AVS1996 Session MS-MoP: Manufacturing Science and Technology Group Poster Session

Monday, October 14, 1996 5:30 PM in Ballroom A
Monday Afternoon

Time Period MoP Sessions | Topic MS Sessions | Time Periods | Topics | AVS1996 Schedule

MS-MoP-1 Education and Training in Semiconductor Manufacturing Processes through Physically-based Dynamic Simulation
G. Lu, M. Oveissi, D. Eckard, G. Rubloff (North Carolina State University)
We are developing PC-based applications of dynamic equipment and process simulation for manufacturing education and training. Because they reflect quantitatively and visually the detailed response of the system to user-initiated actions (e.g., valve opening, reaction rate change with temperature) in real time, they respond realistically to any sequence of learner actions and thus provide a new approach for active learning through "hands-on" virtual operation of sophisticated processing equipment. For the student with relatively little technical background (e.g., manufacturing operators, vocational students), the approach permits the learner to operate equipment and understand its principles, such as vacuum and gas flow, heat transfer, and chemical reactions. The physical sophistication of the simulators also enables engineering design exercises for more sophisticated learners (e.g., graduate students or practicing engineers), such as for process control and for statistical aspects of multi-step manufacturing. The modules are constructed by developing a powerful simulation engine (using VisSim/super TM/ from Visual Solutions, Inc.) and enhancing its user interface through Visual Basic/C++ functionality. Prototypes will be demonstrated at this poster session.
MS-MoP-2 A Wafer-Transfer Robot for Use in Ultra High Vacuum
M. Kanetomo, H. Kashima, T. Suzuki (Hitachi, Ltd., Japan)
A polar co-ordinate wafer-transfer robot has been developed that enables thin-film manufacturing to be done in an ultra high vacuum. This robot features a friction-less spring structure that operates in a vacuum and is magnetically driven by a motor in the atmosphere. To strengthen rigidity in the upper and lower directions when the spring arms are retracted, each arm has ten sheet springs connected together by rigid plates. The shape of the spring arms were optimized by simulating a large deflection model and using the finite element method. The magnetic coupling consists of 12 pairs of permanent magnets arranged in the circumference direction along the vacuum bulkhead. The permanent magnets are primarily composed of rare earth the elements used to prevent degradation of the magnetic characteristics during the baking used to obtain the ultra high vacuum and to obtain a torque of 80 kgcm or more. To achieve highly reliable operation under ultra high vacuum conditions, all moving parts in the vacuum environment have a minimum of two roller bearings. A repeat position precision of 0.4 mm or less was achieved in atmosphere. At a pressure 10\super-7\ Pa, a maximum transfer distance of 800mm and a transfer speed of 400mm/s were achieved. After operation in a cleanroom at a atmosphere pressure, only three particles above 0.1 maicrometer in size were counted. By using this wafer-transfer robot, the area needed for film manufacturing in a multichamber ultra high vacuum can be reduced about 30% compared with using a linear-magnet-coupling wafer-transfer robot.
MS-MoP-3 Development and Characterization of an Advanced Selective Silicon Nitride Etch in a High Density Plasma Reactor
M. Rousey-Seidel, S. Athavale, P. Rajora, S. Lin, J. Almerico (Tegal Corporation)
Advances in semiconductor technology are pressing the limits of traditional isolation technology based on LOCOS oxidation with a silicon nitride mask. One significant limitation has been the availability of an anisotropic selective dry nitride etch with the critical dimension control required for ULSI devices. Leading edge processes require selectivity to thermal oxide > 10:1 and total oxide loss < 50\Ao\. Sub-half micron design rules require minimum micro loading for etch rate and profile as well as nitride sidewall profiles > 85 degrees. An advanced selective nitride etch has been developed and characterized to meet these requirements. Statistical design of experiments methodology is used in this development to ensure robust process design and useful trends for optimization. The optimized process space has been applied to several production device structures. Iterative process improvements have been made to ensure the process will meet a broad range of applications, including LOCOS isolation etch and silicon nitride spacer etch. Data for process characterization and optimization is presented. Low pressure and low bias etching are demonstrated to be important factors for both critical dimension control and high selectivity to gate oxide. New process capabilities are demonstrated for two key ULSI applications for silicon nitride etch.
