for N and P are based on theoretical work
[3]. The phosphorus solubility value was increased by a factor of four to match experimental
values of activation at 1700°C from [4].

A major disadvantage of the Solid model is that it does not reflect the nonequilibrium nature
of dopant activation. For example, with the Solid model, the activation instantly follows the
annealing temperature. In Advanced Calibration, the Transient activation model is switched
on for Al, B, N, and P. Activation of dopants in SiC is a complex process, which includes
dynamic interaction of dopants with different defects and defect agglomerations. It depends on
the implantation conditions, the presence of co-implants, the anneal atmosphere, the ramp
rates, and so on [5][6][7][8][9]. Although the Transient model itself cannot simulate such
complex behavior in detail and cannot capture all the physical effects, it is still more physical
than the Solid model and can emulate, to some extent, the nonequilibrium transient behavior
of the activation process.

The current Advanced Calibration values of the Transient model parameters for Al, B, N,
and P were obtained by qualitative calibration to sheet resistance, Hall measurements, and
capacity measurements described in the literature. Most of the experimental structures had
homogeneously implanted layers, obtained by multiple implantations, and were annealed at
approximately 1500–1700°C [4][8][9][10][11][12][13][14][15], and at a higher temperature
for aluminum [5][16]. In general, qualitative rather than quantitative agreement to published
experiments was achieved. Some experimental trends (such as reduction of activation for Al
with increasing annealing time at 1800°C in [5]) are not captured. Overall calibration quality
is much worse than for the Transient activation model in silicon.

However, taking into account the strong dependency of the activation on particular process
conditions, you should use the Transient activation parameters as a starting point for custom
calibration of dopant activation. Since exact information about amorphization is not available
in most publications (neither is it calibrated in Sentaurus MC for 4H-SiC), the AcInit and
15 –3
AmInit parameters for initial dopant activation are both set to the same value 10 cm ,
which is smaller than the typical background doping concentration in epitaxial layers. The
calibration used the 4H-SiC device simulation parameters from Advanced Calibration for
Device Simulation Version P-2019.03.

It is difficult to interpret the experimental results of dopant activation in SiC for various
reasons:
■ Uncertainties about the value of the Hall scattering factors (especially for holes) [17][18].
■ Temperature dependency of the degeneracy factors of acceptors, which is not included in
the IncompleteIonization model of Sentaurus Device [19][20][21].

Advanced Calibration for Process Simulation User Guide 167

P-2019.03
4: Advanced Calibration for 4H-SiC Process Simulation
Part 1: Basic Model Switches

■ Uncertainties about the ionization levels of active dopants (especially for concentrations
19 –3
above 10 cm ).
■ The presence of high-concentration effects (for example, hopping transport) during the
ionization of active dopants and carrier transport [22].

In addition, ion-implanted layers contain high concentrations of defects, which remain at

relatively high concentrations even after high-temperature anneals. These defects serve as
deep-doping compensation centers and carrier-scattering centers, thereby affecting the net
active doping concentration, the concentrations of the mobile carriers, and the carrier mobility.
An interesting approach for the consideration of such defects is suggested in [23].

NOTE Such defects have not been included in the simulations, which might
explain the systematic underestimation of sheet resistance, which was
observed in the calibrations.

Dopant Transport at the Oxide–SiC Interface

Dopant dose loss is not calibrated for the oxide–SiC (Ox_SiC) interface in TCAD Sentaurus
tools. The existing Parameter Database default parameters of Sentaurus Process are copied
literally from the Oxide_Si interface and, therefore, they do not guarantee the trustworthiness
of simulation results.

Since there are no trustworthy parameters for the segregation of dopants, the simplest possible
HomNeumann boundary condition is set explicitly at the Ox_SiC interface in the Advanced
Calibration file. Of course, you can specify other types of boundary condition (for example,
Segregation or ThreePhaseSegregation) at the Ox_SiC interface as well as the
parameters for the boundary conditions.

Boron Evaporation From the SiC Surface

Significant boron dose loss has been observed experimentally after high-temperature
(> 1500°C) anneal of 4H-SiC [2][12][24]. Advanced Calibration simulates boron dose loss by
boron evaporation at SiC–gas interfaces using natural boundary conditions:
pdbSetSwitch Gas_SiC Boundary BoundaryCondition Natural

The boron evaporation rate is controlled by the Ksurf parameter. The value of Ksurf is set in
Advanced Calibration by the following command:
pdbSetDouble Gas_SiC Boron Ksurf {[expr 1e4*[pdbGet SiC B Dstar]]}

168 Advanced Calibration for Process Simulation User Guide

P-2019.03
 4: Advanced Calibration for 4H-SiC Process Simulation
Part 2: Settings for Monte Carlo Implantation

This value is suitable for 1700°C anneals from [24], where low-temperature preannealing
(2 hours at 900°C) was used to reduce boron dose loss: For high-temperature anneals without
preannealing [12], an order of magnitude larger value of Ksurf is required. For the correct
simulation of boron evaporation at high-temperature (> 1500°C) anneals, it is necessary to
remove the covering oxide from the SiC surface. For all other dopants except for boron, zero
values of Ksurf are set at SiC–gas interfaces by default, which is equivalent to the Neumann
boundary conditions for these dopants.

Part 2: Settings for Monte Carlo Implantation

There has been additional calibration of Sentaurus MC to improve the accuracy of simulations
for implantations of aluminum, boron, and nitrogen into 4H-SiC with a wafer surface
orientation close to (0001).

The calibration is based on comparison of simulation results to SIMS measurements of as-

implanted profiles, either taken from publications [25][26][27][28] or provided by companies
who support the calibration efforts of Synopsys. The data includes implantations at room
temperature as well as implantations into heated substrates (> 300°C ). The calibration focused
on the following range of implantation energies and doses:
■ Aluminum:
• Most implantation energies in the range between 10 keV and 5 MeV
13 –2 14 –2
• Most doses in the range between 10 cm and 5 ×10 cm
■ Boron:
• Energies between 10 keV and 5.8 MeV
13 –2 14 –2
• Most doses in the range between 10 cm and 5 ×10 cm
■ Nitrogen:
• Energies between 10 keV and 5.8 MeV
14 –2 15 –2
• Most doses in the range between 10 cm and 2 ×10 cm

In total, approximately 100 SIMS profiles have been used for the calibration, typically, with a
small wafer miscut angle and a small tilt angle. When using tilted implantations into wafers
with miscut, the results depend on the miscut angle, the tilt angle, the ion beam rotation angle,
the direction of the miscut, and the crystal orientation of the primary wafer flat.

However, not all of the published data used for calibration has all the relevant angles and
orientations explicitly specified. Therefore, assumptions were made about correct angles or
data was omitted from the calibration.

Advanced Calibration for Process Simulation User Guide 169

P-2019.03
4: Advanced Calibration for 4H-SiC Process Simulation
Part 2: Settings for Monte Carlo Implantation

The calibration is widely based on implantation data for off-channel directions. It is probably
less accurate for implantations that are perfectly parallel to the crystal channel directions (such
as tilt zero implantations into wafers without miscut).

For hot implantations, the highest implantation energies of the experimental data that was
referred are approximately 700 keV, and the reproducibility of results for a higher energy than
that might be less reliable.

The displacement threshold and the temperature dependency of damage were adjusted taking
into account experimental data from Kalinina [29] and the references therein. The
displacement threshold in 4H-SiC is larger for Si atoms than for C atoms. It depends on the
direction of the displacement, the nature of the impacting projectile, and the amount of already
accumulated damage. The smallest displacement threshold value in a perfect crystal reported
in the literature (for C atoms) is 18 eV [30]. In the binary collision simulators, typically some
averaged values for the displacement threshold are used. Sentaurus MC allows only a single
constant displacement threshold value in SiC. In Advanced Calibration for 4H-SiC, the
disp.thr value has increased from the default 15 eV to 20 eV to achieve better
correspondence to the real value.

Impact of Wafer Temperature

The substrate temperature during implantation affects lattice self-repair during implantation.
Implantation creates crystal damage such as Si or C interstitials and vacancies. It is possible
that a part of this damage is annealed already during implantation. In particular, if the
implantation is performed into heated wafers (hot implantation), the faster recombination of
interstitials and vacancies at high temperatures reduces the crystal damage during
implantation.

Therefore, a high wafer temperature during implantation reduces crystal damage and possible
amorphization, thereby affecting ion channeling. Lattice vibration also increases at higher
temperatures, but it has only a small impact on implantation and might dominate only at low
damage levels.

During the calibration, it was found that best results are obtained when using slightly different
model parameters for room temperature implantations and hot implantations. Therefore, two
parameter sets were prepared: one for room temperature implantations and one for hot
implantations. Calibration of a sophisticated, smooth dependency of Monte Carlo parameters
on wafer temperature is beyond the scope of this work, due to limited data availability.

Note that different sources of data were used for calibration of parameters for room
temperature and hot implantations. It is possible that differences in results are caused not only
by the temperature difference itself, but also by inconsistencies between the experimental data
used, due to experimental or measurement inaccuracies, or to inaccurate or incomplete

170 Advanced Calibration for Process Simulation User Guide

P-2019.03
 4: Advanced Calibration for 4H-SiC Process Simulation
Part 2: Settings for Monte Carlo Implantation

specification of implantation conditions (such as wafer orientation, rotation angle, or through-

oxide thickness).

Table 8 lists the Sentaurus Process parameters that were changed with respect to the default
values of Sentaurus Process Version P-2019.03 to improve the accuracy of Sentaurus MC.
Some parameters are defined as functions of the implantation energy E . For these, the
formulas are given in Table 8.

The parameters are explained in the Sentaurus™ Process User Guide. For better accuracy,
some parameters are defined as smooth functions of the implantation energy E (unit is keV).
The parameters that are not changed from the defaults are not included in Table 8.

Table 8 Sentaurus Process default parameters and calibrated parameters for room
temperature implantation and hot implantation into SiC
Species Parameter Sentaurus Process Calibrated value: Implantation at room Calibrated value: Hot
default value in temperature (25oC) implantation
Version P-2019.03

All (Al, B, N) disp.thr 15 20 20

Aluminum d.sim 0.225 0.2 0.25

surv.rat 0.275 0.37 0.20

LSS.pre 1 E ≤ 400 : 0.907 0.9

E > 400 : 0.06321*ln(E) + 0.5283

nloc.pre 0.075 0.1 0.1

nloc.exp 0.075 1.2 1.0

Boron surv.rat 0.25 0.8 0.8

LSS.pre 1.35 0.04 1.245

1.245 + -------------------------------------------------------------------------
1 + exp ( – 0.015 × ( E – 2500 ) )

nloc.exp 0.075 0.15 0.075

scr.par 0.5 0.55 0.55

Nitrogen surv.rat 0.5 0.8 0.3

sat.par 1 1 2

LSS.pre 1.5 E ≤ 200 : E ≤ 400 :

1.01 0.6
1.3 + 0.2 ×  ------------------ 1.25 + 0.15 ×  ------------------
200 – E 400 – E
 200   400 
E > 200 : E > 400 :
E – 200 E – 400
1.3 – 0.02 × ----------------------------------- 1.25 – 0.03 × -----------------------------------
20 + ( E – 200 ) 20 + ( E – 400 )

Advanced Calibration for Process Simulation User Guide 171

P-2019.03
4: Advanced Calibration for 4H-SiC Process Simulation
Part 2: Settings for Monte Carlo Implantation

To allow for the definition of energy-dependent parameters, the parameter file uses the callback
procedure _AdvCal::ImpPreProcess, which is called from Sentaurus Process at the very
beginning of the execution of an implant command. The implantation conditions (species,
energy, dose, and so on) are handed over as procedure arguments and are used to define the
parameters dependent on species and energy.

Furthermore, the wafer temperature during implantation is accessed inside the

_AdvCal::ImpPreProcess procedure by the simGetDouble Diffuse temp statement.
If the wafer temperature is < 300°C , the parameters calibrated for room temperature are used.
For wafer temperatures ≥ 300°C , the parameters calibrated for hot implantation are used.

You must specify the implantation temperature (in degree Celsius) in the implant command,
for example:
implant Arsenic dose= 2.0e+13 energy= 160 tilt= 0 temperature= 450

If no temperature is specified, the simGetDouble Diffuse temp function gives the default
implantation temperature (by default, 26.84°C ), which is specified (in kelvin) by:
pdbSet MCImplant Temperature 300.0

To overwrite the settings performed by the _AdvCal::ImpPreProcess procedure, you can

define the UserImpPreProcess procedure in your own specific parameter file. The
UserImpPreProcess procedure uses syntax similar to that of _AdvCal::ImpPreProcess
(see also Example of a User Calibration File on page 150).

Recommendations for Fine-Tuning of Amorphization

The calibration presented in this chapter is based entirely on comparison of simulated profiles
to SIMS data. No calibration has been performed for a possible amorphization of SiC by high-
dose ion implantations, due to lack of suitable experimental data.

Users who want to study amorphization of SiC using Monte Carlo implantation should test the
accuracy of TCAD results against measured data on the existence and thickness of amorphized
layers after implantation. The following parameters are recommended for fine-tuning of
amorphization:
■ surv.rat
■ sat.par

A change of surv.rat has a stronger impact on the channeling tail of an as-implanted profile.

To completely suppress amorphization for hot implantations, set sat.par to a high value (for
example, 10). This suppresses the increase of crystal damage beyond a level of N dens /sat.par,
where N dens is the lattice density.

172 Advanced Calibration for Process Simulation User Guide

P-2019.03
 4: Advanced Calibration for 4H-SiC Process Simulation
Part 3: Initial Conditions After Ion Implantation

Illustration of Calibration Results

Typically, using the calibrated parameters will lead to a significantly better agreement between
simulations results and SIMS data for high-energy implantations. This is illustrated in
Figure 33. A detailed discussion of the simulation accuracy for all of the data studied is beyond
the scope of this user guide.
1e+20
Sentaurus Process default
Advanced Calibration
SIMS
Concentration [cm−3]

1e+19

1e+18

0 0.5 1 1.5
Depth [µm]
Figure 33 Comparison of simulation results for default and calibrated parameters with
nitrogen SIMS from [25]: implantation conditions are N 1000 keV, dose
1.3e15 cm–2, tilt=7, rotation=0, surface orientation (001), miscut=3.5o towards
(110). The direction of the primary wafer flat (110) is assumed, and simulations
have been performed with Sentaurus Process Version J-2014.09.

Part 3: Initial Conditions After Ion Implantation

In most SiC simulations, very simple models such as the default Sentaurus Process Constant
and Solid models for dopant diffusion and activation in SiC are used. Therefore, to reduce the
simulation time, the point-defect transport equations are not solved in SiC by default. To
compute the point-defect transport in SiC, use the following command:
pdbSet SiC Compute.Point.Defect 1

The following command provides correct initialization of the point defects in the beginning of
a diffuse command:
pdbSet Diffuse Use.New.Init.Method 1

Advanced Calibration for Process Simulation User Guide 173

P-2019.03
4: Advanced Calibration for 4H-SiC Process Simulation
References

References
[1] L. Patrick and W. J. Choyke, “Static Dielectric Constant of SiC,” Physical Review B,
vol. 2, no. 6, pp. 2255–2256, 1970.
[2] M. K. Linnarsson et al., “Aluminum and boron diffusion in 4H-SiC,” in MRS
Symposium Proceedings, Silicon Carbide—Materials, Processing, and Devices,
vol. 742, pp. K6.1.1–K6.1.11, 2002.
[3] M. Bockstedte, A. Mattausch, and O. Pankratov, “Solubility of nitrogen and phosphorus
in 4H-SiC: A theoretical study,” Applied Physics Letters, vol. 85, no. 1, pp. 58–60, 2004.
[4] M. Laube et al., “Electrical activation of high concentrations of N+ and P+ ions
implanted into 4H–SiC,” Journal of Applied Physics, vol. 92, no. 1, pp. 549–554, 2002.
[5] Y. Negoro et al., “Electrical activation of high-concentration aluminum implanted in
4H-SiC,” Journal of Applied Physics, vol. 96, no. 9, pp. 4916–4922, 2004.
[6] K. A. Jones et al., “Effects of high-temperature anneals on 4H–SiC Implanted with Al
or Al and Si,” Journal of Applied Physics, vol. 96, no. 10, pp. 5613–5618, 2004.
[7] A. Poggi et al., “Effects of heating ramp rates on the characteristics of Al implanted
4H–SiC junctions,” Applied Physics Letters, vol. 88, no. 16, p. 162106, 2006.
[8] J. Senzaki, K. Fukuda, and K. Arai, “Influences of postimplantation annealing
conditions on resistance lowering in high-phosphorus-implanted 4H–SiC,” Journal of
Applied Physics, vol. 94, no. 5, pp. 2942–2947, 2003.
[9] Y. Negoro, T. Kimoto, and H. Matsunami, “Carrier compensation near tail region in
aluminum- or boron-implanted 4H–SiC (0001),” Journal of Applied Physics, vol. 98,
no. 4, p. 043709, 2005.
17 20 –3
[10] E. M. Handy et al., “Variable-dose ( 10 – 10 cm ) phosphorus ion implantation into
4H–SiC,” Journal of Applied Physics, vol. 88, no. 10, pp. 5630–5634, 2000.
[11] C. P. Morath, Electrical Characterization of Ion-Implanted 4H-Silicon Carbide,
Master’s thesis, Air Force Institute of Technology, Air University, Ohio, USA, March
1999.
[12] T. Troffer et al., “Doping of SiC by Implantation of Boron and Aluminum,” Physica
Status Solidi A, vol. 162, no. 1, pp. 277–298, 1997.
[13] M. V. Rao et al., “Ion-implantation in bulk semi-insulating 4H–SiC,” Journal of Applied
Physics, vol. 86, no. 2, pp. 752–758, 1999.
[14] N. S. Saks et al., “Low-dose aluminum and boron implants in 4H and 6H silicon
carbide,” Journal of Applied Physics, vol. 90, no. 6, pp. 2796–2805, 2001.
[15] N. S. Saks, S.-H. Ryu, and A. V. Suvorov, “Low-dose n-type nitrogen implants in 4H-
SiC,” Applied Physics Letters, vol. 81, no. 26, pp. 4958–4960, 2002.

174 Advanced Calibration for Process Simulation User Guide

P-2019.03
 4: Advanced Calibration for 4H-SiC Process Simulation
References

[16] R. Nipoti et al., “Microwave Annealing of Very High Dose Aluminum-Implanted 4H-
SiC,” Applied Physics Express, vol. 4, no. 11, p. 111301, 2011.
[17] F. Schmid et al., “Hall Scattering Factor for Electrons and Holes in SiC,” Silicon
Carbide, Advanced Texts in Physics, W. J. Choyke, H. Matsunami, and G. Pensl (eds.),
Berlin: Springer, 2004.
[18] A. Koizumi, J. Suda, and T. Kimoto, “Temperature and doping dependencies of
electrical properties in Al-doped 4H-SiC epitaxial layers,” Journal of Applied Physics,
vol. 106, no. 1, p. 013716, 2009.
[19] H. Matsuura, “Investigation of a distribution function suitable for acceptors in SiC,”
Journal of Applied Physics, vol. 95, no. 8, pp. 4213–4218, 2004.
[20] H. Matsuura, “Influence of excited states of a deep substitutional dopant on majority-
carrier concentration in semiconductors,” Physical Review B, vol. 74, no. 24, p. 245216,
2006.
[21] H. Matsuura, “Determination Methods of Densities and Energy Levels of Impurities and
Defects Affecting Majority-Carrier Concentration in Next-Generation
Semiconductors,” Advances in Condensed Matter and Materials Research, vol. 10,
H. Geelvinck and S. Reynst (eds.), Hauppauge, New York: Nova Science Publishers,
pp. 301–356, 2011.
[22] A. Parisini et al., “Remarks on the room temperature impurity band conduction in
heavily Al+ implanted 4H-SiC,” Journal of Applied Physics, vol. 118, no. 3, p. 035101,
2015.
[23] F. Giannazzo, F. Roccaforte, and V. Raineri, “Acceptor, compensation, and mobility
profiles in multiple Al implanted 4H-SiC,” Applied Physics Letters, vol. 91, no. 20,
p. 202104, 2007.
[24] H. Bracht et al., “Diffusion of boron in silicon carbide: Evidence for the kick-out
mechanism,” Applied Physics Letters, vol. 77, no. 20, pp. 3188–3190, 2000.
[25] R. Stief, Dotierung von 4H-SiC durch Ionenimplantation, Ph.D. thesis, University of
Erlangen-Nuremberg, Germany, 1999.
[26] J. Wong-Leung, M. S. Janson, and B. G. Svensson, “Effect of crystal orientation on the
implant profile of 60 keV Al into 4H-SiC crystals,” Journal of Applied Physics, vol. 93,
no. 11, pp. 8914–8917, 2003.
[27] K. Mochizuki et al., “Detailed Analysis and Precise Modeling of Multiple-Energy Al
Implantations Through SiO2 Layers Into 4H-SiC,” IEEE Transactions on Electron
Devices, vol. 55, no. 8, pp. 1997–2003, 2008.
[28] S. Morata, F. Torregrosa, and T. Bouchet, “Simulation of Ion Implantation in SiC:
Dopant Profiling and Activation,” Materials Science Forum, vol. 615–617,
pp. 449–452, March 2009.

Advanced Calibration for Process Simulation User Guide 175

P-2019.03
4: Advanced Calibration for 4H-SiC Process Simulation
References

[29] E. V. Kalinina, “The Effect of Irradiation on the Properties of SiC and Devices Based
on this Compound,” Semiconductors, vol. 41, no. 7, pp. 745–783, 2007.
[30] J. Lefèvre et al., “Silicon threshold displacement energy determined by
photoluminescence in electron-irradiated cubic silicon carbide,” Journal of Applied
Physics, vol. 105, no. 2, p. 023520, 2009.

176 Advanced Calibration for Process Simulation User Guide

P-2019.03
Part II Advanced Calibration in Sentaurus
Process Kinetic Monte Carlo

This part of the Advanced Calibration for Process Simulation User Guide contains the
following chapters:

Chapter 5 Using Advanced Calibration File of Sentaurus Process KMC on page 179

Chapter 6 Contents of Advanced Calibration of Sentaurus Process KMC on page 183

Chapter 7 Guidelines for Additional Calibration on page 239

CHAPTER 5 Using Advanced Calibration File of
Sentaurus Process KMC

This chapter gives an introduction to the use of Advanced

Calibration in a process simulation with Sentaurus Process Kinetic
Monte Carlo (Sentaurus Process KMC).

Advanced Calibration is a selection of models and parameters, which is recommended by

Synopsys to be used for accurate process simulation. In Sentaurus Process KMC, this selection
of models and parameters is contained in a text file, which can be opened with any standard
text editor.

By sourcing the Advanced Calibration file at the beginning of a process simulation, the
standard calibration of Synopsys is selected. If needed, you can change or extend the Advanced
Calibration. This can be performed by either sourcing an additional calibration file that
contains the required parameter changes or editing the Advanced Calibration file with a text
editor.

Location of Advanced Calibration File

The Advanced Calibration file is the ultimate product of Synopsys’ calibration efforts. For each
release of Sentaurus Process, there is a new Advanced Calibration file that includes the best
and latest set of models and parameters. To ensure backward compatibility, previous Advanced
Calibration files are still available.

The files for the Advanced Calibration of Sentaurus Process KMC in this release are located at:
$STROOT/tcad/$STRELEASE/lib/sprocess/TclLib/AdvCal

The default file is named AdvCal_KMC_2019.03.fps. It represents the first version of

Advanced Calibration P-2019.03. Older versions of the Advanced Calibration file can be found
in the same directory. For example, the file AdvCal_KMC_2018.06.fps contains the
Advanced Calibration file for Version O-2018.06 and is available for backward compatibility.

Advanced Calibration for Process Simulation User Guide 179

P-2019.03
5: Using Advanced Calibration File of Sentaurus Process KMC
Using Advanced Calibration

Using Advanced Calibration

To use the Advanced Calibration of Sentaurus Process KMC, you must select the atomistic
mode by using the command SetAtomistic followed by the command
AdvancedCalibration, which is the same one as for the continuum Advanced Calibration.
The command AdvancedCalibration checks whether the atomistic mode is set and loads
the corresponding Advanced Calibration file. Therefore, at the beginning of the input file,
insert the lines:
SetAtomistic
AdvancedCalibration 2019.03

To load Advanced Calibration of Sentaurus Process KMC in continuum mode, you must use
the KMC flag:
AdvancedCalibration KMC

Alternatively, a local copy of this file can be sourced by using:

source ./AdvCal_KMC_2019.03.fps

NOTE The interoperability between Advanced Calibration for continuum

Sentaurus Process and Advanced Calibration for Sentaurus Process
KMC, meaning the mapping of solutions in continuum mode to
particles in atomistic mode, is controlled by the commands PDE2KMC
and KMC2PDE (see Sentaurus™ Process User Guide, Translating
Atomistic and Nonatomistic Information on page 383).

For ultrashallow junction experiments, the recommended minimum simulation size is 500 nm
x 40 nm x 40 nm. The depth (x-coordinate) of at least 500 nm is needed for accurate simulation
of the point-defect evolution, and a surface (y-coordinate and z-coordinate) of 40 nm x 40 nm
is needed for the accurate simulation of the damage accumulation during full cascade Monte
Carlo implantation simulation.

Additional Calibration by Users

Advanced Calibration is based on the assumption that all parameters that are not changed in
the parameter files are the default parameters of Sentaurus Process KMC. To use the Advanced
Calibration file AdvCal_KMC_2019.03.fps, it must be sourced before the real process
description.

180 Advanced Calibration for Process Simulation User Guide

P-2019.03
 5: Using Advanced Calibration File of Sentaurus Process KMC
Additional Calibration by Users

After sourcing AdvCal_KMC_2019.03.fps, you can change the model switches or parameter
values of the physical models. Ideally, this should be performed by experienced users with a
good understanding of the models of Sentaurus Process KMC.

For the process simulation of silicon technology, Advanced Calibration is usually the best
starting point. You can further increase the accuracy for a certain technology by additional fine-
tuning of a few physical parameters.

The best way to perform this is to put all additional calibration in a user calibration file, for
example, my_calibration.fps. This file includes all project-specific changes of physical
parameters with respect to Advanced Calibration.

In the process simulation file, at the beginning of the process simulation, you insert the lines:
SetAtomistic
AdvancedCalibration 2019.03
source ./my_calibration.fps

This approach allows you to:

■ Separate the calibration and the process descriptions completely.
■ Use the Advanced Calibration file as a starting point.
■ Summarize all project-specific calibration in a short and clear text file.

Advanced Calibration for Process Simulation User Guide 181

P-2019.03
5: Using Advanced Calibration File of Sentaurus Process KMC
Additional Calibration by Users

182 Advanced Calibration for Process Simulation User Guide

P-2019.03
CHAPTER 6 Contents of Advanced Calibration
of Sentaurus Process KMC

This chapter explains the contents of the Advanced Calibration file

of Sentaurus Process Kinetic Monte Carlo (Sentaurus Process
KMC) and documents the origin of the parameter values.

