Deposition

CFD-ACE+ is the leading commercial code in the semiconductor industry for high fidelity simulation of deposition and etch of compound semiconductors for:

• ALD
• CVD
• MOCVD or MOVPE
• RTP

CFD-ACE+ accurately simulates the transport of multi-component species and the associated gas-phase and surface reactions in the context of complex geometries. The thermal environment is accurately predicted using advanced radiation models and specialized models for esoteric effects such as thin gaps, rotating parts, and thin films. CFD-ACE+ can be use effectively for both design of equipment and determination of process recipes.

Electromagnetic fields are used extensively in semiconductor processing for plasma generation, induction heating, and flow control/damping. The equations for electromagnetics in CFD-ACE+ are fully coupled with the conservation equations governing flow, energy, and plasma. This allows the analyst to accurately model effects such as coil shape and operating conditions on the process at hand.

Induction Heating: A time varying current generates a time varying magnetic field, which in turn induces eddy currents in conductive materials, generating heat in a process referred to as inductive heating. This technique is often used to heat substrates in thermo-chemical reactors. CFD-ACE+ can model inductive heating for three dimensional reactors with multiple materials and complex coils shapes driven by multiple electrical frequencies.

Two numerical approaches are available: time domain or frequency domain. The time domain approach enables the simulation of multiple frequencies. The frequency domain approach provides for the more rapid solution of a single sinusoidal frequency, as often found in semiconductor applications. Filament models allow for the embedding and simulation of complex three-dimensional coil shapes without having to resort to large grids.

Electroplating Applications

CFD-ACE+ can model electroplating in the context of a full-scale reactor model. In addition to basic heat and fluid flow, the model accounts for: Multi-species transport Multi-species reactions Multi-species deposition Transient effect of growing film on conductivity Pulse and Pulse reversing Turbulence effects Joule Heating Current at the electrodes Primary Secondary (Approximates kinetics at the surface) Tertiary (Includes ion transport and depletion)

ESI Group is releasing CFD-TOPO, a state-of-the-art 3D feature-scale code for simulation of the evolution of features during the fabrication of electronic devices. CFD-TOPO predicts the transport, chemistry, etch and deposition of semiconductor materials on the microscopic scale. It enables prediction of three-dimensional topological evolution for multiple materials during thermal or plasma enhanced fabrication of electronic devices.

Feature scale models simulate etching and deposition of micron-scale features on a wafer or semiconductor device. The CFD-ACE+ GUI provides links to the third-party feature scale codes, SPEEDIE and EVOLVE. These links allow CFD-ACE+ reactor models to pass specie flux information to the respective feature scale model for subsequent prediction of feature evolution. A generic link is also provided for companies with their own proprietary software.

Inductively Coupled Plasma (ICP)

CFD-ACE+ - Simulation of SiO2 deposition in an Inductively Coupled Plasma (ICP) reactor

Capacitively Coupled Plasma (CCP)

CFD-ACE+ 3D - Simulation of Capacitively Coupled Plasma (CCP) for Oxygen at 400 millitorr

Radiation is often the dominant means of energy transport in a semiconductor reactor and is frequently used as the primary heating mechanism.  CFD-ACE+ offers a Discrete Ordinate Method and a Monte Carlo Method for simulating radiative heat transport. These models simulate the propagation of multi-band radiation through materials with disparate optical properties and are especially appropriate to semiconductor applications where the absorptivity of the gas is typically low.

Discrete Ordinate Method (DOM): The CFD-ACE+ Discrete Ordinate Method (DOM) is accurate and straight-forward to use, requiring little interaction from the user other than specification of optical material properties such as emissivity and absorption. DOM provides excellent results for semiconductor applications, except in the event of strong angular reflections (specular effects).

Monte Carlo Method: The Monte Carlo approach offers a highly accurate method of simulating radiative transport for semiconductor processing. It is strongly recommended for applications with dominant specular reflections, nonlinear anisotropic scattering, or path dependent radiative properties. New algorithms for Monte Carlo in CFD-ACE+ have improved its efficiency to the extent that it is nearly as fast as the above Discrete Ordinate Method.

A critical component of a good chemical process model is the accurate formulation of the multi-step chemistry involved. To address this need, ESI Group offers three solutions for developing and providing chemical reaction mechanisms:

• CFD-ACE+ Reaction Libraries
• Mechanism Development
• Links to third-party mechanisms/formats

CFD-ACE+ Reaction Libraries

Libraries of reaction mechanisms for a range of precursors are provided with CFD-ACE+ as a standard feature. These libraries are fully expandable and accessible via the CFD-ACE+ Graphical User Interface (GUI). Specialized mechanisms can be incorporated upon request. These reaction libraries include:

• Nitrides
• III-Vs
• Fluorocarbons
• SiC
• SiO₂

Mechanism Development

ESI Group actively develops mechanisms in-house using ab initio calculations (Computational Chemistry). These ab initio calculations can be used to develop a complete set of mechanisms and/or supplement an existing set. Techniques for simplifying a mechanisms set (Reduced Mechanisms) are also available.

Links to Third-Party Mechanisms & Formats

CFD-ACE+ Version 2003 comes with a translator for the CHEMKIN II format developed by Sandia National Laboratories, allowing the user to read plasma and non-plasma CHEMKIN files directly into CFD-ACE+ for subsequent solution in the context of a CFD-ACE+ model.

In addition, CFD-ACE+ facilitates incorporation of mechanisms and reaction solvers via user-subroutines. These user-subroutines can include very generalized reaction rates. This feature enables third parties to build upon CFD-ACE+ capabilities to create models for specialized applications. For example, STR Inc offers a customized third-party CVD module for the deposition of selected III-V materials.

Reaction Map for MOCVD of GaN

Validation of GaN Mechanism vs. Experimental Data of Chen et. al. 1995

STR Inc Customized Interface for an AIXTRON 200 Reactor