MS-MoP-4 Characterization of a Low Temperature, Low Pressure PECVD TEOS Oxide Deposition Process
L. Arias, S. Selbrede, M. Weise, D. Carl (Mattson Technology, Inc.)
A low temperature (180 \degree\C) and pressure (\<=\ 750 mTorr) TEOS oxide deposition process was developed and characterized in a commercially available PECVD reactor. The reactor uses a dual frequency, capacitively coupled, parallel plate electrode design, which employs multi-station sequential deposition to enhance throughput and uniformity. Deposition rate, uniformity, wet etch rate, Si-OH content, and film stress were characterized as a function of process pressure, gas composition, RF power and temperature. Fourier Transform Infrared Spectroscopy (FTIR) was used to determine the Si-OH content of the deposited films. Production quality oxide films were deposited using low TEOS flowrate (45 sccm), high oxygen flowrate (4000 sccm) and low pressure (500 mTorr). Deposition rate increased linearly with TEOS flowrate and decreases with oxygen flowrate. Deposition rate was weakly dependent on high frequency power and independent of pressure in this low pressure regime. Film thickness uniformity across a 200 mm wafer improved with decreasing TEOS flowrate and pressure. Uniformity was a weak function of oxygen flowrate and high frequency power. Film stress became more compressive with decreasing TEOS flowrate and was a weak function of oxygen flowrate, high and low frequency power, and pressure. A high quality TEOS oxide was deposited in this new processing regime, suitable for integrated circuit applications.
MS-MoP-5 Experimental Investigation of Collisional and Collisionless Heatings in Low Pressure Microwave Plasma Discharges
P. Mak, T. Grotjohn, J. Asmussen, Jr. (Michigan State University)
Recently microwave high density plasma sources have proven useful in semiconductor processing applications, such as submicron etching and thin film deposition. Even with their widely practical usages, questions concerning certain aspects of their operation at low pressure remain. One such vital question is what is the mechanism of microwave heating as pressure is varied from moderate pressure to very low pressures? Is microwave energy supplied through ohmic heating, ECR heating (if magnetic fields are present) or other collisionless mechanisms as pressure is reduced? This paper will present the results of an experimental investigation designed to identify the heating mechanisms in a microwave argon gas discharge as pressure is reduced. Two sets of experiments are performed on a 12.5 cm id. discharge created in a 17.8 cm id. internally tuned microwave applicator. One set of experiment investigates a un-magnetized discharge while the other utilizes strong ECR magnetic fields.The experiments include measurements of (1) magnitude & spatial variation of the impressed electric field, and (2)downstream plasma density & electron temperature as the incident, 2.45 GHz power and pressure vary from 180 to 400 W & 1 - 50 mTorr respectively. Heating mechanisms are identified by comparing the measured impressed electric fields & plasma densities versus pressure and power to both global and numerical plasma models. Both un-magnetized & magnetized discharges are compared to a discharge sustained only with collisional heating. Initial results indicate that both discharges are sustained by collisionless heating processes at pressure below 10 mTorr.
MS-MoP-6 The Study of Plasma-Chemistry Mechanisms for Plasma Processing using Chlorine-based Gas Mixtures: HBr/Cl\sub 2\, BCl\sub 3\/Cl\sub 2\, Cl\sub 2\
J. Shon, R. Larson, E. Meeks (Sandia National Laboratories)
The plasma chemistry of the chlorine-based etching gases BCl\sub 3\/Cl\sub 2\, HBr/Cl\sub 2\ and Cl\sub 2\ will be considered to provide a global view of chlorine plasma processes. The gas-phase excitation mechanisms of ions, radicals, and their surface recombination mechanisms will be discussed. A well-stirred reactor model provides trends of spatially and temporally averaged plasma properties over a range of operating conditions: chamber pressure, power deposition, feed gas mixtures and flow rate. Critical reaction pathways with respect to a range of operating conditions will be discussed according to the sensitivity and rate of production data obtained from a well stirred reactor model. Using time-dependent species concentration information, the differences in response time of radical generation and generation of ions and excited species will also be illustrated. For a HBr/Cl\sub 2\ plasma, the nominal conditions are 1000 Watts of input power, 5 mTorr of pressure, and a gas mixture of HBr/Cl\sub 2\ = 20/80%. The major ions are Cl\super +\, Cl\sub 2\\super +\ and Br\super +\. The results show that approximately Hbr/Cl\sub 2\ = 70/30% will produce the equal amount of Br\super +\ and Cl\super +\ ions. The generation of Br\super +\ and Cl\super +\ are almost exclusively from the ionization of ground states and excited states of Br and Cl, where Br and Cl is mostly formed from the electron dissociation of HBr and Cl\sub 2\, respectively. Cl\sub 2\\ super +\ is mostly from the ionization of ground state Cl\sub 2\.