Overview
The focus of Advanced Calibration is monocrystalline silicon, germanium, and SiGe for all Ge
mole fractions. The calibration for silicon and SiGe with low Ge mole fraction ( ≤ 0.5 ) is the
most mature and reliable. On the other hand, the calibration for pure germanium is less mature
and reliable, and the calibration for SiGe with high Ge mole fraction (> 0.5) is the least mature
and reliable. Most of the model equations and model parameters are taken from reliable
publications. In addition, a rigorous calibration has been performed by Synopsys, based on a
SIMS database.

The Advanced Calibration of Sentaurus Process in continuum mode (see Chapter 2 on page 9),
which has proven good accuracy for a wide range of ultrashallow junction (USJ) experiments,
serves as the starting point for the Advanced Calibration of Sentaurus Process KMC. Wherever
possible, the parameters are inherited from the continuum approach to the kinetic Monte Carlo
(KMC) approach.

However, the calibration of the continuum models is verified and optimized for a high
temperature range ( 700°C – 1100°C ) only, whereas the kinetic Monte Carlo parameters
should be valid below room temperature as well. Moreover, some parameters and models
cannot be translated directly from the atomistic world to the continuum world. Taking into
account these two restrictions, the strategy is to translate the continuum Advanced Calibration
parameters if feasible, and to prefer or calibrate the default kinetic Monte Carlo parameters
where it is favorable.

The Advanced Calibration file AdvCal_KMC_2019.03.fps is divided into the following

sections, which contain numbered subsections and are executed in sequence:
■ Section 1: Silicon material
■ Section 2: Germanium material
■ Section 3: SiGe material
■ Section 4: Epitaxy

Advanced Calibration for Process Simulation User Guide 183

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Overview

The first three sections cover the KMC particle parameters and the lattice KMC (LKMC)
recrystallization parameters. These three sections are divided into two parts:
■ Part 1: Model parameters for damage and point defects
■ Part 2: Model parameters for impurities

The fourth section covers the LKMC epitaxy model parameters for silicon, germanium, and
SiGe.

This chapter is organized slightly differently for simplicity:

■ Part 1: Model Parameters for Implantation Damage and Point Defects on page 185 for the
materials silicon, germanium, and SiGe (of Sections 1–3)
■ Part 2: Model Parameters for Impurities on page 198 for the materials silicon, germanium,
and SiGe (of Sections 1–3)
■ Section 4: Model Parameters for Epitaxial Growth on page 223

Supported Materials
Advanced Calibration for Sentaurus Process KMC supports monocrystalline silicon,
germanium, and SiGe.

Pure silicon is modeled by the Sentaurus Process material Silicon with a native oxide
modeled by the Sentaurus Process material Oxide, while pure germanium is modeled by the
Sentaurus Process material Germanium with the native oxide GeOxide.

SiGe is modeled by the Sentaurus Process material Silicon with a field Germanium, which
Sentaurus Process KMC treats as a nonatomistic impurity in silicon.

NOTE It is not recommended to use the Sentaurus Process material

SiliconGermanium, mainly to prevent any artificial boundaries
between the Silicon and SiliconGermanium materials.

The Ge mole fraction dependency of SiGe can be modeled in two ways:

■ By default, Ge effects are modeled by linear correction factors. Furthermore, the
calibration of Ge effects [1][2] is based on a relatively small set of measured data. It focuses
on Ge concentration ranging between 0% and 50%, and it is not expected to be predictive
for higher concentrations of Ge (> 50%).
■ Optionally, you can enable Si and Ge parameter interpolation for SiGe using the command:
KMC_SiGe_and_Stress_Effect 1 0
The procedure KMC_SiGe_and_Stress_Effect has two Boolean arguments: the first is
Ge_Chem_Eff and the second is Stress_Eff. Setting Ge_Chem_Eff to 1 disables the

184 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

default linear correction factors and uses the interpolation parameters defined for
SiliconGermanium by the following parameter switches:
pdbSet Si Skip.Parameter.Interpolation 0
pdbSet Ge Skip.Parameter.Interpolation 0
The calibration covers the full Ge mole fraction range (0–100%). However, due to the
limited availability of experimental data for the high mole fraction range (> 50% and
< 100%), the calibration is expected to be less mature in this range.

In addition, parameter interpolation for mechanical parameters in compound materials and the
lattice mismatch model are enabled by the following parameter switches:
pdbSet Mechanics Parameter.Interpolation 1
pdbSet Mechanics Lattice.Constant.Mismatch 1

Part 1: Model Parameters for Implantation Damage and

Point Defects
This part describes the model parameters for implantation damage and point defects for the
materials silicon, germanium, and SiGe.

Amorphization and Recrystallization

The amorphization threshold, displacement thresholds, and interstitial–vacancy recombination
in amorphous pockets have been calibrated by the comparison of literature data with simulated
(temperature-dependent and dose rate–dependent) amorphous-crystalline transitions and
amorphization-layer thicknesses generated by silicon, germanium, and carbon implantation
into silicon [3] and into germanium [4][5][6]. The amorphization threshold is set to
22 –3
1.0 × 10 cm for both silicon and germanium, which is slightly lower than the value used
for continuum Advanced Calibration and is in the range of reported values for the critical point-
defect concentration [7].

The binary collision displacement thresholds for most implanted species is 15 eV for both
silicon and germanium substrate materials, which is consistent with reported values [8]. Higher
displacement thresholds are used for light ions such as B, C, and F in combination with the
MCImplant Cascades.Damage.Model Coupled model, which allows species-specific
displacement thresholds so as not to overestimate the simulated damage compared to
experimental data. This is a consequence of the limitations of the standard binary collision
approximation (BCA) model, which assumes that the deposited energy that is lower than the
displacement threshold is dissipated. Unfortunately, the improved BCA model (iBCA) is not a
valid alternative due to its inefficiency in general applications and calibration.

Advanced Calibration for Process Simulation User Guide 185

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

The recrystallization velocity in Advanced Calibration for an undoped amorphous layer on top
of (100)-oriented crystalline substrate is in agreement with the literature for silicon [9] and
germanium [6], both for the standard isotropic and the anisotropic lattice kinetic Monte Carlo
(LKMC) mode. The activation energy for solid phase epitaxial regrowth (SPER) is 2.68 eV in
the case of silicon [9] and 2.17 eV in the case of germanium [10]. In addition for the LKMC
mode, the recrystallization velocity depends on the substrate orientation with approximate
ratios of 20:10:1 for the orientations (100), (110), and (111), respectively, for both silicon [11]
and germanium [6].

The selected LKMC model for (100) SPER is the Planar one differentiating high and low
coordination [12]. To enable the optional LKMC model, you must call the following command:
pdbSet KMC Si Damage SPER.Model LKMC

In the presence of n-type and p-type doping, the recrystallization velocity is enhanced, again
following the literature [9][13][14]. The effect of specific impurities on the recrystallization is
described in Recrystallization on page 214.

During recrystallization, impurities and dopants can be redistributed by the recrystallization

front. The selected redistribution model during recrystallization is the so-called Hops model.
The amount of redistribution is impurity dependent or doping dependent, and is described in
Recrystallization on page 214.

Amorphous Silicon and Germanium

The defects in amorphous silicon, which are the dangling bonds of threefold-coordinated Si
atoms and the floating bonds of fivefold-coordinated atoms, are simulated explicitly by
enabling the amorphous.bonds model for silicon damage, which follows the
literature [15][16].

The dangling bond and floating bond density after amorphization is set to 1.8%, and their
diffusivities are assumed to be the same with a migration barrier of 2.6 eV, both as proposed
by [15]. Moreover, the dangling bonds can be annihilated with floating bonds and bind to
impurity atoms. For details about the interaction of dangling bonds with B atoms, see Boron
on page 200.

For simplicity, the same parameters are assumed for amorphous germanium. However, the
verification of this assumption has not been possible due to a lack of corresponding
experimental data for amorphous germanium.

186 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

Diffusion, Generation, and Recombination

For silicon, the transport capacity for free interstitials:
Di*Cistar = 1.59e25 exp(-4.702eV/kT) cm-1 s-1

is inherited from the Advanced Calibration of continuum Sentaurus Process (see Bulk
Parameters for Free Interstitials on page 18).

The macroscopic diffusivity of continuum Sentaurus Process corresponds to the microscopical

diffusivities of each charge state of Sentaurus Process KMC in the following way:
Di = (D(I0)C(I0)+D(I+)C(I+)+D(I-)C(I-))/(C(I0)+C(I+)+C(I-))

For simplicity, the same diffusivities for charge states –1, 0, +1 and lower ones for charge states
–2, +2 are assumed. Taking the diffusivity from the continuum Advanced Calibration (gained
from the high-temperature experiments of Bracht et al. [17]), you have:
Di = 51 * exp(-1.77eV/kT) cm2 s-1

which is reasonable for high temperatures, but too small at room temperature. Therefore, a
lower migration energy must be chosen [18], namely, 0.7 eV in the case of Advanced
Calibration. The migration prefactor has been adjusted to obtain the same diffusivity as for the
Advanced Calibration of continuum Sentaurus Process at 950°C . The resulting diffusivity is:
Di = 0.002 * exp(-0.7eV/kT) cm2 s-1

which is the better choice for the full temperature range.

Derived from the transport capacity and the diffusivity, the equilibrium concentration is:
Cistar = 1.59e25 * exp(-4.702eV/kT) / 0.002 / exp(-0.7eV/kT) cm-3
= 7.95e27 * exp(-4.002eV/kT) cm-3

The interstitial formation energy of 4.0 eV is in agreement with the literature [19].

In Sentaurus Process KMC, the interfaces set the equilibrium concentrations for point defects.
According to the literature [20], the following relation is valid:
Cistar = 2/alpha/alpha*6/lambda * D0FS * exp(-EForm/kT)

where the silicon lattice constant is alpha=5.43e-8cm, the jump distance is

lambda=3.84e-8cm, and the Sentaurus Process KMC point-defect interface model prefactor
is D0FS and the formation energy is EForm. Therefore, the parameters are:
Cistar = 2/alpha/alpha*6/lambda * 75000 * exp(-4.00eV/kT) cm-3

Advanced Calibration for Process Simulation User Guide 187

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

For germanium, the transport capacity for free interstitials:

Di*Cistar = 1.59e25 exp(-4.38eV/kT) cm-1 s-1

is inherited from the Advanced Calibration of continuum Sentaurus Process (see Bulk
Parameters for Free Interstitials on page 18) as well.

For germanium, there is a lack of experimental data on self-interstitial properties. Therefore,

only the interstitial diffusivity is estimated, based on the calculated migration energy (1.2 eV)
of uncharged interstitials calculated by Vanhellemont et al. [21] and based on the assumption
that the prefactor is the same as for silicon, for simplicity. The equilibrium concentration is
estimated, based on the formation energy (3.18 eV) calculated by Vanhellemont et al. [22] and
again based on the assumption that the prefactor is the same as for silicon, for simplicity.

For vacancies in silicon, the transport capacity is also inherited from the Advanced Calibration
of continuum Sentaurus Process (see Bulk Parameters for Free Vacancies on page 19):
DvCvstar = 4.60e22 * exp(-4.14eV/kT) cm-1 s-1

and the default diffusivity:

Dv = 5e-8 * exp(-0.4eV/kT) cm2 s-1

for neutral vacancies has been chosen with the migration barrier of 0.4 eV [18].

Therefore, the equilibrium concentration is:

Cvstar = 9.20e29 * exp(-3.74eV/kT) cm-3
= 2/alpha/alpha*6/lambda * 8.68e6 * exp(-3.74eV/kT) cm-3

Again, the vacancy formation energy of 3.74 eV is in agreement with the literature [19].

For vacancies in germanium, the transport capacity is inherited from the Advanced Calibration
of continuum Sentaurus Process (see Bulk Parameters for Free Vacancies on page 19) as well:
DvCvstar = 1.18e24 * exp(-3.09eV/kT) cm-1 s-1

The formation energy of the dominating double negative-charged vacancy of 2.87 eV is based
on [23]. Assuming the same prefactor as for silicon for simplicity, the equilibrium
concentration is:
Cvstar = 9.20e29 * exp(-2.87eV/kT) cm-3

The formation energy for the neutral vacancy is derived from the following relation [24]
depending on the local Fermi level eF and the ionization energies e0:
Ef(V0) = Ef(V--) + 2*eF - e0(V-) - e0(V--) = 2.87 + 0.74 - 0.3 - 0.1 = 3.21

188 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

The vacancy diffusivity of neutral charge and (single and double) negative charge is derived
from the vacancy equilibrium concentration and the transport capacity that is based on the
experimental vacancy-mediated self-diffusion coefficient of germanium
–2 –1
( 13.6 × exp ( – 3.09 eV/kT ) ) cm s following [25]. The resulting migration barrier (0.22 eV)
is also consistent with calculations by [23].

For the Si–SiO2 and Ge–GeO2 interfaces, the Allcharges model is selected to allow for the
emission and capture of all the charge states of point defects. The recombination length of point
defects at Si–SiO2 and Ge–GeO2 interfaces is assumed to be 1 nm for continuum and kinetic
Monte Carlo. The one at Si–gas and Ge–gas interfaces is set to 1 nm as well. Parameters for
Si–nitride and Ge–nitride interfaces are inherited from Si–SiO2 and Ge–GeO2 interfaces.

The bulk recombination of interstitials and vacancies in the Advanced Calibration of

continuum Sentaurus Process is diffusion limited. In Sentaurus Process KMC, interstitials and
vacancies form amorphous pockets when they are close (within capture radius) during their
migration. The interstitial and vacancy recombine, and the amorphous pocket is dissolved if no
more interstitials or vacancies are captured by the amorphous pocket beforehand. However, in
contrast to the continuum Advanced Calibration, the I–V recombination is not instantaneous
due to the energy barrier [26]. Therefore, the formation of I–V pairs is diffusion limited and the
recombination of I–V pairs in silicon in the equation recombination rate is:
v = 5.0e-4 * exp(-0.43eV/kT) cm2/s

The corresponding one for germanium is:

v = 5.0e-3 * exp(-0.60eV/kT) cm2/s

The generation of I–V pairs has not been implemented in Sentaurus Process KMC due to its
small contribution, whereas in continuum Advanced Calibration, the generation of I–V pairs is
simulated.

Oxidation causes the injection of interstitials at the exposed surface. The calibration of
Sentaurus Process KMC follows that for the continuum of interstitial injection for dry
oxidation of silicon (see Oxidation-Enhanced Diffusion on page 20). No interstitial injection
for oxidation of germanium is assumed.

Charge States
The charge levels for self-interstitials and vacancies have been studied [18]. Accordingly, the
concentration for positively charged interstitials is:
C(I+) = C(I0) * exp(-(Ef - e(I+))/kT)

Advanced Calibration for Process Simulation User Guide 189

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

where the Fermi level is Ef and the energy level is e(I+), which is the energy needed to take
an electron from the neutral interstitial measured from the valence band (Sentaurus Process
KMC parameter e0(IP)).

In continuum Sentaurus Process, the concentration for positively charged interstitials is:
C(I+) = C(I0) * k0(I+) exp(-(Ef - Ei + kE(I+))/kT)

where the Fermi level is Ef, the intrinsic level is Ei, and the charge-state parameter is
k(I+)=k0(I+)*exp(-kE(I+)/kT).

Therefore, the following relation is valid:

exp(e(I+))/kT) = k0(I+) exp((Ei - kE(I+))/kT) (Eq x)

The intrinsic level depends on the band gap Eg, and the effective state density of the conduction
(Nc) and valence (Nv) bands [27]:
Ei(T) = Eg(T)/2 + kT/2 * ln(Nv/Nc)

In summary, the electronic levels in continuum Sentaurus Process refer to the intrinsic level,
while the electronic levels in atomistic Sentaurus Process are measured from the valence band
edge. This means that due to the temperature dependence of the energy difference of the
intrinsic and valence band levels, an exact translation of the charge-state parameters from
continuum Sentaurus Process to Sentaurus Process KMC is not possible.

Therefore, the electronic levels for interstitials and vacancies in silicon are taken as
recommended in the literature [18]. In the temperature range from 700°C to 1100°C , the
electronic level for I+ of 0.4 eV corresponds to k0(I+)=12 and kE(I+)=0.23eV, and the one
for I– of 1.0 eV corresponds to k0(I-)=0.12 and kE(I-)=0.4eV.

The electronic levels for interstitials and vacancies in germanium are calibrated with guidance
from the literature [22][23].

Extended Defects
In Sentaurus Process KMC, the evolution of extended defects of self-interstitials from small
clusters to {311} defects and faulted dislocation loops is taken into account [28]. Small clusters
have irregular shapes and are amorphous pockets in the terminology of kinetic Monte Carlo.
Amorphous pockets capture any point defect (I and V) within their capture radius.

For silicon, the binding energies are taken from the parameter array Eb.Cluster and are
based on the suggestions of [29] assuming a self-interstitial formation energy of 3.8 eV (see
Table 9 on page 191). The interstitial emission prefactor D0.Cluster is 135.

190 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

Table 9 Binding energies in keV for small interstitial clusters in silicon

N= 2 3 4 5 6 7 8 9 10 11 12 13 14

Cowern [29] 2.64 2.77 2.67 2.48 2.46 3.24 2.40 2.69 2.73 2.92 3.00 3.06

Advanced 2.40 2.50 2.77 2.67 2.48 2.46 3.24 2.40 2.69 2.73 2.92 3.00 3.06
Calibration

Above a threshold, the extended defects in silicon form rod-like {311} defects. The minimal
size of a {311} Min.AP.To.311.Size has been chosen to be 15. {311} defects capture any
neutral interstitial with which they are in contact. The binding energy of this size is 2.94 keV,
the one of size 60 is 2.96 keV, and the ones for the intermediate sizes are linearly interpolated.
Above the size of 60, the binding energies of size n are computed using the following equation:
Eb(n) = Eb_L - (Eb_L - Eb_S) * (n^a -(n-a)^a)/(2^a-1)

where Eb_L (Eb.LargeCluster) is equal to 3.17 keV, Eb_S (Eb.SmallestCluster) is

equal to 2.64 keV, and a (Exponent.Cluster) is equal to 3/4. The interstitial emission
prefactor is the same as for amorphous pockets.

When {311} defects in silicon grow large enough, they transform into dislocation loops. The
threshold size for this transformation Min.311.To.111Loop.Size is equal to 4500. In this
calibration, dislocation loops cannot be formed directly from amorphous pockets, as the
prefactor D0.AP.To.111Loop is set to 0.

Dislocation loops capture any incoming neutral interstitial. The binding energies are:
Eb(n) = Ef(I) + Ef(DL(n-1)) - Ef(DL(n))

with the dislocation loop formation energies:

Ef(DL(n)) = π γ R^2 + a^2 μ /6(1- υ )R * log(8R/b) -nEf(I)

taken from the literature [30], which are Sentaurus Process KMC defaults. The interstitial
emission prefactor D0.Loop is 1e6 in silicon.

In germanium, small clusters and extended defects consisting of interstitials also have been
observed [31][32][33][34][35]. As for silicon, Sentaurus Process KMC takes the binding
energies from the parameter array Eb.Cluster and calculates the binding energies for defects
above the size of 60. The binding energies and the interstitial emission prefactor D0.Cluster
of 5 have been calibrated against literature data [33].

According to the literature [34][35], the observed extended defects in germanium are not {311}
defects, but small dislocation loops. To account for this, the transformation of amorphous
pockets to {311} defects is suppressed by setting D0.AP.To.311 to 0, and the transformation
to dislocation loops is allowed with D0.AP.To.111Loop equal to 1e12 and
E.AP.To.111Loop equal to 1.3 eV in germanium. The minimal size of a dislocation loop

Advanced Calibration for Process Simulation User Guide 191

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

Min.AP.To.111Loop.Size has been chosen to be 33. Finally, the dislocation loop formation
energies and the interstitial emission prefactor D0.Loop of 1e6 also have been calibrated
against literature data [33].

The binding energies for small vacancy clusters up to size 10 in silicon are taken from the
ATOMICS research project [36]. The vacancy emission prefactor D0.Cluster is 5. Voids are
assumed to be vacancy clusters of size 20 and higher.

For germanium, the binding energies for small vacancy clusters up to size 4 are based on the
literature [37]. While the binding energies of larger clusters are inherited from the silicon
calibration. The vacancy emission prefactor D0.Cluster is 5e-2. Again, voids are assumed to
be vacancy clusters of size 20 and higher.

SiGe

Linear Germanium Correction Factors

By default, the Ge effects are modeled by linear correction factors for silicon parameters. The
calibration is valid for low Ge mole fractions ranging between 0% and 50%.

Amorphization and Recrystallization

It is assumed that there is a reduced interstitial–vacancy recombination rate during ion

implantation and annealing in SiGe compared to pure silicon, according to [38]. This is
performed by introducing a Ge effect for interstitial–vacancy recombination in amorphous
pockets:
pdbSet KMC Si Damage Eb_AmorphousPocketGe [expr 1.50/5.e22]

Band Gap

Germanium reduces the band gap of silicon. The formula for Ge-induced bandgap narrowing
for Ge mole fractions below 85%, which is identical to the one used for continuum Sentaurus
Process (see Effect of Germanium and Stress on page 38), is:
dE(BandGap) = [Ge]/5.e22*(0.33*([Ge]/5.e22)-0.55) eV/cm3

and is implemented by defining:

pdbSet KMC Si BandGap GeNarrowing [expr -0.55/5.e22]
pdbSet KMC Si BandGap GeNarrowing2 [expr 0.33/5.e22/5.e22]

192 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

Diffusion, Generation, and Recombination

In the literature, there is no agreement on the impact of Ge (for low Ge mole fractions) on point-
defect diffusion and generation except for the vacancy equilibrium concentration.
Computational studies of vacancies in SiGe have found an ~1.0 eV vacancy formation energy
decrease in Ge [39][40][41], in line with the Advanced Calibration parameters for silicon and
germanium. For computational efficiency, to prevent the abundance of vacancies in SiGe, a
value lower than the published ones is used in the Advanced Calibration of Sentaurus Process
KMC, also lower than the one for continuum Advanced Calibration (see Impact of Stress on
Point-Defect Parameters on page 48):
pdbSet KMC Si Vac EfGe V [expr -0.50/5.e22]

No Ge effect on the interstitial equilibrium concentration is assumed. Based on molecular

dynamics calculations [42], the diffusivities of interstitials and vacancies are almost unchanged
for low Ge mole fractions, but they change rapidly for higher ones. Therefore, no linear Ge
effect is used for the interstitial and vacancy diffusivity.

Extended Defects

The {311} interstitial clusters are less stable in the presence of Ge, and the transformation of
{311} defects into dislocation loops is faster [36][43][44]. The corresponding parameter
adjustments have been calibrated by Synopsys based on experimental data generated within the
ATOMICS research project [2].

The binding energies of small interstitial clusters, {311} defects, and dislocation loops are
lowered in the presence of Ge:
pdbSet KMC Si Int Eb_ClusterGe [expr -0.20/5.e22]
pdbSet KMC Si Int Eb_311Ge [expr -0.20/5.e22]
pdbSet KMC Si Int Eb_LoopGe [expr -0.50/5.e22]

The transformation from {311} defects to dislocation loops occurs earlier in the presence of
Ge:
pdbSet KMC Si Int E_311toLoopGe [expr -1.00/5.e22]

The binding energies of small vacancy clusters and voids are higher in the presence of Ge:
pdbSet KMC Si Vac Eb_ClusterGe [expr 0.20/5.e22]
pdbSet KMC Si Vac Eb_VoidGe [expr 0.20/5.e22]

Silicon and Germanium Parameter Interpolation

Optionally, you can enable Si and Ge parameter interpolation for SiGe using the command:
KMC_SiGe_and_Stress_Effect 1 0

Advanced Calibration for Process Simulation User Guide 193

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

This disables the default linear correction factors and uses the interpolation parameters defined
for SiliconGermanium. The calibration covers the full Ge mole fraction range (0–100%).

In general, prefactor parameters are interpolated in logarithmic scale, and energy parameters
are interpolated in linear scale.

Amorphization and Recrystallization

According to [38], the interstitial–vacancy recombination rate during ion implantation and
annealing in SiGe is reduced, compared to the rate in pure silicon. This is in line with Advanced
Calibration, where the interstitial–vacancy recombination rate at room temperature for small
amorphous pockets is ~1000 times smaller in germanium compared to silicon. The SiGe
interpolation for the interstitial–vacancy recombination rate prefactor is parabolic and is
calibrated based on a limited experimental dataset for Ge implantation into Ge only [45][46].
Parabolic SiGe parameter interpolation is used for the displacement threshold in B
implantation, which has been calibrated against experimental data from AMAT-VSE.

The (100) recrystallization of germanium is much faster than the one of silicon. For SiGe, the
recrystallization activation energy is interpolated parabolically with slightly higher or constant
values for Ge mole fractions up to 40% and lower values for higher Ge mole fractions,
following [47]. The KMC and all LKMC recrystallization prefactors for the different
orientations are interpolated parabolically (in logarithmic scale) to match the regrowth rates of
SiGe of different Ge mole fractions [46][48]. As a result, the recrystallization velocity is faster
for SiGe of all Ge mole fractions compared to silicon.

Band Gap

Germanium reduces the band gap of silicon. The energy bandgap change along with the Ge
mole fraction shows a very nonlinear curve due to the transition between the X-valley ( x < 0.8 )
and the L-valley ( x > 0.8 ) [49]. With SiGe parameter interpolation enabled, the Ge-induced
bandgap narrowing model is used, which is identical to the one used for continuum Sentaurus
Process (see Impact of Germanium on Electrostatic Potential on page 41) and is implemented
by defining:
pdbSet Si KMC BandGap Alloy.Narrowing.Model
pdbSet Si KMC BandGap Eg0 1.170
pdbSet Ge KMC BandGap Eg0 0.719

Diffusion, Generation, and Recombination

For vacancies, both equilibrium concentration and diffusivity are higher in germanium than in
silicon. However, for interstitials, only the equilibrium concentration is higher, but the
diffusivity is lower in germanium than in silicon. Parabolic SiGe parameter interpolation in
linear scale is used for both the formation energies and the migration barriers of point defects,
while prefactors are interpolated linearly in logarithmic scale.

194 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

In addition, electronic levels at T=0 K are interpolated linearly between silicon and
germanium.

The interpolation between silicon and germanium was calibrated based on SiGe interdiffusion
experiments (see SiGe Interdiffusion). Following Castrillo et al. [50], vacancies with a double
negative charge are the dominant contributor to the point-defect transport capacity starting at
medium mole fractions (~0.3) up to mole fraction 1.

NOTE SiGe oxidation is not accurately modeled by KMC. Since Sentaurus

Process KMC does not allow for point-defect injection correction
(needed for vacancies) and for parameter interpolation (needed for
OED), SiGe interdiffusion during oxidation is incorrect and, therefore,
the Ge mole-fraction oxidation rate is incorrect as well (see the
continuum model in SiGe Oxidation on page 46).

Extended Defects

The {311} interstitial clusters are less stable with higher Ge mole fractions, and the
transformation of amorphous pockets and {311} defects into dislocation loops is faster
[36][43][44]. Parabolic SiGe parameter interpolation is used for small and {311} defect
binding energies, the minimum {311} cluster size, and the interstitial emission prefactor from
dislocation loops. Other point-defect cluster parameters are interpolated linearly.

