Follows 1. This document can be downloaded from: http://www.cs.sandia.gov/CRF/mpsalsa.html 2. This work was partially funded by Department of Energy, Mathematical, Information, and Computational Sciences Division, and was carried out at Sandia National Laboratories, operated for the US Department of Energy under contract no. DE-ACO4-94AL85000. 3. Parallel Computational Sciences Department (org. 9221). 4. Parallel Computing Sciences Department (org. 9226). 5. Chemical Processing Science Department (org. 1126). Acknowledgments We would like to thank Professor Michael Jensen for preparing a number of the fluid mechanics examples and for urging the development of the output routines, and Aaron Thomas for benchmarking an early version of the code. We would also like to thank Ed Boucheron for identifying many instances of undesirable functionality so that we could remove them from the code. Finally, we would like to thank Rod Schmidt for his careful reading of this document. Ab...

We su"#LE) two approaches to integrated muegrated processsimusEL-L (IMPS),particu;qL# relevant to integrated circuE (IC) fabrication, in which models forequqV--' (m) andfeatu` (lm) scales are solvedsimuEVLVqLE)Lq The first approachuro reguch grids, and is applied tolow-pressu` chemical vapor deposition (LPCVD) of silicon dioxide from tetraethoxysilane (T OS). The second approachupr uroachEq;`# meshes, and is applied to electrochemical deposition ( CD) of copper. The goal is to develop approaches to estimate "loading" in these processes; i.e., the e#ects of pattern density and topography on local deposition rates. This is accomplished by resolving pattern (mesoscopic, mm) scales, which are betweenequeenEq (0.1--1 m) andfeatu# scales 0:1--1 lm). In this work, wefocu on steady-statesimu-stat resu-s We close with a fewthouxx# on extending IMPS to the grain scale, and the conversion of discrete atomistic representations tocontinuE representations of islandsduand deposition. # 2002 Elsevier Science B.V. All rights reserved.

The reactor network synthesis problem involves determining the type, size, and interconnections of the reactor units, optimal concentration and temperature profiles, and the heat load requirements of the process. A general framework is presented for the synthesis of optimal chemical reactor networks via an optimization approach. The possible design alternatives are represented via a process superstructure which includes continuous stirred tank reactors and cross flow reactors along with mixers and splitters that connect the units. The superstructure is mathematically modeled using differential and algebraic constraints and the resulting problem is formulated as an optimal control problem. The solution methodology for addressing the optimal control formulation involves the application of a control parameterization approach where the selected control variables are discretized in terms of time invariant parameters. The dynamic system is decoupled from the optimization and solved as a func...

The MPSalsa reacting flow code has been developed at Sandia National Laboratories to compute coupled three-dimensional fluid flow and detailed reaction chemistry for modeling reacting flow systems, such as Chemical Vapor Deposition (CVD) reactors. MPSalsa has been developed to take advantage of the tremendous computational speed, memory, and scalability of distributed memory parallel computers. An unstructured finite element discretization allows for the modeling of complex geometries. Complex physics for general mixtures of ideal gases, including thermal diffusion, mixture averaged and true multicomponent diffusion, variable physical properties and both gas and surface reactions can be modeled. In this brief manuscript we discuss the reacting flow model and numerical method and summarize representative calculations using MPSalsa for the CVD growth of Gallium Arsenide in a horizontal reactor with a tilted rotating substrate.

The dominant computational cost in modeling turbulent combustion phenomena numerically with high fidelity chemical mechanisms is the time required to solve the ordinary differential equations associated with chemical kinetics. One approach to reducing that computational cost is to develop an inexpensive surrogate model that accurately represents evolution of chemical kinetics. One such approach, PRISM, develops a polynomial representation of the chemistry evolution in a local region of chemical composition space. This representation is then stored for later use. As the computation proceeds, the chemistry evolution for other points within the same region are computed by evaluating these polynomials instead of calling an ordinary differential equation solver. If initial data for advancing the chemistry is encountered that is not in any region for which a polynomial is defined, the methodology dynamically samples that region and constructs a new representation for that region. The utility of ...