MS-MoP-7 Chemical Vapor Cleaning of Iron and Copper from Silicon Wafer Surfaces
E. Robertson III, S. Beck, M. George, D. Moniot, D. Bohling (Air Products & Chemicals, Inc.); S. Bedge (Lehigh University)
The removal of metal contamination from silicon surfaces becomes more critical as the minimum feature size for devices decreases. Chemical vapor cleaning (CVC) techniques offer the possibility of greater efficacy with lower chemical and water usage than its liquid phase counterparts. In previous work, CVC was shown to be effective for removing copper, iron, and nickel from intentionally contaminated surfaces using the chelating ligand 1,1,1,5,5,5-hexafluoro-2,4-pentanedione (H\super +\hfac). In this paper, we discuss the results of studies using H\super +\hfac and other ligands, e.g. trifluoro acetic acid (H\super +\tfaa), and trimethylsilyl-1,1,1,5,5,5-hexafluoro-2,4-pentanedione (SEE\super +\hfac), to elucidate mechanisms responsible for the removal of iron and copper from the surface of silicon wafers. Investigations of metal removal rates are conducted using two types of surfaces: 1) surfaces intentionally contaminated with sub-monolayer concentrations of iron or copper, and 2) thin films of iron or copper oxides. Surface analysis performed by XPS, TXRF, TOF-SIMS, and SIMS provide the metal atom surface concentration and the nature of the chemical species involved in the surface reactions. The results show that the removal rate depends on the ligand. For example, H\super +\hfac is more effective than H\super +\tfaa in removing iron. We have also observed that the removal rate varies with the orientation of the sample in the gas flow stream and speculate that this dependence results from a change in the boundary layer with the change in the orientation. The differences in removal rates observed for iron and copper may be due to several mechanisms: steric hindrance, decomposition of the ligand, and mass transfer limited reactions.
MS-MoP-8 Wafer Contamination by Showerhead-Accelerated Particles in an RFParallel Plate Reactor
S. Choi, D. Rader, R. Buss (Sandia National Laboratories)
The contamination of semiconductor wafers by particles accelerated in showerhead gas injection nozzles is a serious concern during plasma processing. The trajectories of particles in plasma processing reactors are determined by external forces (the most important being neutral fluid drag, thermophoresis, electrostatic, viscous ion drag, and gravitational), particle inertia, and Brownian motion. In this paper, we present a numerical study of the transport of particles entering the reactor through a showerhead, including particle interaction with the plasma. We compare to results from a research reactor which is configured to measure the particle penetration depth into the plasma, for particles injected through a showerhead. The particles accelerate through the showerhead nozzles and can obtain enough kinetic energy to penetrate deep into the plasma, even against the ion wind force pushing the particles back toward the showerhead. The particle motion is monitored by a l! aser scattering imaging system. To analyze the experimental results, we have used two university codes (HPEM and DTS from the University of Illinois) to model the plasma and particle transports. The simulation has been performed over a wide range of conditions of pressure, gas, particle size, wafer temperature, and plasma power to obtain optimum conditions for minimizing wafer contamination by particles that are introduced through the gas inlet nozzles. * This work performed at Sandia National Laboratories supported by the U.S. Department of Energy under contract DE-AC04-94AL85000.
MS-MoP-9 Investigative Techniques to Maximize Device Yields during Ion Implantation Processing
M. Edgell, S. Smith, V. Chia (Charles Evans & Associates)
Maximizing yields and reliability of IC devices is crucial to remain competitive in the semiconductor industry. This is clearly reflected by the recent trend of processing facilities placing the responsibility on ion implanter manufacturers to provide improved equipment that duplicates existing device characteristics obtained by older equipment. This shift in emphasis from equipment specification to device characteristic specification means that the analytical characterization focuses on diagnosing problems relating to device failures as opposed to ion implanter failures. The conclusion from diagnosing the device failure using the new equipment is then to improve the ion implanter performance by monitoring this particular device characteristics. In other words, only the problems associated with the ion implanter that affects the device reliability needs to be investigated. SIMS is presently the most commonly used analytical tool for implanter characterization due to it high sensitivity for most elements. The diversity of conventional SIMS to include Magnetic SIMS, Quadrupole SIMS and SurfaceSIMS shows that this technique is still maturing. However, complementary analytical techniques are required to effectively investigate the source of the device problem and subsequently to improve the ion implanter performance. This paper reviews the applications of SIMS, Total X-Ray Fluorescence (TXRF) and Time-of-Flight SIMS (TOF-SIMS) to investigate a variety of device failures that are traced to ion implanter problems.