SiGe Interdiffusion

In undoped SiGe, the interdiffusion of Si and Ge atoms is a consequence of point-defect

diffusion [50] (see SiGe Interdiffusion on page 43). Contributions from a direct exchange
mechanism are neglected. SiGe interdiffusion is described as the sum of a contribution from
vacancy diffusion and a contribution from interstitial diffusion. The KMC interdiffusion model
accounts for different Si and Ge self-diffusivities by defining the relative probability of an I (or
a V) moving a Ge or Si atom [50][51][52]. Therefore, changes to the interstitial and vacancy
transport capacity due to excess point defects after implantation, or due to Fermi-level effects
in the case of doping, directly lead to altered SiGe interdiffusion. In addition, when taking into
account stress effects in KMC (see Stress Effects on page 196), interdiffusion is naturally stress
dependent because the diffusivity and equilibrium concentration of point defects are modified
in strained SiGe as a function of pressure.

The parameters for SiGe interdiffusion have been extracted and calibrated against data from
the literature [53][54][55][56][57][58]. While the (dominant) interdiffusivity from vacancies
can be extracted reliably, the parameter error might be larger for the smaller component of
interdiffusivity from the diffusion of interstitials. The biggest uncertainties in calibration exist
for highly n-doped SiGe, where interdiffusion is believed to have a dominant contribution from
P–V or As–V pairs, which is not yet included in the calibration, due to a lack of experimental
data suitable for calibration.

Advanced Calibration for Process Simulation User Guide 195

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

Stress Effects
The stress effects on damage, diffusion, and binding in Advanced Calibration for Sentaurus
Process KMC are only taken into account if the KMC Stress switch is enabled. This is
accomplished by the command:
KMC_SiGe_and_Stress_Effect 0 1

The procedure KMC_SiGe_and_Stress_Effect has two Boolean arguments: the first is

Ge_Chem_Eff and the second is Stress_Eff, which is dedicated for stress effects.

Alternatively, the KMC Stress switch can be enabled directly:

pdbSet KMC Stress 1

The calibration of stress effects [1][2] is based on a relatively small set of measured data. The
models have not yet been tested rigorously against SIMS data for all dopant species and
nonequilibrium annealing conditions, and against electrical data from real-device fabrication
processes.

Amorphization and Recrystallization

No stress effect on damage accumulation is implemented in this calibration. Stress effects on

damage accumulation are not clearly experimentally evidenced and, at least, appear not to be
the primary cause of enhanced damage production in the case of implantation into strained
SiGe compared to relaxed Si [59].

The stress dependency of SPER is calibrated only for LKMC. The stress effect on in-plane
uniaxial stresses is nonlinear and is simulated by a model assuming that {100} events occur
through a dual-timescale atomistic mechanism [12][60]. In addition, compressive hydrostatic
pressure results in an enhancement of the (100) regrowth velocity [60].

Band Gap
The bandgap narrowing for silicon due to stress is calculated according to the deformation
potential theory as for continuum Sentaurus Process (see Impact of Stress on Electrostatic
Potential on page 47). The deformation potential constants for Si and Ge proposed by Van de
Walle were chosen for Advanced Calibration. For strained SiGe, the model uses a simple linear
interpolation between Si and Ge. The model is enabled by:
pdbSet KMC Si BandGap FullNarrowing 1

196 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 1: Model Parameters for Implantation Damage and Point Defects

Diffusion, Generation, and Recombination

Compressive hydrostatic pressure decreases the equilibrium concentration of interstitials and

increases the equilibrium concentration of vacancies. Similarly, the diffusivities of point
defects are modified as a function of pressure. The activation volumes for the above prefactors
were derived from the induced strain values calculated by Diebel [61][62], and the value for
pure germanium has been adjusted to match experimental data reported by Kawamura et al.
[63], in analogy to continuum Advanced Calibration (see Impact of Stress on Point-Defect
Parameters on page 48).

In addition, interstitials show an anisotropy in the diffusivity as a function of biaxial strain

according to [62]. Biaxial tension (as in strained silicon on SiGe) leads to significantly higher
in-plane diffusion compared to the perpendicular direction.

The diffusivities and equilibrium concentrations of interstitials and vacancies in silicon,

germanium, and SiGe are modified in the presence of stress in the following way:
pdbSetArray KMC Si Int VD { IMM -0.00014,-0.00007
IM -0.00014,-0.00007
I -0.00014,-0.00007
IP -0.00014,-0.00007
IPP -0.00014,-0.00007 }
pdbSetArray KMC Ge Int VD { IMM -0.00014,-0.00007
IM -0.00014,-0.00007
I -0.00014,-0.00007
IP -0.00014,-0.00007
IPP -0.00014,-0.00007 }
pdbSetArray KMC Si Vac VD { VMM 0.0012,0.0012
VM 0.0012,0.0012
V 0.0012,0.0012
VP 0.0012,0.0012
VPP 0.0012,0.0012 }
pdbSetArray KMC Ge Vac VD { VMM 0.0008,0.0008
VM 0.0008,0.0008
V 0.0008,0.0008
VP 0.0008,0.0008
VPP 0.0008,0.0008 }
pdbSetArray KMC SiGe Vac VD.X2 { VMM 0.0019,0.0019
VM 0.0019,0.0019
V 0.0019,0.0019
VP 0.0019,0.0019
VPP 0.0019,0.0019 }
pdbSet KMC Si Int VF I -0.0012
pdbSet KMC Ge Int VF I -0.0028
pdbSet KMC Si Vac VF V 0.0013
pdbSet KMC Ge Vac VF V 0.0013

Advanced Calibration for Process Simulation User Guide 197

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Extended Defects

The {311} interstitial clusters are less stable in the presence of compressive hydrostatic
pressure, and the transformation of {311} defects into dislocation loops is faster [36][43][64].

The binding energies of small interstitial clusters, {311} defects, and dislocation loops are
lowered in the presence of compressive hydrostatic pressure:
pdbSet KMC Si Int VFCluster -0.01
pdbSet KMC Si Int VF311 -0.01
pdbSet KMC Si Int VFLoop -0.02

The transformation from {311} defects to dislocation loops occurs earlier in the presence of
compressive hydrostatic pressure:
pdbSet KMC Si Int VF311toLoop -0.04

The binding energies of small vacancy clusters and voids are not dependent on stress in this
calibration.

Part 2: Model Parameters for Impurities

This part describes the model parameters for impurities for the materials silicon, germanium,
and SiGe.

Diffusion
The effective diffusivity of an interstitial-mediated and a vacancy-mediated dopant X in
crystalline silicon or germanium is given by the sum of the contribution of all mobile species
and can be reduced in continuum models using the Maxwell–Boltzmann approximation to:
D(X-) = Si [D(X-I0) + D(X-I+)(p/ni) + D(X-I++)(p/ni)^2] +
Sv [D(X-V0) + D(X-V+)(p/ni) + D(X-V++)(p/ni)^2]

for the negatively charged immobile substitutional dopant X– and:

D(X+) = Si [D(X+I0) + D(X+I-)(n/ni) + D(X+I--)(n/ni)^2] +
Sv [D(X+V0) + D(X+V-)(n/ni) + D(X+V--)(n/ni)^2]

for the positively charged immobile substitutional dopant X+, where Si is the interstitial
supersaturation, Sv is the vacancy supersaturation, and n, p, and ni are the electron, hole, and
intrinsic concentration, respectively.

198 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

According to [18], the relation between the above diffusivity components and the
microscopical parameters is:
D(X-I0) = Uc * DiCi* * 1/Vbk(Xi-) * Vm(Xi-)
D(X-I-) = Uc * DiCi* * 1/Vbk(Xi-) * Vm(Xi0) * exp((e0(Xi-)-ei)/kT)
D(X-V0) = Uc * DvCv* * 1/Vbk(Xv-) * Vm(Xv-)
D(X-V-) = Uc * DvCv* * 1/Vbk(Xv-) * Vm(Xv0) * exp((e0(Xv-)-ei)/kT)

and:
D(X+I0) = Uc * DiCi* * 1/Vbk(Xi+) * Vm(Xi+)
D(X+I-) = Uc * DiCi* * 1/Vbk(Xi+) * Vm(Xi0) * exp((ei-e0(Xi+))/kT)
D(X+V0) = Uc * DvCv* * 1/Vbk(Xv+) * Vm(Xv+)
D(X+V-) = Uc * DvCv* * 1/Vbk(Xv+) * Vm(Xv0) * exp((ei-e0(Xv+))/kT)

where the effective capture volume is Uc = 3*lambda^3 (lambda=3.84nm is the distance

in the silicon lattice from one atom to its second-nearest neighbor), the breakup frequencies are
Vbk, the migration frequencies are Vm, the electronic levels are e0, and the intrinsic Fermi level
is ei (at a given temperature). The breakup frequencies are:
Vbk(Xi-) = Db(Xi-) * exp(-(Eb(Xi-)+Em(I0))/kT)
Vbk(Xv-) = Db(Xv-) * exp(-(Eb(Xv-)+Em(V0))/kT)
Vbk(Xi+) = Db(Xi+) * exp(-(Eb(Xi+)+Em(I0))/kT)
Vbk(Xv+) = Db(Xi+) * exp(-(Eb(Xv+)+Em(V0))/kT)

where the migration energies are Em, the binding prefactor is Db, and the energy is Eb.

This means that the diffusivities of Sentaurus Process Advanced Calibration can be translated
to a fixed relation of migration and binding parameters for Sentaurus Process KMC: The bigger
the binding energy, the bigger the migration energy to obtain the same diffusivity. For this
calibration of Sentaurus Process KMC, the binding prefactor and energy as well as the
electronic levels were first chosen reasonably, and the migration prefactors and energies were
deduced. Then, the optimal combination of binding energy and electronic levels was searched,
based on SIMS calibration, while the migration energies were shifted accordingly to keep the
same diffusivity.

Since kinetic Monte Carlo can also simulate diffusion during implantation at room
temperature, the consideration of as-implanted SIMS data is crucial for calibration of the
binding energies. For the situation of full cascade implantation, many impurities pair with the
abundant point defects because of their high implant energy and, therefore, they can diffuse
even at room temperature if the migration energies are small. Higher binding energies,
implying higher migration energies, can suppress migration at room temperature.

Advanced Calibration for Process Simulation User Guide 199

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Boron

For silicon, the boron diffusivities of Pichler [65], on which also the continuum Advanced
Calibration Version P-2019.03 is based (see Boron Diffusion Coefficient on page 22):
D(B-I0) = 0.123 * exp(-3.566eV/kT)
D(B-I+) = 4.210 * exp(-3.671eV/kT)
D(B-I++) = 39.8 * exp(-4.373eV/kT)

translate to microscopical diffusivities:

Vm(Bi-) = 4.55e-5 * exp(-0.364eV/kT)
Vm(Bi0) = 2.82e-3 * exp(-0.707eV/kT)
Vm(Bi+) = 1.29e-1 * exp(-1.912eV/kT)

where the binding prefactors equal 1.0 and the energies equal 0.8 eV for B– plus a neutral
interstitial and 0.1 eV for B– plus a vacancy and, for the following electronic levels for
boron–interstitial pairs and boron–vacancy pairs:
pdbSetDoubleArray KMC Si B e0 {BiM 0.8
BiP 1.04
BVM 0.1
BVP -5.0}

The electronic levels at T=0 for the boron–interstitial pairs with a negative and a positive charge
are taken from the literature [18], while the ones for boron–vacancy pairs are calibrated. The
intrinsic Fermi level, E fi ( T ) = E g ( T ) ⁄ 2 – 0.5 ln ( N c ⁄ N v ) k B T , is used for the translation of
the diffusivity from continuum to KMC simulation.

The binding energies for the negative-charged boron–interstitial and boron–vacancy pairs are
derived from the following relation [18]:
Eb(Bi-) = Eb(Bi0) - e0(Bi-) + e0(I+) = 1.20 - 0.80 + 0.40 = 0.8
Eb(BV-) = Eb(BV0) - e0(BV-) + e0(V+) = 0.17 - 0.10 + 0.03 = 0.1

with a binding energy of a neutral boron–interstitial pair of 1.2 eV that is close to the reported
value of 1.1 eV [66], and with a binding energy of a neutral boron–vacancy pair of 0.17 eV in
agreement with the literature [67].

The substitutional B– as well as the B– paired with any V are immobile.

In addition to diffusion in crystalline silicon, Advanced Calibration enables boron diffusion in

amorphous silicon mediated by dangling bonds [15][16]. The microscopical diffusivity is set
to:
Vm(Bi) = 72* exp(-2.8/kT)

200 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

where the binding prefactors equal 1.0, and the energies equal 0.2 eV [66] for B plus a dangling
bond. The coefficient for dangling bond creation per B atom in amorphous silicon is set to 0.5
to obtain good accuracy with SIMS for boron implantation into amorphous silicon
[16][68][69]. However, for amorphization of boron marker layers, a coefficient of 1.0 gives
better accuracy, as reported in [15].

For germanium, the boron diffusivities of Uppal [70], on which also the continuum Advanced
Calibration Version P-2019.03 is based (see Boron Diffusion Coefficient on page 22):
D(B-I0) = 1.77e5 * exp(-4.65eV/kT)
D(B-I+) = 1.00e4 * exp(-4.65eV/kT)
D(B-I++) = 1.00e4 * exp(-4.65eV/kT)

translate to microscopical diffusivities:

Vm(Bi-) = 65.5 * exp(-2.87eV/kT)
Vm(Bi0) = 3.70 * exp(-2.70eV/kT)
Vm(Bi+) = 3.70 * exp(-2.73eV/kT)

where the binding prefactors equal 1.0, and the energies equal 1.4 eV for B– plus a neutral
interstitial, 0.1 eV for B– plus a vacancy, and for the following electronic levels for
boron–interstitial pairs and boron–vacancy pairs:
pdbSetDoubleArray KMC Ge B e0 {BiM 0.2
BiP 0.4
BVM 0.1
BVP -5.0}

The electronic levels and binding energies for the boron–interstitial and boron–vacancy pairs
are calibrated. The substitutional B– as well as B– paired with any V are immobile in
germanium, similar as in silicon.

Arsenic

For silicon, the arsenic diffusivities are close to the diffusivities reported by Martin-Bragado et
al. [71], which are not the same but are similar to continuum Advanced Calibration Version
P-2019.03 (see Arsenic Diffusivity on page 26):
D(As+I0) = 0.1350 * exp(-3.64eV/kT)
D(As+I-) = 4200.0 * exp(-4.80eV/kT)
D(As+V0) = 0.0780 * exp(-3.70eV/kT)
D(As+V-) = 10.900 * exp(-4.24eV/kT)
D(As+V--) = 7.500 * exp(-4.80eV/kT)

corresponding to microscopical diffusivities:

Vm(Asi+) = 5.0e-5 * exp(-0.40eV/kT)
Vm(Asi0) = 1.6 * exp(-1.79eV/kT)

Advanced Calibration for Process Simulation User Guide 201

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Vm(AsV+) = 3.8e-4 * exp(-1.90eV/kT)

Vm(AsV0) = 5.4e-2 * exp(-2.51eV/kT)
Vm(AsV-) = 9.8 * exp(-1.04eV/kT)

where the binding prefactor equals 1.0 and the energy equals 0.76 eV for As+ plus a neutral
interstitial, and the binding prefactor equals 0.4 eV and the energy equals 1.34 eV for As+ plus
a neutral vacancy, for the following electronic levels:
pdbSetDoubleArray KMC Si As e0 {AsiP 0.2
AsiM 5.0
AsVP 0.6
AsVM 1.1}

The electronic level at T=0 for arsenic–interstitial pairs and arsenic–vacancy pairs is calibrated.
Compared to [71], the binding prefactors are doubled. In addition, the binding energy of the
arsenic–interstitial pair is recalibrated based on low-temperature diffusion experiments. In
agreement with [71], the binding energy for the arsenic–vacancy pair is based on [72] and is
derived from the following relation [18]:
Eb(AsV+) = Eb(AsV0) + e0(AsV+) - e0(V-) = 1.34 + 0.6 - 0.6 = 1.34

leading to a binding energy of As+ plus a neutral vacancy that is close to another reported
value [73].

The substitutional As+ as well as As+ paired with I–– are immobile.

According to measurements by Larsen et al. [74], the diffusivity of As increases sharply for
20 –3
regions with high As concentration (> 2.0 × 10 cm ). This effect has been attributed to
As–V percolation, which is not simulated with Advanced Calibration for Sentaurus Process
KMC. Another explanation is enhanced diffusion due to mobile clusters, in particular, As2V
[71], as discussed in Arsenic on page 201.

For germanium, the arsenic diffusivities, dominated by As+V– – pairs, are based on Zographos
and Erlebach [75] and are identical to the diffusivities of the continuum Advanced Calibration
Version P-2019.03 (see Arsenic Diffusivity on page 26):
D(As+I0) = 0.1350 * exp(-3.64eV/kT)
D(As+I-) = 4200.0 * exp(-4.80eV/kT)
D(As+V0) = 0.0780 * exp(-3.70eV/kT)
D(As+V-) = 10.900 * exp(-4.24eV/kT)
D(As+V--) = 2.0e3 * exp(-2.90eV/kT)

corresponding to microscopical diffusivities:

Vm(Asi+) = 5.0e-5 * exp(-1.22eV/kT)
Vm(Asi0) = 1.6 * exp(-2.55eV/kT)
Vm(AsV+) = 3.8e-4 * exp(-2.17eV/kT)

202 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Vm(AsV0) = 5.4e-2* exp(-2.78eV/kT)

Vm(AsV-) = 9.8 * exp(-1.31eV/kT)

where the binding prefactors equal 1.0, and the energy equals 0.76 eV for As+ plus a neutral
interstitial, 1.68 eV for As+ plus a neutral vacancy, and for the following electronic levels:
pdbSetDoubleArray KMC Ge As e0 {AsiP 0.2
AsiM 5.0
AsVP 0.3
AsVM 0.5}

The electronic levels for arsenic–vacancy pairs are based on [76]; those for arsenic–interstitial
pairs are inherited from silicon. The binding energy for the arsenic–vacancy has been
calibrated and is in the range of reported values [76][77]. The binding energy for
arsenic–interstitial pairs are again inherited from silicon.

The substitutional As+ as well as As+ paired with I– – are immobile, as in silicon.

Phosphorus

For silicon, the phosphorus diffusivities, with their interstitial-mediated part taken from
Pichler [65]:
D(P+I0) = 0.453 * exp(-3.482eV/kT)
D(P+I-) = 1.610 * exp(-3.647eV/kT)
D(P+V0) = 7.8e-5 * exp(-3.594eV/kT)
D(P+V-) = 7.8e-6 * exp(-3.385eV/kT)
D(P+V--) = 1.6e-3 * exp(-3.469eV/kT)

translate to microscopical diffusivities, with phosphorus–vacancy pair migration barriers taken

from Sahli et al. [72]:
Vm(Pi+) = 1.68e-4 * exp(-1.310eV/kT)
Vm(Pi0) = 5.96e-4 * exp(-1.805eV/kT)
Vm(PV+) = 1.00e-6 * exp(-1.194eV/kT)
Vm(PV0) = 1.00e-6 * exp(-1.115eV/kT)
Vm(PV-) = 2.00e-4 * exp(-0.939eV/kT)

where the binding prefactors equal 1.0 and the energies equal 1.83 eV for P+ plus a neutral
interstitial and equal 1.34 eV for P+ plus a neutral vacancy and, for the following electronic
levels for the phosphorus interstitials and vacancy pairs:
pdbSetDoubleArray KMC Si P e0 {PiP 0.1
PiM 5.0
PVP 0.3
PVM 0.69}

Advanced Calibration for Process Simulation User Guide 203

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

The electronic levels at T=0 for the phosphorus–interstitial or phosphorus–vacancy pairs are
based on the literature [72].

The binding energies for the phosphorus–interstitial and phosphorus–vacancy pairs are based
on [72], and are derived from the following relation [18]:
Eb(Pi+) = Eb(Pi0) - e0(Pi+) + e0(I-) = 0.83 - 0.10 + 1.0 = 1.73
Eb(PV+) = Eb(PV0) - e0(PV+) + e0(V-) = 1.04 - 0.30 + 0.6 = 1.34

However, a small adjustment of +0.1 eV to the phosphorus–interstitial binding energy was

selected.

The substitutional P+ is immobile as well as the P+ is paired with I––.

For germanium, the phosphorus diffusivities, dominated by P+V–– pairs based on Zographos
and Erlebach [75] and with phosphorus–interstitial contribution inherited from silicon:
D(P+I0) = 0.453 * exp(-3.482eV/kT)
D(P+I-) = 1.610 * exp(-3.647eV/kT)
D(P+V0) = 1.0e-4 * exp(-3.000eV/kT)
D(P+V-) = 1.0e-4 * exp(-3.000eV/kT)
D(P+V--) = 2.0e-3 * exp(-2.100eV/kT)

translate to microscopical diffusivities:

Vm(Pi+) = 1.68e-4 * exp(-1.63eV/kT)
Vm(Pi0) = 5.96e-4 * exp(-2.07eV/kT)
Vm(PV+) = 4.90e-7 * exp(-1.34eV/kT)
Vm(PV0) = 4.90e-7 * exp(-1.61eV/kT)
Vm(PV-) = 9.80e-6 * exp(-0.48eV/kT)

where the binding prefactors equal 1.0, and the energies equal 1.33 eV for P+ plus a neutral
interstitial, equal 1.55 eV for P+ plus a neutral vacancy, and for the following electronic levels
for the phosphorus interstitials and vacancy pairs:
pdbSetDoubleArray KMC Ge P e0 {PiP 0.10
PiM 5.00
PVP 0.10
PVM 0.60}

The electronic levels for phosphorus–vacancy pairs are based on [76]; those for
phosphorus–interstitial pairs are inherited from silicon. The binding energy for the
phosphorus–vacancy has been calibrated and is in the range of reported values [76][77]. The
binding energy for phosphorus–interstitial pairs has been calibrated.

The substitutional P+ as well as P+ paired with I– – are immobile, as in silicon.

204 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Indium

For silicon, the diffusivities of Pichler [65]:

D(In-I0) = 3.13 * exp(-3.668eV/kT)
D(In-I+) = 6.45 * exp(-3.752eV/kT)

translate to microscopical diffusivities:

Vm(Ini-) = 1.15e-3 * exp(-0.426eV/kT)
Vm(Ini0) = 2.39e-3 * exp(-0.730eV/kT)

where the binding prefactors equal 1.0, and the energies equal 0.76 eV for In plus a neutral
interstitial and equal 1.25 eV for In plus a vacancy, for the following electronic levels:
pdbSetDoubleArray KMC Si In e0 {IniM 0.7
IniP -5.0
InVM 1.0
InVP -5.0}

The electronic levels at T=0 for the indium–interstitial and indium–vacancy pairs are
calibrated.

The substitutional In– as well as the In– paired with I++ or any V are immobile.

NOTE Advanced Calibration for Sentaurus Process KMC does not include any
parameters for indium in germanium.

Carbon

For silicon, the diffusivity of continuum Advanced Calibration Version P-2019.03 (see Carbon
Diffusion on page 34):
D(C-I0) = 6.11 * exp(-3.293eV/kT)

translates to microscopical diffusivities:

Vm(Ci-) = 2.26e-3 * exp(-0.791eV/kT)

where the binding prefactor equals 1.0 and the binding energy equals 1.5 eV for C plus a neutral
interstitial [78]. In addition, C–V pairing is allowed with the binding prefactor equal to 1.0 and
the binding energy equal to 0.25 eV for C plus a neutral vacancy. The substitutional C and the
C–V pair are both immobile.

NOTE Advanced Calibration for Sentaurus Process KMC does not include any
parameters for carbon in germanium.

Advanced Calibration for Process Simulation User Guide 205

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Fluorine

Current understanding attributes fluorine diffusion in silicon primarily to the migration of

interstitial fluorine from a bond-centered site or a tetrahedral site, both of which are also the
ground-state configurations of fluorine [79][80]. To comply with this model for this calibration
of Sentaurus Process KMC, a very stable and mobile Fi of neutral charge has been
assumed [81], similar to the advanced fluorine model of continuum Advanced Calibration
Version P-2019.03 (see Fluorine Diffusion and Clustering on page 74). Note that in the context
of Sentaurus Process KMC, interstitial fluorine (Fi) and fluorine–interstitial pairs (FI) are the
same.

NOTE By default, Sentaurus Process KMC assigns the particle name F to the
interstitial fluorine. To correctly rename the interstitial fluorine to Fi in
Advanced Calibration, the Monte Carlo implantation name of fluorine
is redefined by the command:
set MCnameOf(Fluorine) "FI"

The microscopical diffusivity for Fi is assumed to be:

Vm(Fi) = 1.00e-6 * exp(-0.600eV/kT)

where the binding prefactors equal 1.0 and the energies equal 4.59 eV for F plus a neutral
interstitial and 1.95 eV for F plus a neutral vacancy as reported in [81].

The substitutional F and the F paired with V are immobile.

In addition to diffusion in crystalline silicon, Advanced Calibration enables fluorine diffusion

in amorphous silicon mediate by dangling bonds similar as for boron [15][16]. The
microscopical diffusivity with activation energy taken from the literature [82] is set to:
Vm(Fi) = 0.15* exp(-2.2/kT)

where the binding prefactors equal 1.0, and the energies equal 0.2 eV for F plus a dangling
bond. The coefficient for dangling bond creation per F atom in amorphous silicon is set to 0.5
to obtain good accuracy with SIMS for fluorine implantation into amorphous silicon [69].

NOTE Advanced Calibration for Sentaurus Process KMC does not include any
parameters for fluorine in germanium.

Nitrogen

Current understanding attributes nitrogen diffusion in silicon primarily to the migration of

interstitial nitrogen from a bond-centered site or a split-interstitial site, both of which are also
the ground-state configurations of nitrogen, and to the migration of nitrogen dimer [83][84].

206 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Following the continuum model proposed by Adam et al. [85], a very stable and mobile Ni of
neutral charge has been assumed for this calibration of Sentaurus Process KMC. The N2I2
nitrogen dimer diffusion, which is enabled by default and modeled by the special dopant Nn,
is switched off for simplicity by disabling the following nonstandard interaction:
pdbSet KMC Si N SpecialReaction N,N,NnV false

NOTE By default, Sentaurus Process KMC assigns the particle name N to the
interstitial nitrogen. To correctly rename the interstitial nitrogen to Ni in
Advanced Calibration, the Monte Carlo implantation name of nitrogen
is redefined by the command:
set MCnameOf(Nitrogen) "Ni"

Based on [84], the microscopical diffusivities for Ni and NV are defined to be:
Vm(Ni) = 1.70e-3 * exp(-0.560eV/kT)
Vm(NV) = 1.00e-4 * exp(-1.360eV/kT)

where the binding prefactors equal 1.0, and the energies equal 3.5 eV [83] for N plus a neutral
interstitial and 1.7 eV [67] for N plus a neutral vacancy.