Classical stability and convergence theory are based on linear and local linearized analysis as the time steps and grid spacings approach zero. This theory offers no guarantee for nonlinear stability and convergence to the correct solution of the nonlinear governing equations. The use of numerical dissipation has been the key mechanism in combating numerical instabilities. Aside from acting as a post-processor step, most filters serve as some form of numerical dissipation. Without loss of generality, “numerical-dissipation/filter ” is, hereafter, referred to as “numerical dissipation”. Proper control of the numerical dissipation to accurately resolve all relevant multiscales of complex flow problems while still maintaining nonlinear stability and efficiency for long-time numerical integrations poses a great challenge to the design of numerical methods. The required type and amount of numerical dissipation are not only physical problem dependent, but also vary from one flow region to another. This is particularly true for unsteady high-speed shock/shear/boundary-layer/turbulence/acoustics interactions and/or combustion problems since the dynamics of the nonlinear effect of these flows are not well-understood, while longtime integrations of these flows have already stretched the limit of the current available supercomputers and the existing numerical methods. It is of paramount importance to have proper control on the type and amount of numerical dissipation in regions where it is needed but nowhere else. Inappropriate type

In numerical simulation ofcombustion models,solution of the chemical kinical isoften the mostexpen;J; part of thecalculation sinc accurate description ofkinbO; mechanOp inhanO large nrgep of speciesan reactionT leadin to a large set of coupled ODEs,often too complex to beconTOpRkT in theirenirpJz alon with a detailed flowsimulation Henu the nep forrepresenRkT the complex chemical reaction by simple reduced models, whichcan retain connTz;;pR accuracy whilerenepkkT computationk feasibility. Realistically,unal different conentic an at different poine in time, different reaction becomeimportan; which hasbeen exploited to developan adaptive scheme such that the reducedreaction model adapts itself to thechanb#; reactorconorpkT;zz methodology is developed in this paper to con;zpRk reduced mechanJHz bysolvin an optimization problem, where the objective is todetermin theran# of conOk;bpR alon thereaction trajectory over which a prespecified nresp ofreaction can predict the actual profilewithin an allowable toleranep Suchan adaptive reducedmechanpJ isthen coupled with the reactive flow algorithm, which selectsan appropriate mechanri depennr on reactorconorpJJ an inorpJJ thecorresponkbJ ODEs for the specified valid randp These ideas are demonstrated using the mechanzp of CO=H 2 combustion in air.

Detailed simulation of reactive flow systems using complex kinetic mechanisms consisting of hundreds of species is a computationally demanding task. In practice, to alleviate the computational complexity, the reactive flow models use a skeletal kinetic model instead of the detailed mechanism. However, the reduced chemistry models can accurately predict the detailed model only over a limited range of conditions. Since the reactive flow simulation encounters a broad range of conditions, using a single reduced model throughout the simulation incorporates inaccuracy into the predictive capacity of the flow model. In this paper, an adaptive scheme is presented, which aims at developing different reduced models to address the changing conditions of the flow simulation, thereby maintaining high accuracy throughout the simulation with relatively simple chemistry models. A new approach is developed for the analysis of the range of validity of the reduced model, which is highly nonconvex, for which the conventional techniques do not perform well. Finally, a procedure to implement the adaptive chemistry in different flow simulations is presented.

I hereby declare that all information in this document has been obtained and presented in accordance with academic rules and ethical conduct. I also declare that, as required by these rules and conduct, I have fully cited and referenced all material and results that are not original to this work.

Previous work done at Sandia on reaction mechanisms for the chemical vapor deposition (CVD) of silicon oxide from tetraethoxysilane (TEOS) and ozone is documented and tested in computational models at Watkins-Johnson. Recommendations for future work in this area are discussed. 4 Report of Work done for Technical Assistance Agreement 1269 between Sandia National Laboratories and the Watkins-Johnson Company: "Chemical Reaction Mechanisms for Computational Models of SiO 2 CVD" I. Introduction The use of computational modeling to improve equipment and process designs for chemical vapor deposition (CVD) reactors is becoming increasingly common. Commercial codes are available that facilitate the modeling of chemically-reacting flows, but chemical reaction mechanisms must be separately developed for each system of interest. One of the products of the Watkins-Johnson Company (WJ) is a reactor marketed to semiconductor manufacturers for the atmospheric-pressure chemical vapor deposition (AP...