MS-MoP-10 Hardmask Metal Etching at Low Pressures
A. Kornblit, J. Colonell, N. Ciampa, J. Lee (Bell Laboratories)
As device critical dimensions decrease below 0.5 micron and metalization schemes become more complex, tougher requirements are placed on metal patterning. One way to address these issues is to use a hard mask material, such as a silicon-dioxide layer, during the metal etch. However metal etch processes using photoresist mask to do not produce the same desired results as when a hard mask is used. In this paper we report the results of a survey of various chemistries to etch metal stacks consisting of aluminum and TiN. We used a commerical high-density etcher (Lam Alliance TCP) capable of handling large gas throughput at pressures of 1-2 mTorr. The basic chemistry consisted of Cl\sub 2\ and various additives such as BCl\sub 3\, N\sub 2\, HBr and HCl. The effects of the additives, source power, bias power, pressure, flow and temperature were investigated. The range of power investigation was 200 to 100 W for the source and 50 to 150 W for the bias. The pressure range was 1 to 10 mTorr, with total flows varying between 40 to 120 sccm. The temperature range was 25 to 50 deg. C. With the proper additive, good profiles with good mask and underlying oxide selectivities can be obtained. Because of the reduced aspect ratios when using a hardmask, aspect ratio dependent etching (ARDE) is reduced significantly enabling the patterning of lines with spacings as low as 0.25 micron. The effects of chemistry on resistance to corrosion was explored as well.
MS-MoP-11 Characterization of Ultrathin SiO\sub 2\ Films Grown by Rapid Thermal Oxidation
Y. Hu, S. Tay (AG Associates); C. Zhao, K. Hebert, E. Irene (University of North Carolina, Chapel Hill)
The time-temperature product during gate dielectric formation is very important in controlling the extent of vertical diffusion profiles in active devices as well as the extent of lateral diffusion in traditional isolation schemes. Rapid thermal oxidation (RTO) is emerging as an important method in the formation of <8 nm gate dielectrics for 0.25 \mu\m ULSI technology. The merits of RTO arise from its high temperature process (>1050\super o\C) which results in greater noise-immunity and hot carrier reliability. By rapid thermal annealing the RTO gate oxide in N\sub 2\O ambient, as much as 7 atomic percent of nitrogen can be incorporated into the gate oxide to form a diffusion barrier against boron from the P+-polysilicon gate electrode. Ultrathin RTO dielectric films grown in O\sub 2\ and N\sub 2\O have been characterized using single wavelength ellipsometry (SWE), spectroscopic ellipsometry (SE) and tunneling current oscillations. RTO oxide thicknesses, 4-6 nm as determined by quantum oscillation technique, were used in SWE analysis, from which a refractive index of 1.894 was determined for the ultrathin oxide films. A two film model was employed in the SE analysis. An interfacial layer, 0.4-0.6 nm thick, consisting of 40% a-Si and 60% SiO\sub 2\ under the bulk SiO\sub 2\ (n=1.465), accounts for the higher index value of the total films. In this paper, the characteristics of the ultrathin RTO oxide films will be compared to other low thermal budget dielectrics including ECR plasma oxides.
MS-MoP-13 Method for Sputter Yield Extraction for HDP CVD Simulators
P. Kapur, D. Bang, J. McVittie, K. Saraswat (Stanford University); T. Mountsier (Novellus)
High Density Plasma Chemical Vapor Deposition (HDP-CVD) is a method of current interest to use for intermetal and interlevel dielectric gapfill in semiconductor circuits. In HDP-CVD, the silicon dioxide film is often sputter-etched as it is being deposited. The resputtering of the deposited film significantly affects the final film profile. Consequently, it is important that the topography simulators which model HDP-CVD should take the effects of resputtering into consideration. A novel method to extract the angle dependent sputter yield for HDP-CVD processes is presented. The method consists of extracting the sputter yield distribution from a semi-circular test structure on a typical silicon wafer. This yield distribution is, then, used to simulate the final silicon dioxide profile. The validity of the distribution is determined by comparing the simulated and the experimental profiles. This method is fully compatible with semiconductor manufacturing technology and requires no equipment modifications. An example of this method of sputter yield extraction is demonstrated. The experimental data was obtained on a HDP-CVD system and SPEEDIE was used as the process simulator.