The substitutional N is immobile.

Nitrogen is known to introduce deep levels into the electronic band gap and, therefore, does not
serve as a dopant in silicon.

NOTE Advanced Calibration for Sentaurus Process KMC does not include any
parameters for nitrogen in germanium.

Clusters
In Advanced Calibration for continuum Sentaurus Process, the transient dopant cluster model
is, in general, used for dopant activation in crystalline silicon and germanium, which
incorporates only impurities and no point defects in the dopant clusters (see Defect Cluster
Models in Silicon and Germanium on page 13). For some impurities, the advanced models
offer dopant clustering with point defects (see Part 4: Comprehensive and Slow Models on
page 68). In Sentaurus Process KMC, this phenomenon can be explained by a dopant–point
defect clustering mechanism or dopant precipitation or both. Obviously, a straight translation
from the continuum activation model to the kinetic Monte Carlo cluster reactions is only
possible if a comprehensive dopant clustering model is available. Nevertheless, additional
dopant cluster calibration for Advanced Calibration of Sentaurus Process KMC is performed
to obtain a good agreement with SIMS and sheet resistance measurements.

Advanced Calibration for Process Simulation User Guide 207

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

For all impurity clusters, the corrections for the binding energies with bandgap narrowing are
taken into account using the parameter flag:
pdbSet KMC Si BandGap Correct.Complex true

Boron

Based on [86][87][88], a comprehensive model for boron–interstitial clustering (BIC) in

silicon is implemented. The allowed BIC configurations are BnIm with 0 < n < 9 and
0 ≤ m ≤ 9.

The potential energies and capture volumes of B2, B2I, B2I2, B2I3, B3, B3I, B3I2, and B3I3 as well
as the boron–interstitial pair and the interstitial emission factors were translated from the BIC
model of [87], being part of the comprehensive models of continuum Advanced Calibration
(see Boron–Interstitial Clusters on page 71). Since this continuum BIC model includes only the
configurations B2, B2I, B2I2, B3I, B3I2, and B3I3, the potential energies and capture volumes of
the configurations BI2, B2I3, and B3 are calibrated in kinetic Monte Carlo only. The potential
energies and capture volumes of all other BIC configurations are inherited from [88]. To obtain
a good agreement with a wide range of SIMS, the boron–interstitial pair emission factor is
adjusted.

In addition to clustering in crystalline silicon, Sentaurus Process KMC allows dopant clusters
in amorphous silicon. In the case of boron, B2 and B3 clusters including one or two dangling
bonds are allowed to form with potential energies calibrated to SIMS [15][68].

In the literature [89], segregation of boron to end-of-range defects is observed. To account for
such an effect, the mobile boron–interstitial pairs can react with extended defects such as {311}
defects and dislocation loops. The pair breaks up and the interstitial is incorporated into the
extended defect, while the boron is regarded as substitutional.

For germanium, the same BIC model is implemented with identical cluster configurations and
capture volumes. The potential energies of the BICs as well as the boron–interstitial pair and
the interstitial emission factors were calibrated based on SIMS and sheet resistance data from
the literature [90][91].

Arsenic
Based on the density functional theory (DFT) calculations of Sahli et al. [72] and Harrison et
al. [92], arsenic forms stable clusters with vacancies and with interstitials in silicon. The
allowed As–V clusters are As2V, As3V, and As4V. The allowed As–I clusters are:
■ AsnI with 1 < n < 6
■ AsnI2 with 0 < n < 7

208 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

■ AsnI3 with 0 < n < 7

■ AsnI4 with 1 < n < 5

The allowed arsenic agglomerates are As2, As3, and As4. The corresponding potential energies
are taken from [71], [72], or [92]. The capture volumes as well as the emission factors are
calibrated.

The formation energies of As3V, As4, and As4V are even lower than the isolated substitutional
As (Ef(AsnV) = Epot(AsnV + Ef(V) < 0) in silicon and, therefore, these are the most stable
clusters. Including AsnIm clusters in the arsenic-clustering model improves the accuracy for
lower thermal budgets and higher interstitial supersaturations compared to the model where
arsenic forms only stable clusters with vacancies, but not with interstitials [93].

Following [71] and [94], As2V clusters are defined as mobile clusters:
pdbSet KMC Si As Dm_Complex As2V 4.0
pdbSet KMC Si As Em_Complex As2V 2.0

With As2V clusters being mobile, the high concentration effect of As diffusion in silicon [74]
can be explained [71].

For germanium, the same As–V and As–I cluster model is implemented with identical cluster
configurations and capture volumes. The potential energies of As2V, As3V, and As4V are
similar to the ones reported in [77], while the potential energies of the remaining arsenic
clusters as well as the arsenic–point defect pair and the interstitial emission factors were
calibrated. Unlike for silicon, As2V clusters are defined as immobile clusters.

Phosphorus
For phosphorus, an approach was chosen following the suggestion of [72]. Phosphorus forms
stable clusters with vacancies and with interstitials. The allowed P–V clusters are P2V, P3V, and
P4V. The allowed P–I clusters are PI2, P2I, P2I2, P3I, P3I2, P4I, and P4I2. The allowed phosphorus
agglomerates are P2, P3, and P4. The corresponding potential energies are based on DFT
calculations of [72], while the capture volumes as well as the emission factors are calibrated
by comparison with experimental data.

Unlike the neutral P–V and P–-I clusters, the phosphorus agglomerates P2, P3, and P4 are
assumed to carry a single positive charge like substitutional phosphorus. This assumption is
based on DFT calculations by Synopsys following [95].

To allow phosphorus deactivation at high concentrations and low thermal budgets, the
percolation model for diffusion-less deactivation by immobile reactants is enabled.

Advanced Calibration for Process Simulation User Guide 209

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

The allowed reactions are:

■ P + P <=> P2
■ P + P2 <=> P3
■ P + P3 <=> P4
■ P + P2V <=> P3V
■ P + P3V <=> P4V

In analogy to arsenic, P2V clusters are defined as mobile clusters and, to a lesser extent, also
P3V and P4V clusters. Since P–V clusters are mobile, the high concentration effect of
phosphorus diffusion in silicon can be explained.

To account for segregation of phosphorus to end-of-range defects in silicon, the mobile

phosphorus–interstitial pairs can react with {311} defects, resulting in the pair being trapped
at the edge of the extended defect. The corresponding binding energy and emission prefactor
have been calibrated.

For germanium, the same P–V and P–I cluster model is implemented with identical cluster
configurations, charges, and capture volumes. The potential energies of P2V, P3V, and P4V are
similar to the ones reported in [77], while the potential energies of the remaining phosphorus
clusters as well as the phosphorus–point defect pair and the interstitial emission factors were
calibrated. In line with silicon, the percolation model for diffusion-less deactivation by
immobile reactants is enabled with identical parameter values.

Arsenic and Phosphorus

According to Sahli et al. [72], arsenic and phosphorus can form mixed clusters in silicon with
binding energies similar to pure arsenic or phosphorus clusters. Therefore, mixed
arsenic–phosphorus clusters with interstitials or vacancies are allowed in this calibration. The
allowed P–As–V clusters are PAsV, PAs2V, P2AsV, PAs3V, P2As2V, and P3AsV. The allowed
P–As–I clusters are PAsI, PAs2I, and P2AsI. The allowed P–As agglomerates are PAs, PAs2, and
P2As. The corresponding potential energies are taken from [72], while the capture volumes as
well as the emission factors are inherited from pure arsenic and phosphorus clusters.

Following [71], PAsV clusters are defined as mobile clusters:

pdbSet KMC Si P Dm_Complex PAsV 0.6
pdbSet KMC Si P Em_Complex PAsV 2.0

With PAsV clusters being mobile, the high concentration effect of As diffusion in silicon [74]
can be explained [71].

For germanium, the same mixed arsenic–phosphorus cluster model is implemented with
identical cluster configurations and capture volumes. The potential energies of PAsV, PAs2V,

210 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

P2AsV, PAs3V, P2As2V, and P3AsV are similar to the ones reported in [77], while the potential
energies of the remaining arsenic clusters as well as the arsenic–point defect pair and the
interstitial emission factors were calibrated. Unlike for silicon, PAsV clusters are defined as
immobile clusters.

Indium

In this calibration, indium is assumed to form indium–interstitial and indium–vacancy clusters

in silicon. The allowed In–V clusters are In2V, In3V, and In4V. The allowed In–I clusters are
InI2, In2I, and In3I. The allowed In agglomerates are In2, In3, and In4. The corresponding
potential energies and capture volumes as well as the emission factors are calibrated by
comparison with experimental data.

In the literature [96], segregation of indium to end-of-range defects is observed. To account for
such an effect, the mobile indium–interstitial pairs can react with extended defects such as
{311} defects and dislocation loops, resulting in the pair being trapped at the edge of the
extended defect. The corresponding binding energy and emission prefactor have been
calibrated.

Carbon

The model for carbon clustering in silicon in this calibration mainly follows the comprehensive
model for carbon–interstitial clustering as proposed in [78]. The allowed carbon–interstitial
cluster configurations are CI2, C2, C2I, C2I2, C2I3, C3, C3I, C3I2, C3I3, C3I4, C4I1, C4I2, C4I3, C4I4,
C4I5, C5I3, C5I4, C5I5, C5I6, C6I4, C6I5, C6I6, and C6I7. In addition, the carbon–vacancy clusters
C2V and C3V are allowed. The corresponding potential energies and capture volumes as well
as the emission factors are calibrated by comparison with experimental data.

To account for segregation of carbon to end-of range defects in silicon, the mobile
carbon–interstitial pairs can react with {311} defects and dislocation loops [78]. The pair
breaks up and the interstitial is incorporated into the extended defect, while the carbon is
regarded as substitutional.

Boron and Carbon

According to [97][98], boron and carbon can form relatively stable mixed clusters in silicon.
Therefore, mixed boron–carbon clusters with interstitials are allowed in silicon in this
calibration. The allowed B–C–I clusters are BCI, B2CI, BC2I, BC2I2, BC3I2, BC3I3, BC3I4,
BC4I3, BC5I6, and BC5I7. The allowed B–C agglomerates are BC, BC2, and B2C. The
corresponding potential energies as well as the emission factors are calibrated, while the
capture volumes are inherited from pure boron and carbon clusters.

Advanced Calibration for Process Simulation User Guide 211

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Fluorine

In the literature [79][80], strongly bound fluorine–vacancy complexes in silicon are proposed.
The F–V structures considered there are interstitial fluorine atoms decorating one or two
vacancies. In the context of Sentaurus Process KMC, these structures are considered to be
actually fluorine–interstitial clusters [81]. This means, for example, that the cluster type F3V
of [79][80], consisting of three interstitial F and one V, corresponds to an F3I2 in Sentaurus
Process KMC. The allowed F–I clusters in silicon are FI2, F2, F2I, F3, F3I, F3I2, F4, F4I, F4I2,
F4I3, F5, F5I, F5I2, F5I3, F6, F6I1, F6I2, F6I3, F6I4, F7I3, F7I4, F8I4, F8I5, and F9I5. The allowed F–V
clusters are FV2, F2V, F2V2, and F3V. The allowed fluorine agglomerates are F2, F3, F4, F5, and
F6. The corresponding potential energies are taken from [81]. The capture volumes as well as
the emission factors are calibrated.

Nitrogen
According to the literature [65], nitrogen can cluster with interstitials and vacancies in silicon.
The nitrogen dimer N2I2 is supposed to be very stable and even mobile [84]. In this calibration,
nitrogen dimer is allowed to diffuse in silicon by setting:
pdbSet KMC Si N Dm_Complex N2I2 67.0e-5
pdbSet KMC Si N Em_Complex N2I2 2.38

All other nitrogen–interstitial clusters are assumed to be immobile. The allowed N–I clusters
are NI2, N2I, N2I2, and N2I3. The only allowed N agglomerate is N2. The corresponding
potential energies and capture volumes as well as the emission factors are calibrated by
comparison with experimental data.

Segregation
The Si–SiO2, Ge–SiO2, and Ge–GeO2 interface model accounting for dopant segregation is the
three-phase segregation model, similar to the one of continuum Sentaurus Process and [99].
The Sentaurus Process KMC parameters EMax.Surf and C0Max.Surf correspond to the
continuum parameter CMax, which defines the number of trap sites at the interface. While the
interface segregation of the continuum model is controlled by dopant trapping and the
parameters of the emission rates for each side of the interface, the parameters of the kinetic
Monte Carlo interface model are the barrier and binding energies of dopants on each side of
the interface.

In the following sections, calibration of Si–SiO2, Ge–SiO2, and Ge–GeO2 segregation is

explained for different dopants and impurities. In general, parameters for Si–nitride and
Ge–nitride interfaces are inherited from Si–SiO2 and Ge–GeO2 interfaces and are not explained
explicitly.

212 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Boron

At the Si–SiO2 interface, the number of interface traps for boron is temperature independent
and corresponds to the one for Advanced Calibration for continuum Sentaurus Process at
647°C (see Boron Dose Loss on page 25). Interface barrier and binding energies for boron
were calibrated based on a collection of boron SIMS profiles, for both crystalline and
amorphous silicon.

The parameters of the Ge–SiO2 and Ge–GeO2 interfaces are set to the same values as for the
Si–SiO2 interface.

Since the interface barrier for boron trapping at the silicon side is relatively low, the limiting
factor for boron dose loss, in general, is the number of interface traps. In the presence of
fluorine trapped in the interface (such as in the case of BF2 implantation or fluorine
coimplantation after annealing), the number of boron interface traps is increased by the
Factor.Max.Surf and Exp.Max.Surf parameters. As a result, boron dose loss is enhanced,
similar as for continuum Sentaurus Process (see Fluorine Diffusion and Clustering on
page 74).

Arsenic
At the Si–SiO2 interface, the number of interface traps for arsenic is temperature independent
and corresponds to the one for Advanced Calibration for continuum Sentaurus Process at
436°C (see Arsenic Dose Loss on page 28). Interface barrier and binding energies for arsenic
were calibrated based on a collection of arsenic SIMS profiles.

The parameters of the Ge–SiO2 and Ge–GeO2 interfaces are set to the same values as for the
Si–SiO2 interface.

Phosphorus
At the Si–SiO2 interface, the number of interface traps for phosphorus is much higher than the
value from Advanced Calibration for continuum Sentaurus Process (see Phosphorus Dose Loss
on page 30), because of the lack of P2 accumulation at the interface for KMC. The interface
barrier and binding energies for phosphorus were calibrated based on a collection of
phosphorus SIMS profiles. In particular, a negative barrier for P–I pairs was chosen to allow
for strong phosphorus dose loss.

The parameters for interface barriers, interface binding energies, interface traps, and emission
to oxide for the Ge–SiO2 and Ge–GeO2 interfaces were calibrated based on a collection of
phosphorus SIMS profiles.

Advanced Calibration for Process Simulation User Guide 213

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Indium

The number of interface traps at the Si–SiO2 interface for indium cannot be taken from
Advanced Calibration for continuum Sentaurus Process, where no third phase is modeled (see
Indium Parameters on page 33), and is set to a high value to allow for strong dose loss. The
Si–SiO2 interface barrier and binding energies for indium were calibrated based on a collection
of indium SIMS profiles.

Carbon

The number of interface traps at the Si–SiO2 interface for carbon cannot be taken from
Advanced Calibration for continuum Sentaurus Process, where no third phase is modeled.

The number of interface traps, the interface barrier, and the binding energies for carbon have
been calibrated based on a collection of carbon SIMS profiles.

Fluorine

The number of Si–SiO2 interface traps, the Si–SiO2 interface barrier, and the binding energies
for fluorine have been calibrated to allow strong dose loss from silicon. In addition, fluorine
diffusion in oxide is strongly reduced to save CPU time.

Nitrogen

The number of Si–SiO2 interface traps, the Si–SiO2 interface barrier, and the binding energies
for nitrogen have been calibrated to allow strong dose loss from silicon. In addition, the
nitrogen emitted from the interface to the oxide is evaporated, and the nitrogen diffusion in
oxide is strongly reduced to save CPU time.

Recrystallization
Solid phase epitaxial regrowth (SPER) can affect the impurity distribution. The
recrystallization front can move dopants away, changing the concentration profiles [100].
Sentaurus Process KMC can model this effect by sweeping the dopant concentration stored in
amorphous defects when the amorphous defects recrystallize [3]. The parameters
Pref.RecrystDeposit and Ener.RecrystDeposit define the probability of a dopant
remaining in the same position after the recrystallization front passes and, therefore, the
parameters control the number of swept dopants.

In Sentaurus Process, the initial level of active concentration in amorphized regions after
recrystallization can be specified per dopant as AmInit. For Sentaurus Process KMC, this

214 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

corresponds to the prefactor C0.Recryst.Max.Active and the activation energy

E.Recryst.Max.Active. For this calibration, the deposited dopant-cluster types during
SPER are specified by Recryst.Deposit.Complex because the Recryst.Max.Size
parameter is unset. The improved algorithm enabled by the Fix.Recryst.Max.Active
parameter is used for better resolution of the activation level for lower concentrations.

Boron

According to [100], no boron is swept by the recrystallization front in silicon. Therefore, the
Pref.RecrystDeposit value is set to 100% [3]. The maximum active boron concentration
20
after silicon recrystallization is temperature dependent, with 2.4 × 10 at 550°C and
20 –2
5.4 × 10 cm at 1050°C close to reported values [101]. These values are higher than the
one for continuum Advanced Calibration (see Boron–Interstitial Clusters on page 71). The
exceeding boron concentration is deposited as B2 clusters [102]. Finally, boron has no effect on
the silicon recrystallization velocity other than the doping effect (see Amorphization and
Recrystallization on page 185).

As for silicon, no boron is swept by the recrystallization front in germanium. Therefore, the
Pref.RecrystDeposit value is set to 100%. The maximum active boron concentration after
20 –2
germanium recrystallization is 5.0 × 10 cm [91] for all temperatures. The exceeding boron
concentration is deposited as B2 clusters, in analogy to silicon. Finally, boron has no effect on
the germanium recrystallization velocity other than the doping effect (see Amorphization and
Recrystallization on page 185).

Arsenic
Following [100], arsenic is swept by the recrystallization front in silicon The
Pref.RecrystDeposit value is set to 20%. The maximum active arsenic concentration after
20 –2
silicon recrystallization is temperature dependent, with 6.0 × 10 cm at 550°C and
21 –2
1.3 × 10 cm at 1050°C . These values are higher than the one for continuum Advanced
Calibration. The exceeding arsenic concentration is deposited as As2 and As3 clusters in the
ratio 1:1. Finally, arsenic is assumed to have no effect on the recrystallization velocity other
than the doping effect (see Amorphization and Recrystallization on page 185).

As for silicon, arsenic is assumed to be swept by the recrystallization front in germanium. The
Pref.RecrystDeposit value is set to 50%. The maximum active arsenic concentration after
20 –2
germanium recrystallization is 1.0 × 10 cm for all temperatures. The exceeding arsenic
concentration is deposited as As2 clusters only. Finally, arsenic has no effect on the germanium
recrystallization velocity other than the doping effect (see Amorphization and
Recrystallization on page 185).

Advanced Calibration for Process Simulation User Guide 215

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Phosphorus

For silicon, the Pref.RecrystDeposit value for phosphorus is set to 10%. The maximum
active phosphorus concentration after silicon recrystallization is temperature dependent, with
21 –2 21 –2
1.1 × 10 cm at 550°C and 2.4 × 10 cm at 1050°C . These values are higher than the
one for continuum Advanced Calibration. The exceeding phosphorus concentration is
deposited as P3 clusters. Finally, phosphorus has no effect on the recrystallization velocity other
than the doping effect (see Amorphization and Recrystallization on page 185).

As for silicon, phosphorus is assumed to be swept by the recrystallization front in germanium.

The Pref.RecrystDeposit value is set to 25%. The maximum active phosphorus
concentration after germanium recrystallization is temperature dependent, with
20 –2 20 –2
1.5 × 10 cm at 400°C and 2.4 × 10 cm at 550°C. The exceeding phosphorus
concentration is deposited as P2 clusters only. Finally, phosphorus has no effect on the
germanium recrystallization velocity other than the doping effect (see Amorphization and
Recrystallization on page 185).

Indium

Following [100], [103], and [104], indium is swept by the recrystallization front in silicon, and
its incorporation during SPER of silicon is temperature dependent [105]. The
Ener.RecrystDeposit value is set to 0.3 eV. The maximum active indium concentration
20 –2
after silicon recrystallization is temperature dependent, with 1.1 × 10 cm at 550°C and
20 –2
2.4 × 10 cm at 1050°C . These values are higher and more realistic than the one for
continuum Advanced Calibration. The exceeding indium concentration is deposited as In2
clusters. Finally, indium has no effect on the recrystallization velocity other than the doping
effect (see Amorphization and Recrystallization on page 185).

Carbon

For silicon, the Pref.RecrystDeposit value for carbon is set to 50%. The maximum
substitutional carbon concentration after silicon recrystallization is temperature dependent,
20 –2 21 –2
with 6.0 ×10 cm at 550°C and 1.3 ×10 cm at 1050°C . The exceeding carbon
concentration is deposited as C2 [78] and C3. Finally, carbon has an impurity effect on the
recrystallization velocity with a recrystallization activation energy of 2.8 eV for 100% carbon.

Fluorine

Following [106], fluorine is swept by the recrystallization front in silicon, and its incorporation
during SPER of silicon is temperature dependent. The Ener.RecrystDeposit value is set to
18 –2
0.5 eV. Concentrations below 1 ×10 cm are deposited as interstitial fluorine, while the rest
is deposited as F4I3 clusters.

216 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Finally, fluorine has an impurity effect on the recrystallization velocity with a recrystallization
activation energy of 3.1 eV for 100% fluorine [9].

Nitrogen

In this calibration, no nitrogen is swept by the recrystallization front in silicon. Concentrations

19 –2
below 1 ×10 cm are deposited as interstitial nitrogen, while the rest is deposited as N2I
clusters.

Epitaxy
In Sentaurus Process, selective epitaxial growth (SEG) of silicon, SiGe, and germanium with
in situ doping can be simulated by specifying total chemical concentrations of dopants in the
epi.doping argument of the diffuse command. The initial level of active concentration in
epitaxially grown regions can be specified per dopant as EpiInit.

For Sentaurus Process KMC and using LKMC for SEG, this corresponds to the prefactor
C0.Epi.Max.Active and the activation energy E.Epi.Max.Active. The deposited
inactive dopant-cluster types during SEG are specified by Epi.Deposit.Complex.

Boron

For the initial activation in the case of in situ boron-doped epitaxial growth of silicon or
20 –3 20 –3
germanium, an activation level of 4 ×10 cm or 5 ×10 cm , respectively, is assumed. The
exceeding boron concentration is deposited as B2 clusters.

Arsenic

For the initial activation in the case of in situ arsenic-doped epitaxial growth of silicon or
20 –3 20 –3
germanium, an activation level of 2.5 ×10 cm or 1.0 ×10 cm , respectively, is assumed.
The exceeding arsenic concentration is deposited as As2 clusters.

Phosphorus
For the initial activation in the case of in situ phosphorus-doped epitaxial growth of silicon or
20 –3
germanium, an activation level of 1.27 ×10 cm is assumed. The exceeding phosphorus
concentration is deposited as P4V clusters.

This assumption follows the literature [107][108][109], which attributes the incomplete
activation and tensile strain measured after high-concentration Si:P epitaxy to the presence of
the pseudocubic Si3P4 compound, corresponding to a P4V cluster in the KMC formalism. (The

Advanced Calibration for Process Simulation User Guide 217

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Si3P4 compound consists of four substitutional P, three Si lattice atoms, and one Si vacancy.
The KMC formalism ignores the lattice Si; therefore, P4V.)

Carbon

The maximum substitutional carbon concentration in epitaxially grown silicon is assumed to

20 –3
be 8 ×10 cm , which corresponds to 1.6%. The exceeding carbon concentration is deposited
as C2 clusters.

SiGe

Linear Germanium Correction Factors

By default, the Ge effects are modeled by linear correction factors for silicon parameters. The
calibration is valid for low Ge mole fractions ranging between 0% and 50%.

For low Ge mole fractions, the chemical effect of Ge on dopants is indirect by point defects,
which themselves encounter a Ge chemical effect, and is direct mainly for dopant–defect pair
diffusion [1][2]. There is no experimental evidence for any chemical effect of Ge on dopant
solubility.

Boron

Boron diffusion is retarded in relaxed SiGe in comparison to the one in Si (see Germanium
Effect on Dopant Diffusivity on page 44). The linear expression for the diffusivity activation
energy was derived by Ahn [110] and confirmed for low mole fractions by measurements
performed within the ATOMICS research project [36]. In this calibration, the migration barrier
and the formation energy of B–I pairs are increased in the presence of Ge:
pdbSet KMC Si B EmGe BiM [expr 0.23/5.e22]
pdbSet KMC Si B EmGe Bi [expr 0.23/5.e22]
pdbSet KMC Si B EmGe BiP [expr 0.23/5.e22]
pdbSet KMC Si B EfGe BiM [expr 0.30/5.e22]

To allow for boron segregation into SiGe [111], the formation energy of substitutional B
decreases in the presence of Ge [51]:
pdbSet KMC Si B EfGe B [expr -0.50/5.e22]

In addition, boron diffusion in amorphous SiGe is retarded in comparison to amorphous

silicon [112]. Again, the migration barrier and the formation energy of boron and dangling-
bond pairs increase in the presence of Ge.

218 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Arsenic

Arsenic diffusion is enhanced in relaxed SiGe in comparison to the one in Si [113][114][115]

(see Germanium Effect on Dopant Diffusivity on page 44). In this calibration, the migration
barrier of As–I pairs is increased in the presence of Ge, while the migration barrier of As–V
pairs is lowered in the presence of Ge. The corresponding pair formation energies are
independent of Ge content.

Phosphorus

Phosphorus diffusion is enhanced in relaxed SiGe in comparison to the one in Si [114][116].

In this calibration, the migration barrier of P–I pairs is decreased in the presence of Ge, while
the migration barrier of P–V pairs is unchanged in the presence of Ge. The P–I pair formation
energy is independent of Ge content, while the formation energy of P–V pairs is increased in
the presence of Ge.

Silicon and Germanium Parameter Interpolation

Optionally, you can enable Si and Ge parameter interpolation for SiGe using the command:
KMC_SiGe_and_Stress_Effect 1 0

This disables the default linear correction factors and uses the interpolation parameters defined
for SiliconGermanium. The calibration covers the full Ge mole fraction range (0–100%).

In general, prefactor parameters are interpolated in logarithmic scale, and energy parameters
are interpolated in linear scale.

Boron

According to experimental work [36][117], interstitial-meditated B diffusion in relaxed SiGe

is retarded with increasing Ge content up to ~50%. For higher Ge content, boron diffusion is
again enhanced [118]. Therefore, a parabolic dependency on Ge mole fraction of the diffusivity
is assumed for the full Ge mole fraction range instead (see Germanium Effect on Dopant
Diffusivity on page 44). However, the calibration based on SIMS for Si, SiGe [118][119], and
Ge [34][70] revealed no need for parabolic interpolation of the microscopic parameters for the
boron–interstitials (migration energy, prefactor, and binding energy) but for the ionization
energies only. The nonlinear effect of Ge on macroscopic boron diffusivity in SiGe is a
consequence of the parabolic dependency of interstitials on the Ge mole fraction.

For BIC potential energies and the initial activation after SPER, linear interpolation is assumed.
The boron–interstitial cluster emission prefactor is interpolated parabolically.

Advanced Calibration for Process Simulation User Guide 219

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

Arsenic

Arsenic diffusion is enhanced in relaxed SiGe in comparison to the one in silicon [113]
[114][115] (see Germanium Effect on Dopant Diffusivity on page 44). In Advanced
Calibration, only the migration rate of As+V– – is larger in germanium than in silicon. For all
other As–V and As–I pairs, they are larger in silicon than in germanium. Based on SIMS data
provided to Synopsys by AMAT-VSE, parabolic interpolation for the arsenic–interstitial and
arsenic–vacancy migration energy has been calibrated.

For the arsenic-cluster potential energies, the cluster emission prefactors, and the initial
activation after SPER, parabolic interpolation is assumed. Moreover, the prefactor for As2V
migration is interpolated parabolically to limit cluster diffusion in SiGe.

For the interface with SiO2, parabolic interpolation is used for the number of interface traps and
the As+V– – energy barrier.

Phosphorus

As for arsenic, phosphorus diffusion is enhanced in relaxed SiGe in comparison to the one in
silicon [114][116] (see Germanium Effect on Dopant Diffusivity on page 44). In Advanced
Calibration, the migration rate of P +V– – and P–I pairs is larger in germanium than silicon. For
all other P–V pairs, they are larger in silicon than in germanium. Based on SIMS data provided
to Synopsys by AMAT-VSE [46] and SIMS data from the literature [120], parabolic
interpolation for the phosphorus–interstitial and phosphorus–vacancy migration energy has
been calibrated.

For the phosphorus-cluster potential energies, the cluster emission prefactors, and the initial
activation after SPER, parabolic interpolation is assumed.

For the interface with SiO2, parabolic interpolation is used for the P+V– – energy barrier.

Stress Effects
The stress effects on dopants in Advanced Calibration for Sentaurus Process KMC are taken
into account only if the KMC Stress switch is enabled. This is accomplished by the command:
KMC_SiGe_and_Stress_Effect 0 1

The procedure KMC_SiGe_and_Stress_Effect has two Boolean arguments: the first is

Ge_Chem_Eff and the second is Stress_Eff, which is dedicated for stress effects.

Alternatively, the KMC Stress switch can be enabled directly:

pdbSet KMC Stress 1

220 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

The calibration of stress effects [1][2] is based on a relatively small set of measured data. The
models have not yet been tested rigorously against SIMS data for all dopant species and
nonequilibrium annealing conditions, and against electrical data from real-device fabrication
processes.

The stress effect on dopants is indirect by stress-dependent point defects and is direct mainly
for dopant–defect pair diffusion [1][2]. Dopant diffusion in Si or SiGe can be anisotropic in the
presence of biaxial strain and, therefore, anisotropic diffusion tensors are used [121]. Dopant
solubilities are changed in the presence of stress according to [122]. However, in this
calibration, the dopant–defect cluster stability is kept independent of stress due to the limited
availability of experimental data for calibration.

Boron

According to first-principles density functional theory calculations [61][123][124], boron

diffusion in silicon shows an anisotropy as a function of biaxial strain. Biaxial tension (as in
strained Si on SiGe) leads to significantly higher in-plane diffusion compared to the
perpendicular direction. The diffusion barrier along the strain plane is decreased (with a value
similar to continuum Advanced Calibration (see Impact of Pressure on Dopant Diffusivity on
page 49)), while the barrier in the vertical direction remains unchanged:
pdbSetArray KMC Si B VD { BiM -0.0024,0.0
Bi -0.0024,0.0
BiP -0.0024,0.0 }

The formation energy of B–I pairs is independent of strain. To allow for boron segregation into
SiGe [51], the formation energy of substitutional B decreases in the presence of strain:
pdbSetDoubleArray KMC Si B VF { B -0.03
BiM 0.0 }
pdbSetDoubleArray KMC Ge B VF { B -0.03
BiM 0.0 }

Arsenic

According to first-principles density functional theory calculations [110][124], arsenic in

silicon undergoes a total diffusivity enhancement under compressive strain [115] and little
change under tensile strain [125]. The migration barrier of As–I pairs is increased under
compressive strain, while the migration barrier of As–V pairs is lowered under compressive
strain, with values similar to continuum Advanced Calibration (see Impact of Pressure on
Dopant Diffusivity on page 49).

Advanced Calibration for Process Simulation User Guide 221

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Part 2: Model Parameters for Impurities

The diffusivity enhancement or retardation is isotropic and larger for the vacancy mechanism
than for the interstitial mechanism:
pdbSetArray KMC Si As VD { AsiP -0.0040,-0.0040
Asi -0.0040,-0.0040
AsVP 0.0018,0.0018
AsV 0.0018,0.0018
AsVM 0.0018,0.0018 }
pdbSetArray KMC Ge As VD { AsiP -0.0040,-0.0040
Asi -0.0040,-0.0040
AsVP 0.0018,0.0018
AsV 0.0018,0.0018
AsVM 0.0018,0.0018 }

In this calibration, the formation energies of As–I and As–V pairs are independent of strain. To
allow for arsenic segregation out of SiGe, the formation energy of substitutional As increases
slightly in the presence of strain:
pdbSetDoubleArray KMC Si As VF { Asi 0.008
AsiP 0.0
AsVP 0.0 }
pdbSetDoubleArray KMC Ge As VF { Asi 0.008
AsiP 0.0
AsVP 0.0 }

Phosphorus

According to first-principles density functional theory calculations [110], phosphorus

undergoes a total diffusivity retardation under compressive strain and enhancement under
tensile strain. The migration barrier of P–I pairs is increased under compressive strain, while
the migration barrier of P–V pairs is lowered under compressive strain, with values similar to
continuum Advanced Calibration (see Impact of Pressure on Dopant Diffusivity on page 49).
While strain has negligible impact on in-plane diffusivity, out-of-plane diffusivity is a modest
function of strain:
pdbSetArray KMC Si P VD { PiP -0.0022,-0.0022
Pi -0.0022,-0.0022
PVP 0.0040,0.0040
PV 0.0040,0.0040
PVM 0.0040,0.0040 }
pdbSetArray KMC Ge P VD { PiP -0.0011,-0.0011
Pi -0.0011,-0.0011
PVP 0.0020,0.0020
PV 0.0020,0.0020
PVM 0.0020,0.0020 }

222 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Section 4: Model Parameters for Epitaxial Growth

In this calibration, the formation energies of P–I and P–V pairs are independent of strain. To
allow for phosphorus segregation out of SiGe [120], the formation energy of substitutional P
increases slightly in the presence of strain:
pdbSetDoubleArray KMC Si P VF { Pi 0.004
PiP 0.0
PVP 0.0 }
pdbSetDoubleArray KMC Ge P VF { Pi 0.04
PiP 0.0
PVP 0.0 }

Indium

According to first-principles density functional theory calculations [110], indium undergoes a

total diffusivity retardation under compressive strain and enhancement under tensile strain. The
migration barrier of In–I pairs is increased under compressive strain, with values similar to
continuum Advanced Calibration (see Impact of Pressure on Dopant Diffusivity on page 49).
The diffusivity enhancement or retardation is isotropic:
pdbSetArray KMC Si In VD { IniM -0.0009,-0.0009
Ini -0.0009,-0.0009
IniP -0.0009,-0.0009 }

In this calibration, the formation energy of In–I pairs is independent of strain:

pdbSetDoubleArray KMC Si In VF { IniM 0.0 }

Section 4: Model Parameters for Epitaxial Growth

This section describes the model parameters for epitaxial growth.

Selective epitaxial growth is simulated by LKMC, which must be enabled by:

pdbSet KMC Epitaxy true

The default LKMC model for epitaxial growth is the Coordinations.Planes model.

Coordinations.Planes Model
The default model for selective epitaxial growth simulation is the one reported by
Martin-Bragado and Moroz [126]. The (default) model switch is:
pdbSet LKMC Epitaxy.Model Coordinations.Planes

Advanced Calibration for Process Simulation User Guide 223

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Section 4: Model Parameters for Epitaxial Growth

This computationally efficient atomistic model can reproduce the growth rates and shapes of
selectively grown Si and SiGe. The growth process, which in reality depends on many aspects
such as adsorption, desorption, surface diffusion, simultaneous etching, and chemistry as well
as gas composition and pressure, is modeled simply by surface orientation–dependent net
growth rates. Therefore, the calibration being part of Advanced Calibration is only a baseline
one, but it cannot predict the absolute growth of all epitaxial processes due to the limited set of
parameters included in the model. To fit the absolute growth, the recommended scaling factors
are either the epi.thickness argument of the diffuse Epi command or the parameter:
pdbSet KMC Si Epitaxy prefactor.thickness <value>

The parameters for epitaxial growth of undoped (100) silicon and (100) germanium are based
on low-temperature chemical vapor deposition (CVD) experiments [127][128][129]. The
growth rates show an exponential dependency on temperature in the hydrogen
desorption–limited regime of silicon ( ≤ 800°C ) and germanium ( ≤ 700°C ). The activation
energies associated with the growth processes are 1.9 eV for silicon (dichlorosilane [127] and
disilane [129]) and 0.4 eV for germanium (germane [127]). The prefactors for the different
orientations and the correction energy for {311} planes are defined to have average values and
to be in line with [126].

Linear SiGe parameter interpolation is assumed for epitaxial growth. The growth prefactors are
interpolated in logarithmic scale, and the activation energies are interpolated in linear scale. In
general, SiGe growth is faster than silicon growth and is performed at lower temperatures.

Coordinations Model
An alternative model for selective epitaxial growth simulation is enabled by:
pdbSet LKMC Epitaxy.Model Coordinations

In contrast to the Coordinations.Planes model, the growth rates of the Coordinations

model are based on species-dependent bonds and atomistic gas molecule arrival rates. The
Coordinations model requires the specification of epitaxial gas conditions by specifying
ambients in the diffuse, temp_ramp, or gas_flow commands. In addition, it is assumed
that the net growth process is anisotropic and depends on the total binding energy of the site,
which is the summation of all neighbor binding energies up to the third nearest neighbors of
the diamond lattice [130]. Again, growth rates and faceting of selectively grown Si and SiGe
can be reproduced, but many aspects of the growth process such as desorption, surface
diffusion, simultaneous etching, and chemistry are ignored.

The parameters for epitaxial growth of undoped (100) silicon and (100) germanium are again
based on low-temperature CVD experiments and are defined for the ambients dichlorosilane
[127], germane [127], silane [128], and disilane [129].

224 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Section 4: Model Parameters for Epitaxial Growth

The growth rates show an exponential dependency on temperature in the hydrogen

desorption–limited regime of silicon ( ≤ 800°C ) and germanium ( ≤ 700°C ).

The calibrated parameters are set as the defaults of Sentaurus Process and, therefore, they are
not included in the Advanced Calibration file.

Coordinations.Reactions Model
A second alternative model for selective epitaxial growth simulation is enabled by:
pdbSet LKMC Epitaxy.Model Coordinations.Reactions

In line with the Coordinations model, the growth rates of the

Coordinations.Reactions model are based on species-dependent bonds and atomistic gas
molecule arrival rates. The Coordinations.Reactions model requires the specification of
epitaxial gas conditions by specifying ambients in the diffuse, temp_ramp, or gas_flow
commands. Its anisotropic net growth process depends on the total binding energy of the site,
which is the summation of all neighbor binding energies up to the third nearest neighbors of
the diamond lattice [130].

In addition, surface chemistry such as molecule adsorption, desorption, dissociation, and

etching is modeled explicitly to take into account the hydrogen passivation of the surface and
desorption. This allows a calibration for the entire temperature range including the hydrogen
desorption–limited regime at low temperatures and the precursor supply- and dissociation-
limited regime at high temperatures [46]. However, surface diffusion ignored for efficiency,
which will likely lead to rougher surfaces in the simulation compared to reality.

Epitaxial Growth
The parameters for epitaxial growth of undoped (100) silicon and (100) germanium are again
based on CVD experiments for the ambients dichlorosilane, germane, silane, disilane,
phosphine, and diborane.

Silicon Growth

Epitaxial growth of undoped silicon for dichlorosilane ambients with H2 carrier gas is
calibrated against data from the literature [127][131][132]. In this calibration, dichlorosilane
(SiH2Cl2) adsorption is modeled by decomposition into the surface passivating products
silylene (SiH*), hydrogen (H*), and chlorine (Cl*). Silylene decomposes into Si and a
passivating H* by a surface reaction; however, for this specific reaction, H* is supposed to
desorb immediately.

Advanced Calibration for Process Simulation User Guide 225

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
Section 4: Model Parameters for Epitaxial Growth

Epitaxial growth of undoped silicon for silane and disilane ambients with and without H2
carrier gas is calibrated against data from the literature [133][134][135]. Silane (SiH4) and
disilane (Si2H6) adsorption is modeled by decomposition into the surface passivating products
silylene (SiH2* or SiH3*) and H*. Silylenes decompose into Si and two to three passivating H*
by a surface reaction; however, for this specific reaction, H* is supposed to desorb
immediately.

In situ doped epitaxial growth of silicon for phosphine and diborane ambients with H2 carrier
gas is calibrated against data from the literature [136][137]. In line with silane and disilane, the
adsorption of phosphine (PH3) and diborane (B2H6) is modeled by decomposition into the
surface passivating products PH2* or BH2*, and H*. PH2* and BH2* decompose into P or B
and two passivating H* by surface reactions; however, for these specific reactions, H* is
supposed to desorb immediately. In addition, surface segregation for P is assumed as suggested
in the literature [136][138]. The exchange of the subsurface P with the surface Si leads to a
P-rich surface that influences the epitaxial growth and limits P incorporation in the bulk.

Germanium and SiGe

To simulate germanium or SiGe epitaxial growth using the Coordinations.Reactions

model, it is recommended to switch on an optional parameter set by calling:
LKMC_SiGe_Epitaxy

This enables a ten times higher hydrogen adsorption and desorption rate compared to the
standard rates, which are needed to realize simulated germanium growth rates that are not
growth time dependent. The higher hydrogen adsorption and desorption rates allow for a quick
saturation of hydrogen coverage of the germanium surface and, therefore, stabilize the growth
velocity. As a consequence, other adsorption rates have been adjusted to compensate for the
higher hydrogen reaction rates. However, silicon growth can also be accurately simulated with
the parameter set of LKMC_SiGe_Epitaxy, but the simulation times are one order of
magnitude slower compared to the standard parameter set due to many more hydrogen events.

Epitaxial growth of undoped germanium for germane ambients with H2 carrier gas is calibrated
against data from the literature [139][140]. Germane (GeH4) adsorption is modeled by
decomposition into the surface passivating products germylene (GeH2*) and H*. Germylene
decomposes into Ge and two passivating H* by a surface reaction; however, for this specific
reaction, H* is supposed to desorb immediately.

Epitaxial growth of undoped SiGe for dichlorosilane and germane ambients with H2 carrier gas
is calibrated against data from the literature [132][133][140][141][142]. To have both the
growth rate and the grown Ge mole fraction in the correct range, surface segregation for Ge
must be assumed as suggested in the literature [138]. The exchange of the subsurface Ge with
the surface Si leads to a Ge-rich surface that influences the epitaxial growth and limits Ge
incorporation in the bulk. In situ doped epitaxial growth of SiGe for diborane ambients with H2
carrier gas is calibrated against data from the literature [143].

226 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
Section 4: Model Parameters for Epitaxial Growth

Moreover, LKMC_SiGe_Epitaxy accounts for a Ge mole fraction–dependent lattice density

when transferring the LKMC particles to KMC by setting:
pdbSet LKMC Lattice.Density.Correction 1

Therefore, to activate both Ge parameter interpolation and Ge mole fraction–dependent lattice

density for KMC, it is recommended to also call:
KMC_SiGe_and_Stress_Effect 1 *

Etching

Etching of undoped silicon for HCl ambients with H2 carrier gas is calibrated against data from
the literature [144][145][146]. In this calibration, HCl adsorption is modeled by decomposition
into the surface passivating products hydrogen (H*) and chlorine (Cl*). The etching process
itself is modeled by an etch reaction by a surface passivating species SiHCl* – Cl* and a silicon
surface atom form SiHCl* – which desorbs. The inclusion of H* in this reaction is neglected.

The faster etching of (110) silicon compared to the (100) silicon is achieved by a strong
bonding of SiHCl* to the nearest neighbor (2 in the case of (100) silicon and 1 in the case of
(110) silicon).

For low temperatures, the process is SiHCl* formation/desorption limited as well as hydrogen
passivation limited, as shown in experiments [144] where the orientation dependency is strong
for low temperatures and low for high temperatures. For high temperatures, the process is HCl
adsorption limited, as demonstrated in experiments [145] where the etch rates depend strongly
on the HCl/H2 gas flow ratio at high temperatures.

Etching of undoped germanium and SiGe for HCl ambients with H2 carrier gas is calibrated
against data from the literature [147] in analogy to the one for silicon. The accuracy of
simulations with and without LKMC_SiGe_Epitaxy is comparable.

NOTE The current calibration for epitaxial growth and etching is based on
simplistic assumptions for the surface chemistry and does not utilize the
full potential of the Coordinations.Reactions model. More
complex and physical surface reactions could be used, but accurate
calibration is currently trailing.

Advanced Calibration for Process Simulation User Guide 227

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
References

References
[1] N. Zographos et al., “Process Modeling of Chemical and Stress Effects in SiGe,” in 19th
International Conference on Ion Implantation Technology (IIT), Valladolid, Spain,
pp. 212–216, June 2012.
[2] N. Zographos and I. Martin-Bragado, “Process modeling of stress and chemical effects
in SiGe alloys using kinetic Monte Carlo,” Journal of Computational Electronics,
vol. 13, no. 1, pp. 59–69, 2014.
[3] N. Zographos and I. Martin-Bragado, “A Comprehensive Atomistic Kinetic Monte
Carlo Model for Amorphization/Recrystallization and its Effects on Dopants,” in MRS
Symposium Proceedings, Doping Engineering for Front-End Processing, vol. 1070,
p. 1070-E03-01, 2008.
[4] T. E. Haynes and O. W. Holland, “Comparative study of implantation-induced damage
in GaAs and Ge: Temperature and flux dependence,” Applied Physics Letters, vol. 59,
no. 4, pp. 452–454, 1991.
[5] S. Koffel et al., “Amorphization kinetics of germanium during ion implantation,”
Journal of Applied Physics, vol. 105, no. 1, p. 013528, 2009.
[6] B. L. Darby et al., “Self-implantation energy and dose effects on Ge solid-phase
epitaxial growth,” Nuclear Instruments and Methods in Physics Research B, vol. 269,
no. 1, pp. 20–22, 2011.
[7] L. Pelaz et al., “Atomistic modeling of amorphization and recrystallization in silicon,”
Applied Physics Letters, vol. 82, no. 13, pp. 2038–2040, 2003.
[8] J. J. Loferski and P. Rappaport, “Radiation Damage in Ge and Si Detected by Carrier
Lifetime Changes: Damage Thresholds,” Physical Review, vol. 111, no. 2, pp. 432–439,
1958.
[9] G. L. Olson and J. A. Roth, “Kinetics of Solid Phase Crystallization in Amorphous
Silicon,” Materials Science Reports, vol. 3, pp. 1–78, 1988.
[10] G.-Q. Lu, E. Nygren, and M. J. Aziz, “Pressure-enhanced crystallization kinetics of
amorphous Si and Ge: Implications for point-defect mechanisms,” Journal of Applied
Physics, vol. 70, no. 10, pp. 5323–5345, 1991.
[11] L. Csepregi et al., “Substrate-orientation dependence of the epitaxial regrowth rate from
Si-implanted amorphous Si,” Journal of Applied Physics, vol. 49, no. 7, pp. 3906–3911,
1978.
[12] B. Sklenard et al., “An atomistic investigation of the impact of in-plane uniaxial stress
during solid phase epitaxial regrowth,” Applied Physics Letters, vol. 102, no. 15,
p. 151907, 2013.
[13] B. C. Johnson and J. C. McCallum, “Dopant-enhanced solid-phase epitaxy in buried
amorphous silicon layers,” Physical Review B, vol. 76, no. 4, p. 045216, 2007.

228 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[14] D. D’Angelo et al., “Role of the strain in the epitaxial regrowth rate of heavily doped
amorphous Si films,” Applied Physics Letters, vol. 93, p. 231901, December 2008.
[15] S. Mirabella et al., “Mechanism of Boron Diffusion in Amorphous Silicon,” Physical
Review Letters, vol. 100, p. 155901, April 2008.
[16] I. Martin-Bragado and N. Zographos, “Indirect boron diffusion in amorphous silicon
modeled by kinetic Monte Carlo,” Solid-State Electronics, vol. 55, no. 1, pp. 25–28,
2011.
[17] H. Bracht, N. A. Stolwijk, and H. Mehrer, “Equilibrium Concentrations of Intrinsic
Point Defects in Silicon Determined by Zinc Diffusion,” in Proceedings of the Seventh
International Symposium on Silicon Materials Science and Technology (Semiconductor
Silicon), vol. 94-10, San Francisco, CA, USA, pp. 593–602B, May 1994.
[18] I. Martin-Bragado et al., “Physical atomistic kinetic Monte Carlo modeling of Fermi-
level effects on species diffusing in silicon,” Physical Review B, vol. 72, p. 035202, July
2005.
[19] L. Pelaz et al., “Front-end process modeling in silicon,” The European Physical Journal
B, vol. 72, no. 3, pp. 323–359, 2009.
[20] I. Martin-Bragado, Simulación atomística de procesos para Microelectrónica, Ph.D.
thesis, Universidad de Valladolid, Valladolid, Spain, 2004.
[21] J. Vanhellemont, P. Spiewak, and K. Sueoka, “On the solubility and diffusivity of the
intrinsic point defects in germanium,” Journal of Applied Physics, vol. 101, no. 3,
p. 036103, 2007.
[22] J. Vanhellemont et al., “Intrinsic point defect properties and engineering in silicon and
germanium Czochralski crystal growth,” in 5th International Symposium on Advanced
Science and Technology of Silicon Materials, Kona, HI, USA, November 2008.
[23] P. Spiewak, J. Vanhellemont, and K. J. Kurzydlowski, “Improved calculation of
vacancy properties in Ge using the Heyd-Scuseria-Ernzerhof range-separated hybrid
functional,” Journal of Applied Physics, vol. 110, no. 6, p. 063534, 2011.
[24] I. Martin-Bragado et al., “Fermi-level effects in semiconductor processing: A modeling
scheme for atomistic kinetic Monte Carlo simulators,” Journal of Applied Physics,
vol. 98, no. 5, p. 053709, 2005.
[25] J. Vanhellemont and E. Simoen, “Brother Silicon, Sister Germanium,” Journal of The
Electrochemical Society, vol. 154, no. 7, pp. H572–H583, 2007.
[26] L. Pelaz, L. A. Marqués, and J. Barbolla, “Ion-beam-induced amorphization and
recrystallization in silicon,” Journal of Applied Physics, vol. 96, no. 11, pp. 5947–5976,
2004.
[27] S. M. Sze, Physics of Semiconductor Devices, New York: John Wiley & Sons, 2nd ed.,
1981.

Advanced Calibration for Process Simulation User Guide 229

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[28] I. Martin-Bragado et al., “From point defects to dislocation loops: A comprehensive

modelling framework for self-interstitial defects in silicon,” Solid-State Electronics,
vol. 52, no. 9, pp. 1430–1436, 2008.
[29] N. E. B. Cowern et al., “Cluster ripening and transient enhanced diffusion in silicon,”
Materials Science in Semiconductor Processing, vol. 2, no. 4, pp. 369–376, 1999.
[30] F. Cristiano et al., “Formation energies and relative stability of perfect and faulted
dislocation loops in silicon,” Journal of Applied Physics, vol. 87, no. 12, pp. 8420–8428,
2000.
[31] D. P. Hickey et al., “Regrowth-related defect formation and evolution in 1 MeV
amorphized (001) Ge,” Applied Physics Letters, vol. 90, no. 13, p. 132114, 2007.
[32] S. Koffel et al., “End of range defects in Ge,” Journal of Applied Physics, vol. 105,
no. 12, p. 126110, 2009.
[33] E. Napolitani et al., “Transient enhanced diffusion of B mediated by self-interstitials in
preamorphized Ge,” Applied Physics Letters, vol. 96, no. 20, p. 201906, 2010.
[34] F. Panciera et al., “End-of-range defects in germanium and their role in boron
deactivation,” Applied Physics Letters, vol. 97, no. 1, p. 012105, 2010.
[35] S. Boninelli et al., “Role of the Ge surface during the end of range dissolution,” Applied
Physics Letters, vol. 101, no. 16, p. 162103, 2012.
[36] IST Project 027152 ATOMICS, Advanced Front-End Technology Modeling for
Ultimate Integrated Circuits, for more information, go to https://
www.iisb.fraunhofer.de/content/dam/iisb2014/en/Documents/Research-Areas/
Simulation/final-report-atomics.pdf.
[37] A. Chroneos, R. W. Grimes, and H. Bracht, “Fluorine codoping in germanium to
suppress donor diffusion and deactivation,” Journal of Applied Physics, vol. 106, no. 6,
p. 063707, 2009.
[38] R. Kögler et al., “Excess vacancies in high energy ion implanted SiGe,” Journal of
Applied Physics, vol. 101, no. 3, p. 033508, 2007.
[39] G. M. Dalpian et al., “Ab initio calculations of vacancies in SixGe1–x,” Applied Physics
Letters, vol. 81, no. 18, pp. 3383–3385, 2002.
[40] P. Venezuela et al., “Vacancy-mediated diffusion in disordered alloys: Ge self-diffusion
in Si1–xGex,” Physical Review B, vol. 65, no. 19, p. 193306, 2002.
[41] P. Ramanarayanan, K. Cho, and B. M. Clemens, “Effect of composition on vacancy
mediated diffusion in random binary alloys: First principles study of the Si1–xGex
system,” Journal of Applied Physics, vol. 94, no. 1, pp. 174–185, 2003.
[42] M. Haran, J. A. Catherwood, and P. Clancy, “Effects of Ge content on the diffusion of
group-V dopants in SiGe alloys,” Applied Physics Letters, vol. 88, no. 17, p. 173502,
2006.

230 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[43] P. F. Fazzini et al., “Effect of Germanium content and strain on the formation of
extended defects in ion implanted Silicon/Germanium,” Thin Solid Films, vol. 518,
no. 9, pp. 2338–2341, 2010.
[44] R. T. Crosby, Evolution of Self-Interstitials Induced by Ion-Implantation in SiGe Alloys,
Ph.D. thesis, University of Florida, Gainesville, FL, USA, 2005.
[45] A. Belafhaili et al., “Influence of the Germanium content on the amorphization of
silicon–germanium alloys during ion implantation,” Materials Science in
Semiconductor Processing, vol. 16, no. 6, pp. 1655–1658, 2013.
[46] N. Zographos et al., “Multiscale modeling of doping processes in advanced
semiconductor devices,” Materials Science in Semiconductor Processing, vol. 62,
pp. 49–61, May 2017.
[47] K. Y. Suh and H. H. Lee, “Composition dependence of activation energy in solid phase
epitaxial growth of Si1–xGex alloys,” Journal of Applied Physics, vol. 80, no. 12,
pp. 6716–6719, 1996.
[48] T. E. Haynes et al., “Composition dependence of solid-phase epitaxy in silicon-
germanium alloys: Experiment and theory,” Physical Review B, vol. 51, no. 12,
pp. 7762–7772, 1995.
[49] R. Braunstein, A. R. Moore, and F. Herman, “Intrinsic Optical Absorption in
Germanium-Silicon Alloys,” Physical Review, vol. 109, no. 3, pp. 695–710, 1958.
[50] P. Castrillo et al., “Physical modeling and implementation scheme of native defect
diffusion and interdiffusion in SiGe heterostructures for atomistic process simulation,”
Journal of Applied Physics, vol. 109, p. 103502, May 2011.
[51] P. Castrillo et al., “Atomistic Modeling of Defect Diffusion in SiGe,” in International
Conference on Simulation of Semiconductor Processes and Devices (SISPAD), Vienna,
Austria, pp. 9–12, September 2007.
[52] P. Castrillo et al., “Atomistic modeling of defect diffusion and interdiffusion in SiGe
heterostructures,” Thin Solid Films, vol. 518, no. 9, pp. 2448–2453, 2010.
[53] R. Kube et al., “Composition dependence of Si and Ge diffusion in relaxed Si1–xGex
alloys,” Journal of Applied Physics, vol. 107, no. 7, p. 073520, 2010.
[54] Y. Dong et al., “A unified interdiffusivity model and model verification for tensile and
relaxed SiGe interdiffusion over the full germanium content range,” Journal of Applied
Physics, vol. 111, no. 4, p. 044909, 2012.
[55] G. Xia et al., “Strain dependence of Si–Ge interdiffusion in epitaxial Si/Si1–yGey/Si
heterostructures on relaxed Si1–xGex substrates,” Applied Physics Letters, vol. 88, no. 1,
p. 013507, 2006.
[56] G. Xia, J. L. Hoyt, and M. Canonico, “Si-Ge interdiffusion in strained Si/strained SiGe
heterostructures and implications for enhanced mobility metal-oxide-semiconductor
field-effect transistors,” Journal of Applied Physics, vol. 101, p. 044901, February 2007.

Advanced Calibration for Process Simulation User Guide 231

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[57] N. E. B. Cowern et al., “Diffusion in Strained Si(Ge),” Physical Review Letters, vol. 72,
no. 16, pp. 2585–2588, 1994.
[58] H. H. Silvestri et al., “Diffusion of silicon in crystalline germanium,” Semiconductor
Science and Technology, vol. 21, no. 6, p. 758, 2006.
[59] M. Vos et al., “Channeling studies of implantation damage in SiGe superlattices and
SiGe alloys,” Nuclear Instruments and Methods in Physics Research B, vol. 66, no. 3,
pp. 361–368, 1992.
[60] B. Sklenard et al., “Atomistic investigation of the impact of stress during solid phase
epitaxial regrowth,” Physica Status Solidi C, vol. 11, no. 1, pp. 97–100, 2014.
[61] M. Diebel and S. T. Dunham, “Ab-initio Calculations to Predict Stress Effects on
Defects and Diffusion in Silicon,” in International Conference on Simulation of
Semiconductor Processes and Devices (SISPAD), Boston, MA, USA, pp. 147–150,
September 2003.
[62] M. Diebel, Application of Ab-initio Calculations to Modeling of Nanoscale Diffusion
and Activation in Silicon, Ph.D. thesis, University of Washington, Seattle, WA, USA,
2004.
[63] Y. Kawamura et al., “Self-diffusion in compressively strained Ge,” Journal of Applied
Physics, vol. 110, p. 034906, August 2011.
[64] V. Moroz et al., “Dissolution of extended defects in strained silicon,” Journal of
Vacuum Science & Technology B, vol. 26, no. 1, pp. 439–442, 2008.
[65] P. Pichler, Intrinsic Point Defects, Impurities, and Their Diffusion in Silicon,
Computational Microelectronics, Vienna: Springer, 2004.
[66] J. Zhu et al., “Ab initio pseudopotential calculations of B diffusion and pairing in Si,”
Physical Review B, vol. 54, no. 7, p. 4741–4747, 1996.
[67] J. S. Nelson, P. A. Schultz, and A. F. Wright, “Valence and atomic size dependent
exchange barriers in vacancy-mediated dopant diffusion,” Applied Physics Letters,
vol. 73, no. 2, pp. 247–249, 1998.
[68] V. C. Venezia et al., “Boron diffusion in amorphous silicon,” Materials Science and
Engineering B, vol. 124–125, pp. 245–248, December 2005.
[69] R. Duffy et al., “Dopant diffusion in amorphous silicon,” in MRS Symposium
Proceedings, Silicon Front-End Junction Formation—Physics and Technology, San
Francisco, CA, USA, vol. 810, pp. C10.2.1–C10.2.6, April 2004.
[70] S. Uppal et al., “Diffusion of boron in germanium at 800–900oC,” Journal of Applied
Physics, vol. 96, no. 3, pp. 1376–1380, 2004.
[71] I. Martin-Bragado, N. Zographos, and P. Castrillo, “Atomistic modelling and simulation
of arsenic diffusion including mobile arsenic clusters,” Physica Status Solidi A, vol. 211,
no. 1, pp. 147–151, 2014.

232 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[72] B. Sahli et al., “Ab initio calculations of phosphorus and arsenic clustering parameters
for the improvement of process simulation models,” Materials Science and Engineering
B, vol. 154-155, pp. 193–197, December 2008.
[73] D. C. Mueller, E. Alonso, and W. Fichtner, “Arsenic deactivation in Si: Electronic
structure and charge states in vacancy-impurity clusters,” Physical Review B, vol. 68,
no. 4, p. 045208, 2003.
[74] A. N. Larsen et al., “Heavy doping effects in the diffusion of group IV and V impurities
in silicon,” Journal of Applied Physics, vol. 73, no. 2, pp. 691–698, 1993.
[75] N. Zographos and A. Erlebach, “Process simulation of dopant diffusion and activation
in germanium,” Physica Status Solidi A, vol. 211, no. 1, pp. 143–146, 2014.
[76] A. Chroneos, “Dopant-defect interactions in Ge: Density functional theory
calculations,” Materials Science in Semiconductor Processing, vol. 15, no. 6,
pp. 691–696, 2012.
[77] A. Chroneos et al., “Engineering the free vacancy and active donor concentrations in
phosphorus and arsenic double donor-doped germanium,” Journal of Applied Physics,
vol. 104, no. 11, p. 113724, 2008.
[78] N. Zographos and I. Martin-Bragado, “Atomistic Modeling of Carbon Co-Implants and
Rapid Thermal Anneals in Silicon,” in 15th IEEE International Conference on
Advanced Thermal Processing of Semiconductors (RTP), Catania, Italy, pp. 119–122,
October 2007.
[79] M. Diebel and S. T. Dunham, “Ab Initio Calculations to Model Anomalous Fluorine
Behavior,” Physical Review Letters, vol. 93, no. 24, p. 245901, 2004.
[80] M. Diebel and S. T. Dunham, “Reply to Ab Initio Calculations to Model Anomalous
Fluorine Behavior,” Physical Review Letters, vol. 96, p. 039602, January 2006.
[81] K. Vollenweider et al., “Fluorine clustering and diffusion in silicon: Ab initio
calculations and kinetic Monte Carlo model,” Journal of Vacuum Science & Technology
B, vol. 28, no. 1, pp. C1G1–C1G6, 2010.
[82] G. R. Nash et al., “Activation energy for fluorine transport in amorphous silicon,”
Applied Physics Letters, vol. 75, no. 23, pp. 3671–3673, 1999.
[83] P. A. Schultz and J. S. Nelson, “Fast through-bond diffusion of nitrogen in silicon,”
Applied Physics Letters, vol. 78, no. 6, pp. 736–738, 2001.
[84] N. Stoddard et al., “Ab Initio Identification of the Nitrogen Diffusion Mechanism in
Silicon,” Physical Review Letters, vol. 95, no. 2, p. 025901, 2005.
[85] L. S. Adam et al., “Physical integrated diffusion-oxidation model for implanted nitrogen
in silicon,” Journal of Applied Physics, vol. 91, no. 4, pp. 1894–1900, 2002.
[86] L. Pelaz et al., “B diffusion and clustering in ion implanted Si: The role of B cluster
precursors,” Applied Physics Letters, vol. 70, no. 17, pp. 2285–2287, 1997.

Advanced Calibration for Process Simulation User Guide 233

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[87] C. Zechner et al., “Simulation of dopant diffusion and activation during flash lamp
annealing,” Materials Science and Engineering B, vol. 154-155, pp. 20–23, December
2008.
[88] M. Aboy et al., “Kinetics of large B clusters in crystalline and preamorphized silicon,”
Journal of Applied Physics, vol. 110, no. 7, p. 073524, 2011.
[89] J. Xia et al., “Boron segregation to extended defects induced by self-ion implantation
into silicon,” Journal of Applied Physics, vol. 85, no. 11, pp. 7597–7603, 1999.
[90] A. Satta et al., “Diffusion, activation, and recrystallization of boron implanted in
preamorphized and crystalline germanium,” Applied Physics Letters, vol. 87, no.17,
p. 172109, 2005.
[91] B. R. Yates et al., “Anomalous activation of shallow B+ implants in Ge,” Materials
Letters, vol. 65, no. 23–24, pp. 3540–3543, 2011.
[92] S. A. Harrison, T. F. Edgar, and G. S. Hwang, “Interstitial-Mediated Arsenic Clustering
in Ultrashallow Junction Formation,” Electrochemical and Solid-State Letters, vol. 9,
no. 12, pp. G354–G357, 2006.
[93] R. Pinacho et al., “Modeling arsenic deactivation through arsenic-vacancy clusters using
an atomistic kinetic Monte Carlo approach,” Applied Physics Letters, vol. 86, p. 252103,
June 2005.
[94] J. Xie and S. P. Chen, “Diffusion and Clustering in Heavily Arsenic-Doped Silicon:
Discrepancies and Explanation,” Physical Review Letters, vol. 83, no. 9, pp. 1795–1798,
1999.
[95] Y. Park et al., “Dopant diffusion in Si, SiGe and Ge: TCAD model parameters
determined with density functional theory,” IEDM Technical Digest, San Francisco,
CA, USA, pp. 784–787, December 2017.
[96] T. Noda, S. Odanaka, and H. Umimoto, “Effects of end-of-range dislocation loops on
transient enhanced diffusion of indium implanted in silicon,” Journal of Applied
Physics, vol. 88, no. 9, pp. 4980–4984, 2000.
[97] C.-L. Liu et al., “Ab initio modeling and experimental study of C–B interactions in Si,”
Applied Physics Letters, vol. 80, no. 1, pp. 52–54, 2002.
[98] Y. Shimizu et al., “Impact of carbon coimplantation on boron behavior in silicon:
Carbon–boron coclustering and suppression of boron diffusion,” Applied Physics
Letters, vol. 98, no. 23, p. 232101, 2011.
[99] Y.-S. Oh and D. E. Ward, “A Calibrated Model for Trapping of Implanted Dopants at
Material Interface During Thermal Annealing,” in IEDM Technical Digest, San
Francisco, CA, USA, pp. 509–512, December 1998.
[100] V. C. Venezia et al., “Dopant redistribution effects in preamorphized silicon during low
temperature annealing,” in IEDM Technical Digest, Washington, DC, USA,
pp. 489–492, December 2003.

234 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[101] S. H. Jain et al., “Metastable boron active concentrations in Si using flash assisted solid
phase epitaxy,” Journal of Applied Physics, vol. 96, no. 12, pp. 7357–7360, 2004.
[102] D. De Salvador et al., “Experimental evidence of B clustering in amorphous Si during
ultrashallow junction formation,” Applied Physics Letters, vol. 89, p. 241901,
December 2006.
[103] C. Zechner, D. Matveev, and A. Erlebach, “Phase-field model for the dopant
redistribution during solid phase epitaxial regrowth of amorphized silicon,” Materials
Science and Engineering B, vol. 114–115, pp. 162–165, 2004.
[104] O. Dokumaci et al., “Transient Enhanced Diffusion and Dose Loss of Indium in
Silicon,” in MRS Symposium Proceedings, Si Front-End Processing—Physics and
Technology of Dopant-Defect Interactions, San Francisco, CA, USA, vol. 568,
pp. 205–210, April 1999.
[105] R. Duffy et al., “Influence of preamorphization and recrystallization on indium doping
profiles in silicon,” Journal of Vacuum Science & Technology B, vol. 22, no. 3,
pp. 865–868, 2004.
[106] G. Impellizzeri et al., “Fluorine in preamorphized Si: Point defect engineering and
control of dopant diffusion,” Journal of Applied Physics, vol. 99, p. 103510, May 2006.
[107] Z. Ye et al., “High Tensile Strained In-Situ Phosphorus Doped Silicon Epitaxial Film
for nMOS Applications,” ECS Transactions, vol. 50, no. 9, pp. 1007–1011, 2013.
[108] X. Li et al., “Selective Epitaxial Si:P Film for nMOSFET Application: High
Phosphorous Concentration and High Tensile Strain,” ECS Transactions, vol. 64, no. 6,
pp. 959–965, 2014.
[109] S. K. Dhayalan et al., “On the Evolution of Strain and Electrical Properties in As-Grown
and Annealed Si:P Epitaxial Films for Source-Drain Stressor Applications,” ECS
Journal of Solid State Science and Technology, vol. 7, no. 5, pp. P228–P237, 2018.
[110] C. Ahn, Atomic scale modeling of stress and pairing effects on dopant behavior in
silicon, Ph.D. thesis, University of Washington, Seattle, WA, USA, 2007.
[111] R. F. Lever, J. M. Bonar, and A. F. W. Willoughby, “Boron diffusion across
silicon–silicon germanium boundaries,” Journal of Applied Physics, vol. 83, no. 4,
pp. 1988–1994, 1998.
[112] L. A. Edelman et al., “Boron diffusion in amorphous silicon-germanium alloys,”
Applied Physics Letters, vol. 92, no. 17, p. 172108, 2008.
[113] P. Laitinen, Self- and Impurity Diffusion in Intrinsic Relaxed Silicon - Germanium,
Ph.D. thesis, University of Jyväskylä, Finland, January 2004.
[114] S. Eguchi et al., “Comparison of arsenic and phosphorus diffusion behavior in
silicon–germanium alloys,” Applied Physics Letters, vol. 80, no. 10, pp. 1743–1745,
2002.

Advanced Calibration for Process Simulation User Guide 235

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[115] Y.-M. Sheu et al., “Modeling Dopant Diffusion in Strained and Strain-Relaxed Epi-
SiGe,” in International Conference on Simulation of Semiconductor Process and
Devices (SISPAD), Tokyo, Japan, pp. 75–78, September 2005.
[116] N. Zangenberg, Defect and Diffusion Studies in Si and SiGe, Ph.D. thesis, University of
Aarhus, Denmark, January 2003.
[117] P. Kuo et al., “Boron Diffusion in Si and Si1–xGex,” in MRS Symposium Proceedings,
Strained Layer Epitaxy - Materials, Processing, and Device Applications, vol. 379,
pp. 373–378, 1995.
[118] S. Uppal et al., “Diffusion of Boron in Germanium and Si1–xGex (x>50%) alloys,” in
MRS Symposium Proceedings, CMOS Front-End Materials and Process Technology,
vol. 765, p. D6.16.1, January 2003.
[119] C. C. Wang et al., “Boron diffusion in strained and strain-relaxed SiGe,” Materials
Science and Engineering B, vol. 124–125, pp. 39–44, December 2005.
[120] J. S. Christensen et al., “Diffusion of phosphorus in relaxed Si1–xGex films and strained
Si/Si1–xGex heterostructures,” Journal of Applied Physics, vol. 94, no. 10,
pp. 6533–6540, 2003.
[121] I. Martin-Bragado et al., “Anisotropic dopant diffusion in Si under stress using both
continuum and atomistic methods,” Journal of Computational Electronics, vol. 7, no. 3,
pp. 103–106, 2008.
[122] C. Ahn et al., “Stress effects on impurity solubility in crystalline materials: A general
model and density-functional calculations for dopants in silicon,” Physical Review B,
vol. 79, no. 7, p. 073201, 2009.
[123] L. Lin et al., “Boron diffusion in strained Si: A first-principles study,” Journal of
Applied Physics, vol. 96, no. 10, pp. 5543–5547, 2004.
[124] S. T. Dunham et al., “Calculations of effect of anisotropic stress/strain on dopant
diffusion in silicon under equilibrium and nonequilibrium conditions,” Journal of
Vacuum Science & Technology B, vol. 24, no. 1, pp. 456–461, 2006.
[125] N. Sugii et al., “Recrystallization, redistribution, and electrical activation of strained-
silicon/Si0.7Ge0.3 heterostructures with implanted arsenic,” Journal of Applied Physics,
vol. 96, no. 1, pp. 261–268, 2004.
[126] I. Martin-Bragado and V. Moroz, “Modeling of {311} facets using a lattice kinetic
Monte Carlo three-dimensional model for selective epitaxial growth of silicon,” Applied
Physics Letters, vol. 98, no. 15, p. 153111, 2011.
[127] P. M. Garone et al., “Silicon vapor phase epitaxial growth catalysis by the presence of
germane,” Applied Physics Letters, vol. 56, no. 13, pp. 1275–1277, 1990.
[128] M. Liehr et al., “Kinetics of silicon epitaxy using SiH4 in a rapid thermal chemical vapor
deposition reactor,” Applied Physics Letters, vol. 56, no. 7, pp. 629–631, 1990.

236 Advanced Calibration for Process Simulation User Guide

P-2019.03
 6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[129] H.-C. Tseng et al., “Effects of isolation materials on facet formation for silicon selective
epitaxial growth,” Applied Physics Letters, vol. 71, no. 16, pp. 2328–2330, 1997.
[130] R. Chen et al., “A New Kinetic Lattice Monte Carlo Modeling Framework for the
Source-Drain Selective Epitaxial Growth Process,” in International Conference on
Simulation of Semiconductor Processes and Devices (SISPAD), Glasgow, Scotland,
pp. 9–12, September 2013.
[131] D. Dutartre, A. Talbot, and N. Loubet, “Facet Propagation in Si and SiGe Epitaxy or
Etching,” ECS Transactions, vol. 3, no. 7, pp. 473–487, 2006.
[132] J. M. Hartmann et al., “Growth kinetics of Si and SiGe on Si(100), Si(110) and Si(111)
surfaces,” Journal of Crystal Growth, vol. 294, no. 2, pp. 288–295, 2006.
[133] J. M. Hartmann et al., “A benchmarking of silane, disilane and dichlorosilane for the low
temperature growth of group IV layers,” Thin Solid Films, vol. 520 no. 8,
pp. 3185–3189, 2012.
[134] J. M. Hartmann et al., “SiGe growth kinetics and doping in reduced pressure-chemical
vapor deposition,” Journal of Crystal Growth, vol. 236, no. 1–3, pp. 10–20, 2002.
[135] J. M. Hartmann et al., “Potentialities of disilane for the low temperature epitaxy of
intrinsic and boron-doped SiGe,” Thin Solid Films, vol. 557, pp. 19–26, April 2014.
[136] J. M. Hartmann et al., “Disilane-based cyclic deposition/etch of Si, Si:P and Si1–yCy:P
layers: I. The elementary process steps,” Semiconductor Science and Technology,
vol. 28, no. 2, p. 025017, 2013.
[137] Y. Bogumilowicz and J. M. Hartmann, “Reduced-pressure chemical vapor deposition of
boron-doped Si and Ge layers,” Thin Solid Films, vol. 557, pp. 4–9, April 2014.
[138] S. Fukatsu et al., “Self-limitation in the surface segregation of Ge atoms during Si
molecular beam epitaxial growth,” Applied Physics Letters, vol. 59, no. 17,
pp. 2103–2105, 1991.
[139] J. Aubin et al., “Very low temperature epitaxy of Ge and Ge rich SiGe alloys with Ge2H6
in a Reduced Pressure – Chemical Vapour Deposition tool,” Journal of Crystal Growth,
vol. 445, pp. 65–72, July 2016.
[140] Y. Bogumilowicz et al., “High-temperature growth of very high germanium content
SiGe virtual substrates,” Journal of Crystal Growth, vol. 290, no. 2, pp. 523–531, 2006.
[141] J. M. Hartmann, “Low temperature growth kinetics of high Ge content SiGe in reduced
pressure-chemical vapor deposition,” Journal of Crystal Growth, vol. 305, no. 1,
pp. 113–121, 2007.
[142] J. M. Hartmann et al., “Reduced Pressure–Chemical Vapour Deposition of Si/SiGe
heterostructures for nanoelectronics,” Materials Science and Engineering B,
vol. 154–155, pp. 76–84, December 2008.

Advanced Calibration for Process Simulation User Guide 237

P-2019.03
6: Contents of Advanced Calibration of Sentaurus Process KMC
References

[143] J. M. Hartmann et al., “Growth kinetics and boron doping of very high Ge content SiGe
for source/drain engineering,” Journal of Crystal Growth, vol. 310, no. 1, pp. 62–70,
2008.
[144] J. M. Hartmann et al., “HCl selective etching of SiGe versus Si in stacks grown on
(110),” ECS Transactions, vol. 33, no. 6, pp. 985–993, 2010.
[145] N. Loubet, A. Talbot, and D. Dutartre, “Silicon Etching Study in a RT-CVD Reactor
with the HCl/H2 Gas Mixture,” in MRS Symposium Proceedings, vol. 910, January
2006.
[146] V. Destefanis et al., “Low-thermal surface preparation, HCl etch and Si/SiGe selective
epitaxy on (1 1 0) silicon surfaces,” Semiconductor Science and Technology, vol. 23,
no. 10, p. 105018, 2008.
[147] Y. Bogumilowicz et al., “Chemical vapour etching of Si, SiGe and Ge with HCl;
applications to the formation of thin relaxed SiGe buffers and to the revelation of
threading dislocations,” Semiconductor Science and Technology, vol. 20, no. 2,
pp.127–134, 2005.

238 Advanced Calibration for Process Simulation User Guide

P-2019.03
CHAPTER 7 Guidelines for Additional
Calibration

This chapter provides guidelines for additional calibration.

The Advanced Calibration file is the recommended starting point for accurate process
simulation with Sentaurus Process Kinetic Monte Carlo (Sentaurus Process KMC). However,
Advanced Calibration cannot fully replace an additional calibration by the user. With a
customized process calibration, the accuracy can always be further increased for any
technology of interest. A customized calibration of process and device simulation models
needs to be performed by the user or can be requested from Synopsys in the context of a
customer service project.

To further improve the Advanced Calibration, Synopsys appreciates feedback from customers
regarding the accuracy obtained with the parameter files for different process conditions, and
suggestions for improved models or parameter values.

Accuracy and Limitations of Advanced Calibration of

Sentaurus Process KMC
The Advanced Calibration is based on scientific literature of process simulation models and on
a continual calibration effort based on the Synopsys collection of SIMS profiles from state-of-
the-art device manufacturing technology. A good agreement is obtained for a wide range of
SIMS data. However, in some cases, there is a significant mismatch between kinetic Monte
Carlo simulation results obtained with Advanced Calibration and the experimental data, for
similar reasons as for continuum simulations listed in Accuracy and Limitations of Advanced
Calibration of Sentaurus Process on page 107.

In this section, the accuracy of the Advanced Calibration for Sentaurus Process KMC is
discussed. In particular, it will explain for which process conditions the accuracy is limited and
which parameters can be fine-tuned by users to increase the accuracy in a process window of
interest.

Advanced Calibration for Process Simulation User Guide 239

P-2019.03
7: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC

Damage and Point Defects

Amorphization

The calibration of amorphization and recrystallization in silicon follows Zographos and

Martin-Bragado [1]. In Figure 34, the accuracy of the damage accumulation during
implantation is shown by the comparison of literature data with simulated amorphization-layer
thicknesses generated by germanium implantation.
Experiments
Sentaurus MC
80
Crystal−TRIM
Amorphous Layer Thickness [nm]

60

40

20

0
0 20 40 60
Implantation Energy [keV]

Figure 34 Amorphous layer thickness versus Ge implantation energy, dose 1x1015 cm–2:
comparison of experiments (squares) and simulations (solid lines) at dose
rate of 5x1012 cm–2 s–1

Obviously, the implantation dose rate as well as the implantation temperature strongly
influence the amorphization and need to be considered by the user.
12 –2 –1
NOTE The default implantation dose rate is 1 × 10 cm s and is not
changed within Advanced Calibration because it is considered to be a
process parameter.

To adjust the dose rate, define the dose.rate argument of the implant command or redefine
the following procedure before the implant command:
proc DoseRate { dose } {
set refDose 5e13 ; # dose rate of 5e13 cm-2 s-1
expr $dose/$refDose
}

240 Advanced Calibration for Process Simulation User Guide

P-2019.03
 7: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC

If the dose rate and temperature are not known exactly, they can be used as fitting parameters
to adjust the damage and amorphization-layer thicknesses of implantations.

Germanium Preamorphization

In the case of germanium preamorphization implantation, the as-implanted germanium profiles

for Sentaurus MC implantation are often slightly shallower than corresponding SIMS profiles
and, therefore, the amorphization-layer thickness can also be slightly shallower.

The lower amorphization can be compensated partly by lowering the displacement threshold
for the damage generation of Sentaurus MC implantation from 15 eV to 13 eV. However, the
as-implanted germanium profiles for Crystal-TRIM implantation are, in general, slightly
deeper and more accurate compared to SIMS and, therefore, could be used as an alternative.
The difference between the amorphous-layer thicknesses for the two different Monte Carlo
implant engines can be seen in Figure 34 on page 240.

Cold and Hot Implantation

The damage accumulation has been calibrated on data of different implantation dose rates and
temperature [1]. Therefore, the trends for implantation temperature dependency should be
modeled accurately, for example, the increased amorphization during implantation at cold
temperatures (lower than room temperature) [2], the decreased amorphization [3], and
enhanced channeling during implantation at hot temperatures [4].

If the amorphization is underestimated, the displacement threshold for the damage generation
of Sentaurus MC implantation can be decreased:
pdbSet Si B casc.dis 670

If the implant channeling is overestimated, the proportional factor for the switching probability
from the crystalline to the amorphous model can be increased:
pdbSet Si B casc.amo 4.0

Recrystallization

While the recrystallization velocity for an undoped amorphous–silicon layer on top of (100)-
oriented crystalline silicon is well known [5] and might not require any changes, its
enhancement in the presence of n-type and p-type doping is less certain [6][7]. Therefore, the
corresponding silicon damage parameters V0.Recryst.ntype and V0.Recryst.ptype
can be used as fitting parameters. Since the diffusivities of dopants in crystallized and
amorphous silicon are usually very different, the recrystallization velocity can influence the
overall diffusion.

Advanced Calibration for Process Simulation User Guide 241

P-2019.03
7: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC

It is also known that the recrystallization depends on substrate orientation; however, by default,
the recrystallization velocity in Advanced Calibration for Sentaurus Process KMC is
independent of it. To enable the substrate orientation dependency, you should select:
pdbSet KMC Si Damage SPER.Model LKMC

Diffusion, Generation, and Recombination

The parameter values for the equilibrium concentration and diffusivity of point defects have
been inherited from continuum Advanced Calibration and have been chosen by Synopsys as a
careful compromise between various suggestions in recent publications (see Point Defects on
page 108). The I–V pair recombination rate is part of the amorphization calibration (see
Amorphization on page 240).

Changing any of these parameters might affect the diffusion and activation behavior of several
dopants. Therefore, for the purpose of improving the accuracy of diffusion of one dopant, it is
not recommended to change point-defect parameters, due to the possible unwanted effects on
other dopants.

Extended Defects
Like the advanced TED model in the Advanced Calibration of continuum Sentaurus Process
(see Interstitial Clusters on page 69), the comprehensive KMC model for extended defects of
self-interstitials in silicon [8] describes accurately the initial phase of ultrahigh interstitial
supersaturation after ion implantation, which was reported in [9] and is ascribed to the
formation and dissolution of small interstitial clusters (see Figure 35).
108

107
Interstitial Supersaturation

106

105

104

103

102

101
100 101 102 103 104 105
Time [s]

Figure 35 Self-interstitial supersaturation evolution: comparison of experiments (squares)

and simulations (solid lines) for anneals at 600oC (red), 700oC (green), and 800oC
(blue) (experimental data from [9])

242 Advanced Calibration for Process Simulation User Guide

P-2019.03
 7: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC

The binding energies of larger cluster sizes have been calibrated from transmission electron
microscope (TEM) data on the dissolution of {311} defects published by Stolk et al. [10] and
Saleh et al. [11].

The calibration of the transformation rate from {311} defects to dislocation loops is based on
data from the literature [12][13][14][15]. Due to low dislocation loop statistics in KMC, this
calibration is difficult and not as accurate as for {311} dissolution. Therefore, the following
parameters can be used as fitting parameters:
pdbSet KMC Si Int D0.311.To.Loop 4500
pdbSet KMC Si Int E.311.To.Loop 0.0

The parameters for voids have been calibrated on a limited experimental dataset only [16] and,
therefore, are subject to additional calibration in the case of vacancy engineering implantation.

Impurities

Diffusion

In general, the diffusivities are inherited from Advanced Calibration of continuum Sentaurus
Process and are based on the literature and SIMS calibration. For details about the translation
of the diffusivities to the migration and binding parameters for Sentaurus Process KMC, see
Diffusion on page 198.

For fine-tuning of the diffusivity, the binding, migration, and ionization energies of
dopant–point defect pairs can be adjusted slightly. The influence of the binding, migration, and
ionization energies is shown in Figure 36.

1021 1021 1021

1020 1020 1020

Eb+0.2 eV Em−0.2 eV
Concentration [cm−3]
Concentration [cm−3]

Concentration [cm−3]

e0+0.2 eV
1019 1019 1019
Eb Em e0

1018 1018 1018

Eb−0.2 eV
Em+0.2 eV e0−0.2 eV

1017 1017 1017

1016 1016 1016

0 50 100 0 50 100 0 50 100
Depth [nm] Depth [nm] Depth [nm]

Figure 36 The influence of the binding (left), migration (middle), and ionization (right)
energies on boron diffusion in silicon; the process conditions are B 1e15, 0.5 keV
implantation followed by 1050oC spike anneal

Advanced Calibration for Process Simulation User Guide 243

P-2019.03
7: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC

The higher the binding energy, the more stable is the mobile pair and, therefore, the dopant
encounters more diffusion. In contrast, the higher the migration barrier, the less the pair is
mobile and, therefore, the dopant diffuses less.

A lower ionization level of a charge state allows that charge state to be more populated and,
therefore, to have more influence on the overall diffusion. If the different charge states have
different migration rates, changing the ionization levels results in a different overall migration
rate.

NOTE Changing the binding energy and ionization levels of pairs also changes
the impurity cluster formation, in particular, for reactions where the pair
is involved. Since the binding energies of clusters are computed using
the potential energies of the reactant, the potential energy of the pair
–
(given by – E b ( A i ) – e ( – , 0 ) ( A i ) for negative substitutional A ) also
influences the cluster stability. In addition, the migration barrier of the
pair is involved in the emission of pairs from impurity clusters.

Antimony

Advanced Calibration for Sentaurus Process KMC includes no parameters for antimony
diffusion because the default parameters are based already on the translation of the continuum
Advanced Calibration parameters and are verified against SIMS data.

Co-Diffusion of Arsenic and Phosphorus

As elaborated in Co-Diffusion of Arsenic and Phosphorus on page 123, the physics of the co-
diffusion of arsenic and phosphorus in high concentration is very complex for several reasons.
In general, Advanced Calibration for Sentaurus Process KMC underestimates the diffusion of
arsenic in the presence of phosphorus at high concentrations. One reason for this might be the
lack of an arsenic percolation-type diffusion model.

Diffusion in Amorphous Silicon

In Advanced Calibration for Sentaurus Process KMC, boron and fluorine are the only diffusing
species in amorphous silicon. There is no evidence for significant diffusion of other
species [17].

The boron diffusion, in combination with clustering, in amorphous silicon was calibrated [18]
using literature data on completely amorphized SOI structures [19][20] and deeply
preamorphized bulk silicon structures [17]. The calibrated parameters for diffusion and
clustering mediated by dangling bonds allow for good general agreement with experiments.
The transient behavior of boron diffusion is well reproduced by the annihilation of dangling
and floating bonds.

244 Advanced Calibration for Process Simulation User Guide

P-2019.03
 7: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC

The coefficient for dangling bond creation per B atom in amorphous silicon is set to 0.5 in
Advanced Calibration to obtain good accuracy with SIMS for boron implantation into
amorphous silicon [19]. However, for amorphization of boron marker layers, a coefficient of
1.0 gives better accuracy, as reported in [18][20]. Therefore, the parameter:
pdbSet KMC aSi B gamma 1.0

can be used as a fitting parameter, similar to the ifactor for ion implantation in continuum
process simulation.

Activation

The responsible mechanism for deactivation/activation is the clustering/declustering of

dopant–defect clusters. In Sentaurus Process KMC, the parameters for the dopant–defect
cluster model are the potential energies, the emission prefactors, and the capture volumes. For
recalibration, the most recommended fitting parameter is the emission prefactor D0.Cluster
for dopant–defect pairs, while the one for point defects should not be changed for reasons of
microscopic reversability.

Decreasing the emission prefactor results in decreasing the declustering rate, making the
clusters more stable. Moreover, the solubility of impurities is governed by the potential
energies of the clusters. Since most dopants can form several different cluster types, the
potential energies of more than one cluster can influence the overall deactivation/activation.

Therefore, the best way to tune the solid solubility of dopants is either to apply an overall shift
for all cluster energies or to change the potential energy of the most stable clusters. The
influence of the emission prefactor, the potential energies of all clusters, and the potential
energy of the most stable cluster is illustrated in Figure 37.

1021 1021 1021

1020 1020 1020

Concentration [cm−3]

Concentration [cm−3]

Concentration [cm−3]

2*D0 Epot+0.5 eV E(B3I)+0.5 eV

1019 1019 1019
0.5*D0

Epot−0.5 eV
1018 1018 1018
E(B3I)−0.5 eV

1017 1017 1017

1016 1016 1016

0 50 100 0 50 100 0 50 100
Depth [nm] Depth [nm] Depth [nm]

Figure 37 The influence of the BI emission prefactor (left), BIC potential energies (middle),
and B3I potential energy (right) on boron diffusion; the process conditions are B
1e15, 0.5 keV implantation into silicon followed by 1050oC spike anneal

Finally, the capture volumes, which are specified relative to the substitutional dopant, are based
on realistic guesses and can be changed only slightly.

Advanced Calibration for Process Simulation User Guide 245

P-2019.03
7: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC

Segregation at End-of-Range Defects

Segregation of impurities to end-of-range defects is modeled by an interaction of the mobile

impurity–interstitial pairs with extended defects such as {311} defects and dislocation loops.
The pair breaks up and the interstitial is incorporated into the extended defect, while the
impurity is regarded as substitutional. Obviously, the location of the impurity segregation
depends on the end-of-range location, which depends on the amorphization-layer thickness
generated by amorphizing implantation.

For carbon, this model works reasonably accurately [21]. However, in the case of indium, the
segregation effect seems to be stronger [22] and can be modeled by selecting the following
model:
pdbSet KMC Si In 311.Dopant.Model 1
pdbSet KMC Si In Loop.Dopant.Model 1

In this case, the neutral pair does not break up, but it is attached to the extended defect and stays
attached until the dissolution of the extended defect.

Boron Interaction With Fluorine

It is known that the presence of fluorine can reduce diffusion and enhance the activation of
boron. In Advanced Calibration, no direct interaction between boron and fluorine is assumed,
as suggested by [23]. However, direct boron–fluorine interaction has been proposed by
others [24]. Therefore, similar to boron–carbon clusters as proposed by [25] and described in
Boron and Carbon on page 211, you can study the direct interaction of boron with fluorine. To
add fluorine–boron–interstitial clusters, for example BFI and BF, you must define the allowed
reactions:
pdbSet KMC Si B ReactionsPointDefect Bi,F true
pdbSet KMC Si F ReactionsPointDefect Fi,B true
pdbSet KMC Si B ReactionsCluster BF,I true
pdbSet KMC Si B ReactionsCluster BFI,V true

and specify the corresponding binding energy, capture volume, and emission prefactors:
pdbSet KMC Si B Etotal.Complex BF -0.7
pdbSet KMC Si B Etotal.Complex BFI -5.6
pdbSet KMC Si B CaptVol.Complex BF 1.3
pdbSet KMC Si B CaptVol.Complex BFI 1.5
pdbSet KMC Si B D0.Cluster BF,Bi 0.01
pdbSet KMC Si B D0.Cluster BF,Fi 0.0001
pdbSet KMC Si B D0.Cluster BF,I 0.5
pdbSet KMC Si B D0.Cluster BF,V 0.5

246 Advanced Calibration for Process Simulation User Guide

P-2019.03
 7: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC

Dose Loss

The dopant dose loss is controlled by the three-phase segregation model for the silicon–oxide
interface model. Dopant trapping at the interface is diffusion limited; therefore, the
dopant–point defect migration parameters also have an influence on the dose loss. The
recommended fitting parameters for dose loss are the parameters for the maximum surface
concentration of trapped particles C0Max.Surf and EMax.Surf as well as the parameter for
the interface barrier EBarrier_SurfSi.

Recrystallization

For experiments with amorphizing conditions, the recrystallization parameters allow for a wide
range of optimization, but the complex parameter interactions must be taken into account
carefully. The maximum active doping concentration after recrystallization is an influential and
sometimes critical parameter both for activation and diffusion, and the corresponding
parameters C0.Recryst.Max.Active and E.Recryst.Max.Active serve as fitting
parameters.

However, since the activation of doping during SPER usually is temperature dependent [26],
the accurate anneal temperature at SPER is critical. Therefore, the specification of the accurate
temperature profile during the anneal, especially the initial phase of the ramp-up with
temperatures at 500°C and above, is important. In addition, the doping and impurity
concentrations have an effect on the recrystallization velocity and can influence the activation
during SPER. The corresponding parameters V0.Recryst.ntype and V0.Recryst.ptype
are candidates for recalibration parameters. Finally, the portion of swept impurities by the
recrystallization front can also be adjusted from case to case.

Fluorine Effect

A special case is fluorine with its retarding effect on recrystallization [5]. By strongly slowing
down SPER, the presence of (coimplanted) fluorine can not only result in incorporation of an
higher active doping concentration during temperature ramp-up, but also allow more time for
dopant diffusion in amorphous silicon.

In the case of boron, with its higher diffusivity in amorphous silicon than crystalline
silicon [20], the influence of diffusion in amorphous silicon is enhanced. Therefore, for fine-
tuning, you can tune not only the fluorine parameters for redistribution during SPER
(Pref.RecrystDeposit and Ener.RecrystDeposit), but also the fluorine parameters for
slowing down the recrystallization (E.Recryst and E.Recryst.Exponent).

Advanced Calibration for Process Simulation User Guide 247

P-2019.03
7: Guidelines for Additional Calibration
Accuracy and Limitations of Advanced Calibration of Sentaurus Process KMC

BF2 Implantation

While for boron implantation after preamorphization implantation, no boron is moved by the
recrystallization front, this could be the case for BF2. Therefore, you can allow some portion
of the boron swept by the recrystallization front by adjusting Pref.RecrystDeposit.

In addition, the length of boron diffusion can be controlled by the incorporation of fluorine
during SPER. Most fluorine is incorporated into silicon as fluorine interstitials and as so-called
F–V clusters [27][28]. In the context of Sentaurus Process KMC, the F–V clusters are F–I
clusters (see Fluorine on page 206) containing fewer interstitials than fluorine. Finally, the
amount of interstitials incorporated into F–I clusters during SPER also influences the boron
diffusion.

Stress and SiGe Effects

The calibration of the effects of Ge concentration and stress on defect and dopant diffusion, and
clustering is part of Advanced Calibration for Sentaurus Process KMC. By default, the Ge
effects are modeled by linear correction factors for silicon parameters and are valid only for
low Ge mole fractions ranging between 0% and 50%. Optionally, you can enable Si and Ge
parameter interpolation for SiGe using the following command with the first argument set to
true:
KMC_SiGe_and_Stress_Effect 1 0

This disables the default linear correction factors and uses the interpolation parameters defined
for SiliconGermanium. The calibration covers the full range of the Ge mole fraction
(0–100%).

The stress effects on damage, diffusion, and binding in Advanced Calibration for Sentaurus
Process KMC are taken into account only if the KMC Stress switch is enabled. Stress effects
are recommended to be enabled by using the following command with the second argument set
to true:
KMC_SiGe_and_Stress_Effect 0 1

The calibration of Ge and stress effects [29][30] is based on a relatively small set of measured
data. The models have not yet been tested rigorously against SIMS data for all dopant species
and nonequilibrium annealing conditions, and against electrical data from real-device
fabrication processes. Further calibration for specific applications might be required.

Stress Calculation

Sentaurus Process KMC uses the stress provided by Sentaurus Process, but Sentaurus Process
KMC does not compute it. The stress fields are updated from Sentaurus Process for each
diffuse step.

248 Advanced Calibration for Process Simulation User Guide

P-2019.03
 7: Guidelines for Additional Calibration
References

For calculating stress, it is recommended to use the mechanics parameters from Advanced
Calibration for Sentaurus Process in continuum mode. Therefore, Advanced Calibration
should also be called before enabling the KMC mode by SetAtomistic.

For details about the mechanics parameters, see Parameters for Mechanics on page 261.

Dopant Activation Under Stress

In Advanced Calibration for Sentaurus Process KMC, no stress effects are included for
dopant–defect clusters. However, the solid solubility of dopants in silicon depends on the
strain [31] and is modeled in continuum Advanced Calibration (see Impact of Pressure on
Dopant Activation on page 51). In general, for compressive strain, the solubility of atoms
smaller than Si increases; whereas, the solubility of larger atoms decreases. To change the
cluster stability under strain, the stress correction volumes for potential energy can be defined
for individual clusters:
pdbSet KMC Si As VF.Complex As4V -0.01

References
[1] N. Zographos and I. Martin-Bragado, “A Comprehensive Atomistic Kinetic Monte
Carlo Model for Amorphization/Recrystallization and its Effects on Dopants,” in MRS
Symposium Proceedings, Doping Engineering for Front-End Processing, vol. 1070,
p. 1070-E03-01, 2008.
[2] H.-J. Gossmann et al., “Predictive Process Simulation of Cryogenic Implants for
Leading Edge Transistor Design,” in 19th International Conference on Ion Implantation
Technology (IIT), Valladolid, Spain, pp. 225–228, June 2012.
[3] R. D. Goldberg, J. S. Williams, and R. G. Elliman, “Amorphization of silicon by
elevated temperature ion irradiation,” Nuclear Instruments and Methods in Physics
Research B, vol. 106, no. 1–4, pp. 242–247, 1995.
[4] M. Posselt, L. Bischoff, and J. Teichert, “Influence of dose rate and temperature on ion-
beam-induced defect evolution in Si investigated by channeling implantation at different
doses,” Applied Physics Letters, vol. 79, no. 10, pp. 1444–1446, 2001.
[5] G. L. Olson and J. A. Roth, “Kinetics of Solid Phase Crystallization in Amorphous
Silicon,” Materials Science Reports, vol. 3, pp. 1–78, 1988.
[6] B. C. Johnson and J. C. McCallum, “Dopant-enhanced solid-phase epitaxy in buried
amorphous silicon layers,” Physical Review B, vol. 76, no. 4, p. 045216, 2007.
[7] D. D’Angelo et al., “Role of the strain in the epitaxial regrowth rate of heavily doped
amorphous Si films,” Applied Physics Letters, vol. 93, no. 23, p. 231901, 2008.

Advanced Calibration for Process Simulation User Guide 249

P-2019.03
7: Guidelines for Additional Calibration
References

[8] I. Martin-Bragado et al., “From point defects to dislocation loops: A comprehensive

TCAD model for self-interstitial defects in silicon,” in 37th European Solid-State
Device Research Conference (ESSDERC), Munich, Germany, pp. 334–337, September
2007.
[9] N. E. B. Cowern et al., “Energetics of Self-Interstitial Clusters in Si,” Physical Review
Letters, vol. 82, no. 22, pp. 4460–4463, 1999.
[10] P. A. Stolk et al., “Physical mechanisms of transient enhanced dopant diffusion in ion-
implanted silicon,” Journal of Applied Physics, vol. 81, no. 9, pp. 6031–6050, 1997.
[11] H. Saleh et al., “Energy dependence of transient enhanced diffusion and defect kinetics,”
Applied Physics Letters, vol. 77, no. 1, pp. 112–114, 2000.
[12] J. Li and K. S. Jones, “{311} defects in silicon: The source of the loops,” Applied
Physics Letters, vol. 73, no. 25, pp. 3748–3750, 1998.
[13] F. Cristiano et al., “Ion beam induced defects in crystalline silicon,” Nuclear
Instruments and Methods in Physics Research B, vol. 216, pp. 46–56, February 2004.
[14] Y. Lamrani et al., “Direct evidence of the recombination of silicon interstitial atoms at
the silicon surface,” Nuclear Instruments and Methods in Physics Research B, vol. 216,
pp. 281–285, February 2004.
[15] P. Calvo et al., “Thermal evolution of {113} defects in silicon: transformation against
dissolution,” Nuclear Instruments and Methods in Physics Research B, vol. 216,
pp. 173–177, February 2004.
[16] V. C. Venezia et al., “Depth profiling of vacancy clusters in MeV-implanted Si using
Au labeling,” Applied Physics Letters, vol. 73, no. 20, pp. 2980–2982, 1998.
[17] R. Duffy et al., “Dopant diffusion in amorphous silicon,” in MRS Symposium
Proceedings, Silicon Front-End Junction Formation—Physics and Technology,
vol. 810, San Francisco, CA, USA, p. C10.2.1, April 2004.
[18] I. Martin-Bragado and N. Zographos, “Indirect boron diffusion in amorphous silicon
modeled by kinetic Monte Carlo,” Solid-State Electronics, vol. 55, no. 1, pp. 22–28,
2011.
[19] V. C. Venezia et al., “Boron diffusion in amorphous silicon,” Materials Science and
Engineering B, vol. 124–125, pp. 245–248, December 2005.
[20] S. Mirabella et al., “Mechanism of Boron Diffusion in Amorphous Silicon,” Physical
Review Letters, vol. 100, p. 155901, April 2008.
[21] N. Zographos and I. Martin-Bragado, “Atomistic Modeling of Carbon Co-Implants and
Rapid Thermal Anneals in Silicon,” in 15th IEEE International Conference on
Advanced Thermal Processing of Semiconductors (RTP), Catania, Italy, pp. 119–122,
October 2007.

250 Advanced Calibration for Process Simulation User Guide

P-2019.03
 7: Guidelines for Additional Calibration
References

[22] T. Noda, S. Odanaka, and H. Umimoto, “Effects of end-of-range dislocation loops on

transient enhanced diffusion of indium implanted in silicon,” Journal of Applied
Physics, vol. 88, no. 9, pp. 4980–4984, 2000.
[23] G. Impellizzeri et al., “Fluorine in preamorphized Si: Point defect engineering and
control of dopant diffusion,” Journal of Applied Physics, vol. 99, p. 103510, May 2006.
[24] A. Halimaoui et al., “Investigation of Fluorine Effect on the Boron Diffusion by Mean
of Boron Redistribution in Shallow Delta-doped Layers,” in MRS Symposium
Proceedings, Silicon Front-End Junction Formation–Physics and Technology, vol. 810,
San Francisco, CA, USA, p. C3.10, April 2004.
[25] C.-L. Liu et al., “Ab initio modeling and experimental study of C–B interactions in Si,”
Applied Physics Letters, vol. 80, no. 1, pp. 52–54, 2002.
[26] S. H. Jain et al., “Metastable boron active concentrations in Si using flash assisted solid
phase epitaxy,” Journal of Applied Physics, vol. 96, no. 12, pp. 7357–7360, 2004.
[27] M. Diebel and S. T. Dunham, “Ab Initio Calculations to Model Anomalous Fluorine
Behavior,” Physical Review Letters, vol. 93, no. 24, p. 245901, 2004.
[28] M. Diebel and S. T. Dunham, “Reply to Ab Initio Calculations to Model Anomalous
Fluorine Behavior,” Physical Review Letters, vol. 96, p. 039602, January 2006.
[29] N. Zographos et al., “Process Modeling of Chemical and Stress Effects in SiGe,” in 19th
International Conference on Ion Implantation Technology (IIT), Valladolid, Spain,
pp. 212–216, June 2012.
[30] N. Zographos and I. Martin-Bragado, “Process modeling of stress and chemical effects
in SiGe alloys using kinetic Monte Carlo,” Journal of Computational Electronics, July
2013.
[31] C. Ahn et al., “Stress effects on impurity solubility in crystalline materials: A general
model and density-functional calculations for dopants in silicon,” Physical Review B,
vol. 79, no. 7, p. 073201, 2009.

Advanced Calibration for Process Simulation User Guide 251

P-2019.03
7: Guidelines for Additional Calibration
References

252 Advanced Calibration for Process Simulation User Guide

P-2019.03
Part III Advanced Calibration for Mechanics
Simulations

This part of the Advanced Calibration for Process Simulation User Guide contains the
following chapters:

Chapter 8 Using Advanced Calibration File for Mechanics Simulations on page 255

Chapter 9 Contents of Advanced Calibration for Mechanics Simulations on page 259

CHAPTER 8 Using Advanced Calibration File for
Mechanics Simulations

This chapter describes the use of the Advanced Calibration file for
mechanics simulations using Sentaurus Process and Sentaurus
Interconnect.

Advanced Calibration is a selection of models and parameters that is recommended by

Synopsys to be used for accurate process simulation, including mechanics simulation. In
Sentaurus Process and Sentaurus Interconnect, this selection of models and parameters is
contained in a text file, which can be opened with any standard text editor.

By sourcing the Advanced Calibration file at the beginning of a simulation, the standard
calibration of Synopsys is selected. If needed, you can change or extend Advanced Calibration.
This can be performed by either sourcing an additional calibration file, which contains the
required parameter changes, or editing the Advanced Calibration file with a text editor.

Location of Advanced Calibration File

The Advanced Calibration file is the ultimate product of Synopsys’ calibration efforts. For each
release of Sentaurus Process and Sentaurus Interconnect, there is a new Advanced Calibration
file that includes the best and latest set of models and parameters. To ensure backward
compatibility, the previous Advanced Calibration files are still available.

The files for the Advanced Calibration of Sentaurus Process in this release are located at:
$STROOT/tcad/$STRELEASE/lib/sprocess/TclLib/AdvCal

The default file is named AdvCal_2019.03.fps. It represents the first version of Advanced
Calibration Version P-2019.03. Older versions of the Advanced Calibration file can be found
in the same directory. For example, the file AdvCal_2018.06.fps contains the Advanced
Calibration file for Version O-2018.06 and is available for backward compatibility. Mechanical
parameters are located only in the Advanced Calibration file for silicon, SiGe, and germanium
(the default one). Currently, there are no settings related to mechanics in the Advanced
Calibration file for 4H-SiC process simulation.

Advanced Calibration for Process Simulation User Guide 255

P-2019.03
8: Using Advanced Calibration File for Mechanics Simulations
Using Advanced Calibration

The files for the Advanced Calibration of Sentaurus Interconnect in this release are located at:
$STROOT/tcad/$STRELEASE/lib/sinterconnect/TclLib/AdvCal

The default file is named AdvCal_2019.03.sis. It represents the first version of Advanced
Calibration Version P-2019.03. Older versions of the Advanced Calibration file can be found
in the same directory. For example, the file AdvCal_2018.06.sis contains the Advanced
Calibration file for Version O-2018.06 and is available for backward compatibility.

Using Advanced Calibration

To use Advanced Calibration for mechanics simulations in Sentaurus Process and Sentaurus
Interconnect, at the beginning of the input file, insert the line:
AdvancedCalibration 2019.03

Alternatively, this file can be sourced by using:

source $AdvCalDir/AdvCal_2019.03.fps

or:
source $AdvCalDir/AdvCal_2019.03.sis

The recommended set of mechanical models and parameters is located in section 2.15 of the
default Advanced Calibration file (in Sentaurus Process, it is the file for Si-, SiGe-, and Ge-
based technology). In earlier versions of Advanced Calibration (E-2010.12 to K-2015.06), this
set is activated manually by calling the AdvancedCalibrationMechanics procedure.

NOTE The set is applied by

default, and the procedure
AdvancedCalibrationMechanics is deprecated.

Earlier Versions of Advanced Calibration

You can source earlier versions of the Advanced Calibration file by inserting, for example, the
line:
AdvancedCalibration 2018.06

This is converted internally to:

source $AdvCalDir/AdvCal_2018.06.fps

256 Advanced Calibration for Process Simulation User Guide

P-2019.03
 8: Using Advanced Calibration File for Mechanics Simulations
Earlier Versions of Advanced Calibration

After sourcing the Advanced Calibration file, insert the line:

AdvancedCalibrationMechanics

This command calls the procedure including the parameters for mechanics simulation.

For earlier versions of the Advanced Calibration file of Sentaurus Process, see Earlier Versions
of Advanced Calibration on page 6.

Table 10 lists the earlier versions of the Advanced Calibration file that can be loaded with
Sentaurus Interconnect Version P-2019.03.

Table 10 Earlier versions of Advanced Calibration file and their corresponding commands
Advanced Calibration file Corresponding command

AdvCal_2018.06.sis AdvancedCalibration 2018.06

AdvCal_2017.09.sis AdvancedCalibration 2017.09

AdvCal_2016.12.sis AdvancedCalibration 2016.12

AdvCal_2016.03.sis AdvancedCalibration 2016.03

AdvCal_2015.06.sis AdvancedCalibration 2015.06

AdvCal_2014.09.sis AdvancedCalibration 2014.09

AdvCal_2013.12.sis AdvancedCalibration 2013.12

AdvCal_2013.03.sis AdvancedCalibration 2013.03

This possibility is available to provide backward compatibility. You can run simulations with
the latest version of Sentaurus Interconnect, but the simulations can still be based on an old
calibration. For new TCAD projects, it is recommended to load the latest version of Advanced
Calibration.

Advanced Calibration for Process Simulation User Guide 257

P-2019.03
8: Using Advanced Calibration File for Mechanics Simulations
Earlier Versions of Advanced Calibration

258 Advanced Calibration for Process Simulation User Guide

P-2019.03
CHAPTER 9 Contents of Advanced Calibration
for Mechanics Simulations

This chapter describes the models and parameters used in Advanced

Calibration for mechanics simulations.

Overview
A set of recommended mechanical parameters for the materials of the most common
electronics as well as appropriate settings for mechanics simulations are included in
section 2.15 of the Advanced Calibration file. In previous versions of Advanced Calibration
(E-2010.12 to K-2015.06), this set was placed in the AdvancedCalibrationMechanics
procedure. This procedure was not switched on by default and had to be invoked explicitly to
load the modified parameters described here. After sourcing the Advanced Calibration file of
Version P-2019.03, the set is applied by default without calling the procedure
AdvancedCalibrationMechanics.

NOTE The procedure AdvancedCalibrationMechanics is deprecated in

Advanced Calibration.

The set of mechanical parameters contains values from the literature (with corresponding
references for the values that differ from the default ones) for silicon, germanium, polysilicon,
SiO2 (Oxide), Si3N4 (Nitride), TiN (TiNitride), HfO2, GeO2 (GeOxide), and SiC
(SiliconCarbide). The set is incomplete, that is, it does not include all parameters for all the
materials and mechanical models available in Sentaurus Process and Sentaurus Interconnect,
but it serves as an aggregated source of the most recent (and, sometimes, different from the
default) parameters for mechanical simulations.

Switches for Interpolation in Mole Fraction–Dependent

Mechanical Models
The set of recommended mechanical models and parameters includes the parameter
interpolation syntax (see Sentaurus™ Process User Guide, Alloy Materials and Parameter
Interpolation on page 24) as well as the older syntax including deprecated syntax for Ge mole
fraction dependency of mechanical parameters in Ge-doped silicon.

Advanced Calibration for Process Simulation User Guide 259

P-2019.03
9: Contents of Advanced Calibration for Mechanics Simulations
Overview

Using the new syntax is the best choice for devices with SiGe regions. However, the usage of
the parameter interpolation can noticeably increase the simulation time for diffusion in the
presence of Ge in silicon with the concentration > Min.Conv.Conc (or similarly, the presence
of Si in germanium) even in cases where the influence of the Ge content on diffusion is not
important. This might occur, for example, in a simulation of silicon devices with Ge
preamorphization implantation (PAI) simulated with Monte Carlo implantation. Typically,
within Advanced Calibration, these two cases are differentiated by the usage of the
SiGe_and_Stress_Effect procedure:
■ Where the Ge mole fraction is important for diffusion, use SiGe_and_Stress_Effect
with the Ge_Chem_Eff=1 argument.
■ Where the Ge mole fraction is not important for diffusion, you can either:
• Not call SiGe_and_Stress_Effect at all, which corresponds to:
SiGe_and_Stress_Effect 0 0 0 0
• Call SiGe_and_Stress_Effect 0 * * * (where * is either 0 or 1).

Depending on the value of the argument Ge_Chem_Eff, either the modern syntax (for
Ge_Chem_Eff=1) or the older syntax (for Ge_Chem_Eff=0) is used for mechanics
simulations.

The following lines defining the default behavior of Advanced Calibration are included in
section 2.14 of AdvCal_2019.03.fps:
pdbSet Si Skip.Parameter.Interpolation 1
pdbSet Ge Skip.Parameter.Interpolation 1
pdbSet Mechanics Lattice.Constant.Mismatch 0
pdbSet Mechanics Parameter.Interpolation 0

The above lines switch off parameter interpolation in SiGe (as well as interpolation of elastic
constants and the coefficients of thermal expansion in all materials) and switch off the default
lattice mismatch model for alloy materials. The same settings are applied in the procedure
SiGe_and_Stress_Effect when Ge_Chem_Eff=0. If Ge_Chem_Eff=1, the opposite
values are set for the above four parameters.

When the model of parameter interpolation for mechanical parameters in compound materials
is disabled using the following command, the deprecated syntax for interpolation of
mechanical parameters in SiGe becomes active:
pdbSet Mechanics Parameter.Interpolation 0

In addition, the following command activates the old Conc.Strain model for Ge impurity in
silicon:
pdbSet Mechanics Lattice.Constant.Mismatch 0

260 Advanced Calibration for Process Simulation User Guide

P-2019.03
 9: Contents of Advanced Calibration for Mechanics Simulations
Parameters for Mechanics

Besides the usage of the modern or older syntax (and corresponding models), these switches
determine which definition of the Ge mole fraction is used. That is, the models corresponding
to the modern syntax use a more precise definition (see Eq. 2, p. 41). Whereas, the old
mechanical models use the old definition (see Eq. 3, p. 41).

The value of the Si Skip.Parameter.Interpolation parameter controls which

definition of the mole fraction is used for the xMoleFraction field saved into a TDR file after
diffusion for material SiliconGermanium derived from material Silicon.

The following parameter setting leads to the new mole fraction definition being used for saving
the xMoleFraction field:
pdbSet Si Skip.Parameter.Interpolation 0

Both the modern and older mechanics syntax allow similar functionality for SiGe simulations
(and even very close results for low Ge mole fractions), but the modern syntax is superior in
terms of functionality and precision for large mole fractions.

NOTE Versions P-2019.03 to K-2015.06 of the Advanced Calibration file for

Sentaurus Interconnect do not include settings for mole fraction
dependency of mechanical parameters and concentration-dependent
stress models, which are present in the version for Sentaurus Process.

Parameters for Mechanics

The parameters for mechanics simulations are described here.

Suppression of Dilatational Viscosity

In Sentaurus Process and Sentaurus Interconnect, the viscoelastic response can be applied to
both the deviatoric and volumetric parts. In fact, solid materials do not demonstrate dilatational
viscosity. To restrict the viscoelastic response to the deviatoric part only, the following flag is
set:
pdbSet Mechanics NoBulkRelax 1

Mole Fraction–Dependent Mechanics Parameters for SiGe

A simple linear interpolation between parameter values in silicon and germanium is used to
calculate the elastic stiffness constants and the linear thermal expansion coefficient in SiGe.

Advanced Calibration for Process Simulation User Guide 261

P-2019.03
9: Contents of Advanced Calibration for Mechanics Simulations
Parameters for Mechanics

NOTE Due to some limitations in Sentaurus Process, you should use only
temperature-independent coefficients of thermal expansion.

By default and in the case of Ge_Chem_Eff=0, the mole fraction–dependent model is invoked
by the following older syntax:
pdbSetBoolean Si IsCompound 1
pdbSetString Mechanics BCompoundList {Silicon Germanium}
pdbSetDoubleArray SiliconGermanium CompoundInterp {0 0 1 1}

For epitaxially grown materials, the correct thermal stress model is set by the following string:
pdbSetBoolean Mechanics Compound.ThExpCoeff 1

The definition of lattice mismatch strain with respect to the relaxed local lattice constant is set
by the following line:
pdbSet Mechanics Lattice.Mismatch.Strain LCC

Quadratic interpolation for the strain calculation using the lattice mismatch model in binary
Si1–xGex materials (lattice constant a SiGe = 5.431 + 0.201x + 0.026x2 Å) and Si1–xCx materials
(lattice constant a SiC = 5.431 – 2.4542x + 0.59x2 Å) is introduced [1]. In the parameter
Conc.Strain, the strain is specified as a piecewise linear function of the mole fraction (the
older mole fraction definition is used then; see Eq. 3, p. 41). The modern syntax for the
quadratic lattice constant interpolation for the lattice mismatch model in SiGe is given by:
pdbSet SiGe LatticeConstant.X2 -0.026

Cubic Crystal Anisotropy for Silicon and Germanium

For calculating stress, silicon and germanium are treated as anisotropic materials, defined by
the following lines:
pdbSet Si Mechanics Anisotropic 1
pdbSet Ge Mechanics Anisotropic 1

Temperature Dependency of Stiffness Coefficients for

Silicon and Germanium
Temperature dependency of the stiffness coefficients for silicon and germanium have been
introduced using the temperature derivatives at 300 K [2]. Both the modern syntax (using T1)
and the older syntax (using Rate) are provided. The temperature dependency of the elastic
coefficients of crystalline silicon and germanium leads to an approximately 10% reduction of
their values at 1000°C in comparison to the room temperature ones.

262 Advanced Calibration for Process Simulation User Guide

P-2019.03
 9: Contents of Advanced Calibration for Mechanics Simulations
Parameters for Mechanics

Isotropic Elastic Moduli for Germanium

Germanium isotropic elastic moduli are calculated in the Voigt–Reuss–Hill [3] polycrystalline
approximation from its single-crystal anisotropic elastic constants:
pdbSet Ge Mechanics BulkModulus 75.00e10
pdbSet Ge Mechanics ShearModulus 54.44e10

Polysilicon
Polysilicon bulk modulus and shear modulus are calculated from the measured Young’s
modulus and the Poisson ratio [4]. Polysilicon is known to be a plastic material at high
temperatures with a temperature-dependent plasticity threshold [5][6]. It is taken into account
by the Arrhenius-type dependency of the yield stress as follows:
pdbSet PolySi Mechanics FirstYield 5.75e8
pdbSet PolySi Mechanics FirstYieldW -0.23

The temperature-dependent yield stress can be used to increase accuracy of the stress
memorization effect simulation [7]. The plasticity model is not switched on in Advanced
Calibration. If needed, it can be switched on by the line:
pdbSetBoolean PolySilicon Mechanics IsPlastic 1

Temperature dependency has been introduced for the elastic moduli of polysilicon following
the values for crystalline silicon [2]. The given rate values have been derived from the
temperature dependency of the isotropic elastic moduli of polysilicon, which were calculated
from the elastic constants of silicon in the Voigt–Reuss–Hill polycrystalline approximation.
Both the modern syntax (using T1) and the older syntax (using Rate) are provided. The
temperature dependency of the elastic moduli of polysilicon leads to an approximately 10%
reduction of their values at 1000°C in comparison to the room temperature ones.

Viscosity of Oxide and Nitride

Temperature dependency also was included for the oxide and nitride stress-dependent viscosity
activation volumes [8]:
pdbSet Ox Mechanics Vcrit0 1.25e-20
pdbSet Ox Mechanics VcritW 0.35
pdbSet Nit Mechanics Vcrit0 4.7e-21
pdbSet Nit Mechanics VcritW 0.35

Advanced Calibration for Process Simulation User Guide 263

P-2019.03
9: Contents of Advanced Calibration for Mechanics Simulations
Parameters for Mechanics

Isotropic Moduli for Titanium and Titanium Silicide

For Ti (Titanium) and TiSi2 (TiSilicide), isotropic moduli based on the literature [9][10]
are suggested:
pdbSet Ti Mechanics BulkModulus 106.e10
pdbSet Ti Mechanics ShearModulus 42.e10
pdbSet TiSi2 Mechanics BulkModulus 128.3e10
pdbSet TiSi2 Mechanics ShearModulus 59.2e10

The values for TiSi2 are obtained from the measured value of the biaxial elastic modulus
(220 GPa) [10] and the assumed Poisson ratio 0.30. No reliable measurement of the Poisson
ratio for the stable (C54) phase of TiSi2 was found in the literature. The value 0.22 ± 0.07 from
[11] appears to be unreliable because it uses an improbably large value of the elastic modulus
(259 GPa) obtained for bulk material from [12]. The spread of the elastic coefficient values is,
in general, large for silicides: The values can vary by approximately 50% depending on the
exact silicide stoichiometry and process conditions [13].

Amorphous Germanium Oxide

Isotropic elastic moduli and the linear coefficient of thermal expansion (LCTE) are introduced
for vitreous (amorphous) GeO2:
pdbSet GeOx Mechanics BulkModulus 23.87e10
pdbSet GeOx Mechanics ShearModulus 18.10e10
pdbSet GeOx Mechanics ThExpCoeff 9.06e-6

The values of the elastic moduli at 25°C are taken from [14]. These values correspond to the
Young’s modulus of 43.34 GPa, which is smaller than 53.5 GPa ± 10% reported in [15].
LCTE = 9.06e-6 K–1 at 660°C is taken from [16]. In [15], LCTE = 1.0e-6 K–1 was measured
at room temperature and, in addition, LCTE = 7.5e-6 K–1 and 7.7e-6 K–1 were referenced.

Titanium Nitride
Isotropic elastic moduli and the LCTE are introduced for TiN (TiNitride):
pdbSet TiNitride Mechanics BulkModulus 300e10
pdbSet TiNitride Mechanics ShearModulus 180e10
pdbSet TiNitride Mechanics ThExpCoeff 9.35e-6

The structure of TiN films is typically polycrystalline with cubic lattice and preferable [100]
or [111] orientation [17], but for certain deposition conditions, it can be also amorphous [18].
The above TiN elastic moduli are obtained by averaging the measured values for coating films

264 Advanced Calibration for Process Simulation User Guide

P-2019.03
 9: Contents of Advanced Calibration for Mechanics Simulations
Parameters for Mechanics

[19] and Voigt–Reuss–Hill isotropic ones calculated from single-crystal measurements [20].
They result in a Young’s modulus equal to 450 GPa. Elastic moduli of TiN vary substantially
depending on the deposition conditions and the film thickness, for example, experimental
Young’s modulus values vary from 200 GPa to 488 GPa [19]–[26]. LCTE of TiN is taken
from [22].

Hafnium Oxide
Isotropic elastic moduli and the LCTE are introduced for HfO2:
pdbSet HfO2 Mechanics BulkModulus 234.615e10
pdbSet HfO2 Mechanics ShearModulus 109.2e10
pdbSet HfO2 Mechanics ThExpCoeff 8.5e-6

The above elastic moduli are obtained for monoclinic HfO2 in [27], and they correspond to a
Young’s modulus of 283.6 GPa. Depending on the manufacturing conditions, the structure of
HfO2 film can be amorphous or polycrystalline with a monoclinic or an orthorhombic lattice,
or even a combination of the three [28][29]. Reported experimental Young’s modulus values of
HfO2 vary from 152 GPa to 370 GPa [27][30][31][32][33]. The above isotropic LCTE of HfO2
was calculated by averaging the anisotropic LCTE components of monoclinic HfO2 from [34].

Silicon Carbide
Isotropic elastic moduli and the LCTE are introduced for SiC (SiliconCarbide):
pdbSet SiC Mechanics BulkModulus 220.552e10
pdbSet SiC Mechanics ShearModulus 191.106e10
pdbSetDouble SiC Mechanics ThExpCoeff 3.25622e-6

The above SiC isotropic elastic moduli are calculated in the Voigt–Reuss–Hill approximation
from experimental anisotropic elastic constants reported in [35]. The measurements of single-
crystal samples showed that 4H and 6H-SiC have the same elastic constants within
experimental uncertainties [35]. For 3C-SiC, isotropic elastic moduli based on a Young’s
modulus of 410 GPa and a Poisson ratio of 0.19 measured for single-crystal thin-film samples
in [36] are BulkModulus=220e10 and ShearModulus=172e10 dyn/cm2. Measurements
reported in [36] had substantial spread, depending on the structure of the samples. Additional
references on experimental measurements of 3C-SiC elastic properties can be found in [37].
The above isotropic LCTE of SiC was calculated by averaging the anisotropic LCTE
components of hexagonal 4H-SiC at 300 K from [38][39]. Both 6H and 3C polytypes have
similar values of LCTE [40][41][42][43]. The results of LCTE measurements for 3C, 4H, and
6H-SiC were extended analytically for other SiC polytypes in [44].

Advanced Calibration for Process Simulation User Guide 265

P-2019.03
9: Contents of Advanced Calibration for Mechanics Simulations
References

References
[1] D. De Salvador et al., “Lattice parameter of Si1–x–yGexCy alloys,” Physical Review B,
vol. 61, no. 19, pp. 13005–13013, 2000.
[2] S. Adachi, Handbook on Physical Properties of Semiconductors, vol. 1, (p. 57, Table
2.3.5 and p. 142, Table 3.3.9), New York: Springer, 2004.
[3] R. Hill, “The Elastic Behaviour of a Crystalline Aggregate,” The Proceedings of the
Physical Society, vol. 65, no. 389A, pp. 349–354, 1952.
[4] W. N. Sharpe, Jr. et al., “Measurements of Young’s Modulus, Poisson’s Ratio, and
Tensile Strength of Polysilicon,” in Tenth Annual International Workshop on Micro
Electro Mechanical Systems (MEMS), Nagoya, Japan, pp. 424–429, January 1997.
[5] C.-S. Oh, G. Coles, and W. N. Sharpe, Jr., “High Temperature Behavior of Polysilicon,”
in MRS Symposium Proceedings, Nano- and Microelectromechanical Systems (NEMS
and MEMS) and Molecular Machines, vol. 741, Boston, MA, USA, pp. J3.6.1–J3.6.6,
December 2002.
[6] Yu. V. Milman, I. V. Gridneva, and A. A. Golubenko, “Construction of Stress-Strain
Curves for Brittle Materials by Indentation in a Wide Temperature Range,” Science of
Sintering, vol. 39, no. 1, pp. 67–75, 2007.
[7] C. Ortolland et al., “Stress Memorization Technique—Fundamental Understanding and
Low-Cost Integration for Advanced CMOS Technology Using a Nonselective Process,”
IEEE Transactions on Electron Devices, vol. 56, no. 8, pp. 1690–1697, 2009.
[8] V. Senez et al., “Two-Dimensional Simulation of Local Oxidation of Silicon: Calibrated
Viscoelastic Flow Analysis,” IEEE Transactions on Electron Devices, vol. 43, no. 5,
pp. 720–731, 1996.
[9] Materials Properties Handbook: Titanium Alloys, R. Boyer, G. Welsch, and E. W.
Collings (eds.), Materials Park, Ohio: ASM International, 1994.
[10] J. T. Pan and I. Blech, “In situ stress measurement of refractory metal silicides during
sintering,” Journal of Applied Physics, vol. 55, no. 8, pp. 2874–2880, 1984.
[11] P. J. J. Wessels et al., “Stresses in sputtered Ti-Si multilayers and polycrystalline silicide
films,” Journal of Applied Physics, vol. 63, no. 10, pp. 4979–4982, 1988.
[12] G. V. Samsonov and I. M. Vinitskii, Handbook of Refractory Compounds, p. 287, New
York: IFI/Plenum, 1980.
[13] J. F. Jongste et al., “Elastic constants and thermal expansion coefficient of metastable
C49 TiSi2,” Journal of Applied Physics, vol. 73, no. 6, pp. 2816–2820, 1993.
[14] N. Soga, “Pressure Derivatives of the Elastic Constants of Vitreous Germania at 25° ,
–78.5°, and –195.8°C,” Journal of Applied Physics, vol. 40, no. 8, pp. 3382–3385, 1969.

266 Advanced Calibration for Process Simulation User Guide

P-2019.03
 9: Contents of Advanced Calibration for Mechanics Simulations
References

[15] S. Spinner and G. W. Cleek, “Temperature Dependence of Young’s Modulus of

Vitreous Germania and Silica,” Journal of Applied Physics, vol. 31, no. 8,
pp. 1407–1410, 1960.
[16] D. B. Dingwell, R. Knoche, and S. L. Webb, “A Volume Temperature Relationship for
Liquid GeO2 and some Geophysically Relevant Derived Parameters for Network
Liquids,” Physics and Chemistry of Minerals, vol. 19, no. 7, pp. 445–453, 1993.
[17] P. Patsalas, C. Gravalidis, and S. Logothetidis, “Surface kinetics and subplantation
phenomena affecting the texture, morphology, stress, and growth evolution of titanium
nitride films,” Journal of Applied Physics, vol. 96, no. 11, pp. 6234–6246, 2004.
[18] M. Leskelä and M. Ritala, “Atomic layer deposition (ALD): from precursors to thin film
structures,” Thin Solid Films, vol. 409, no. 1, pp. 138–146, 2002.
[19] A. S. Maxwell, S. Owen-Jones, and N. M. Jennett, “Measurement of Young’s modulus
and Poisson’s ratio of thin coatings using impact excitation and depth-sensing
indentation,” Review of Scientific Instruments, vol. 75, no. 4, pp. 970–975, 2004.
[20] J. O. Kim et al., “Elastic constants of single-crystal transition-metal nitride films
measured by line-focus acoustic microscopy,” Journal of Applied Physics, vol. 72,
no. 5, pp. 1805–1811, 1992.
[21] P. Jedrzejowski, J. E. Klemberg-Sapieha, L. Martinu, “Relationship between the
mechanical properties and the microstructure of nanocomposite TiN/SiN1.3 coatings
prepared by low temperature plasma enhanced chemical vapor deposition,” Thin Solid
Films, vol. 426, no. 1–2, pp. 150–159, 2003.
[22] H. O. Pierson, Handbook of Refractory Carbides and Nitrides: Properties,
Characteristics, Processing and Applications, Westwood, New Jersey: Noyes
Publications, p. 193, 1996.
[23] R. Berriche et al., “Evaluation of Hard TiN Coatings by Depth Sensing Indentation and
Scratch Testing Methods,” Advanced Performance Materials, vol. 4, no. 4, pp. 357–370,
1997.
[24] T. Namazu et al., “Direct Measurement Technique of Strain in XRD Tensile Test for
Evaluating Poisson’s Ratio of Micro-Thick TiN Films,” in IEEE International
Conference on Micro Electro Mechanical Systems (MEMS), Maastricht, The
Netherlands, pp. 157–160, January 2004.
[25] K. Holmberg et al., “Residual stresses in TiN, DLC and MoS2 coated surfaces with
regard to their tribological fracture behaviour,” Wear, vol. 267, no. 12, pp. 2142–2156,
2009.
[26] H. Chen et al., “Strength and elastic moduli of TiN from radial x-ray diffraction under
nonhydrostatic compression up to 45 GPa,” Journal of Applied Physics, vol. 107, no. 11,
p.113503, 2010.

Advanced Calibration for Process Simulation User Guide 267

P-2019.03
9: Contents of Advanced Calibration for Mechanics Simulations
References

[27] S. L. Dole, O. Hunter, Jr., and C. J. Wooge, “Elastic Properties of Monoclinic Hafnium
Oxide at Room Temperature,” Journal of the American Ceramic Society, vol. 60,
no. 11–12, pp. 488–490, 1977.
[28] M. C. Cheynet et al., “Crystal structure and band gap determination of HfO2 thin films,”
Journal of Applied Physics, vol. 101, no. 5, p. 054101, 2007.
[29] M. Franta et al., “Microstructure of HfO2 and HfxSi1-xOy Dielectric Films Prepared on
Si for Advanced CMOS Application,” in Sixth International Conference on Advanced
Semiconductor Devices and Microsystems (ASDAM), Smolenice Castle, Slovakia,
pp. 47–50, October 2006.
[30] D. K. Venkatachalam et al., “Nanomechanical properties of sputter-deposited HfO2 and
HfxSi1-xO2 thin films,” Journal of Applied Physics, vol. 110, no. 4, p. 043527, 2011.
[31] K. Tapily et al., “Nanoindentation Investigation of HfO2 and Al2O3 Films Grown by
Atomic Layer Deposition,” Journal of the Electrochemical Society, vol. 155, no.7,
pp. H545–H551, 2008.
[32] K. Tapily et al., “Nanomechanical study of amorphous and polycrystalline ALD HfO2
thin films,” International Journal of Surface Science and Engineering, vol. 5, no 2-3,
pp. 193–204, 2011.
[33] R. G. Munro, “Elastic Moduli Data for Polycrystalline Ceramics,” NISTIR 6853,
National Institute of Standards and Technology, Gaithersburg, Maryland,
https://srdata.nist.gov/CeramicDataPortal/elasticity, 2002.
[34] R. N. Patil and E. C. Subbarao, “Axial Thermal Expansion of ZrO2 and HfO2 in the
Range Room Temperature to 1400°C ,” Journal of Applied Crystallography, vol. 2,
no. 6, pp. 281–288, 1969.
[35] K. Kamitani et al., “The elastic constants of silicon carbide: A Brillouin-scattering study
of 4H and 6H SiC single crystals,” Journal of Applied Physics vol. 82, no. 6,
pp. 3152–3154, 1997.
[36] K. M. Jackson et al., “Mechanical Properties of 3C Thin-Film Silicon Carbide,” in SEM
Annual Conference on Experimental Mechanics, Milwaukee, WI, USA, June 2002.
[37] K. M. Jackson et al., “Mechanical Properties of Epitaxial 3C Silicon Carbide Thin
Films,” Journal of Microelectromechanical Systems, vol. 14, no. 4, pp. 664–672, 2005.
[38] Z. Li and R. C. Bradt, “Thermal expansion of the hexagonal (4H) polytype of SiC,”
Journal of Applied Physics, vol. 60, no. 2, pp. 612–614, 1986.
[39] NIST Structural Ceramics Database (SCD), SCD Citation Number: Z00390 (go to
https://srdata.nist.gov/CeramicDataPortal/Scd/Z00390).
[40] Z. Li and R. C. Bradt, “Thermal Expansion of the Hexagonal (6H) Polytype of Silicon
Carbide,” Journal of the American Ceramic Society, vol. 69, no. 12, pp. 863–866, 1986.
[41] NIST Structural Ceramics Database (SCD), SCD Citation Number: Z00077 (go to
https://srdata.nist.gov/CeramicDataPortal/Scd/Z00077).

268 Advanced Calibration for Process Simulation User Guide

P-2019.03
 9: Contents of Advanced Calibration for Mechanics Simulations
References

[42] Z. Li and R. C. Bradt, “Thermal expansion of the cubic (3C) polytype of SiC,” Journal
of Materials Science, vol. 21, no. 12, pp. 4366–4368, 1986.
[43] NIST Structural Ceramics Database (SCD), SCD Citation Number: Z00011 (go to
https://srdata.nist.gov/CeramicDataPortal/Scd/Z00011).
[44] Z. Li and R. C. Bradt, “Thermal Expansion and Thermal Expansion Anisotropy of SiC
Polytypes,” Journal of the American Ceramic Society, vol. 70, no. 7, pp. 445–448, 1987.

Advanced Calibration for Process Simulation User Guide 269

P-2019.03
9: Contents of Advanced Calibration for Mechanics Simulations
References

270 Advanced Calibration for Process Simulation User Guide

P-2019.03