MS-MoP-14 Flow and Heat Transfer Considerations in Tool Development for 300 mm Processing
M. Meyyappan (Scientific Research Associates)
Plasma tool development for large area processing, especially 300 mm wafers and large display panels, is receiving much attention recently. The cost of design, development, and prototyping new generation equipment is significantly high compared to previous generation efforts. In this regard, computational models have become valuable in reducing the number of trial and error steps. Thus far, many of the published works have focused on the discharge physics aspects of tool development, addressing power coupling, plasma and radical generation, and related issues. In addition to these, flow and heat transfer related considerations are important in equipment design. In this work, we have systamatically invstigated these issues using 2- and 3-dimensional simulations. The model based on Navier-Stokes equations accounts for slip flow and jump in temperature at the walls in the transition flow regime. We have shown previously that this approach accurately reproduces pressure measurements in low pressure, high density plasma reactors, with and without the plasma. Here, a simple model gas phase chemistry and surface reaction are used. Based on the simulations, we develop scaling rules for flow uniformity, uniformity of surface radical concentration, and processing rates, in terms of dimensionless numbers such as Reynolds number, heat and mass Peclet numbers and reactor geometry parameters.
MS-MoP-15 Direct Simulation Monte Carlo (DSMC) Computations of an HDP CVD Reactor
L. Gochberg (Novellus Systems, Inc.); D. Rault (NASA Langley Research Center)
Requirements for the next generation of semiconductor devices continue to demand increased component density and decreased feature size. Also, each thin film layer deposited in the manufacture of a semiconductor device must possess specific and reliable properties to provide the necessary performance. The development of new chemical vapor deposition (CVD) manufacturing equipment to achieve these goals can be made more efficient by using computational modeling of the flows inside these reactors as part of the equipment development process. Computations are performed for the flow inside of a high-density plasma (HDP) CVD reactor designed for simultaneous etch and deposition of a silicon dioxide dielectric thin film. The reactor operates at a substantially reduced pressure compared to ambient (5 millitorr). Based on the mean free path at that pressure, reactor conditions range from near continuum and nearly collisionless flow. Hence, any computations of the flow field will need to address effects of rarefaction as well as be able to compute in the near continuum regime. A particle-based direct simulation Monte Carlo (DSMC) code was used to examine the performance of the HDP reactor from Novellus Systems, San Jose, California. The paper presents neutral flow field results inside the reactor, and for the gas flow inside the injection tubes. The computations are performed on multiple platforms, including HP workstations and a several different Pentium personal computers. Full scale computations for the reactor can be accomplished on a single processor with less than 1,000,000 molecules.
MS-MoP-16 A Model of the Surface Evolution of Pt during Dry Etching
D. Kotecki (IBM Microelectronics Division); S. Hamaguchi, K. Milkove, C. Farrell (IBM T.J. Watson Research Center); C. Wang (IBM Microelectronics Division)
The patterning of noble metals, such as Pt, by dry etching techniques, which are important for integrating high-dielectric and ferroelectric materials into microelectronic devices, has proven to be a challenge due to the low volatility of reactive byproducts. A numerical model, based on a shock tracking algorithm, has been used to predict the evolution of the surface profile during dry etching of Pt. In the model, a mask material, with a specified height, sidewall angle, and pattern density is situated over a layer of Pt. An ion flux is incident on this structure and based on the sputter yield, sticking coefficient, relative etch rates of the various layers, and angles of the evolving structure, material is removed by physical sputtering while layers of Pt and PtCl\sub x\ are redeposited on the surfaces of the etched structures. The profile of the structure obtained after etching through the Pt layer is compared to experimental results obtained for the etching of Pt in both an inert (Ar), and a reactive (Cl\sub 2\, CF\sub 4\) plasma. The results from the model are found to be in qualitative agreement with the observed experimental results and the time evolution profile generated by the model provides insight into the mechanisms of the etching process.
MS-MoP-17 Measurements and Modeling of Pressure Uniformity and Gas Distribution in an Asymmetrically Pumped Inductively Coupled Plasma Etching System
B. McMillin, W. Collison, S. Baldwin, Jr., M. Barnes (Lam Research Corporation)
The etching uniformity in low-pressure inductively coupled plasma (ICP) tools depends to a large extent on the uniformity of the plasma, gas distribution, and neutral transport above the wafer, all of which can be affected by the design of the source coil and the chamber. Recently, Kushner et al.[1] reported 3-D plasma modeling results which investigated the impact of the coil design on the ion flux uniformity in a planar ICP. Their results showed that asymmetries in etch rates could be directly correlated with spatial asymmetries in the ion production rates, and that proper coil design could minimize these effects and result in more uniform etch processes. In this study, we examine the gas distribution and pressure uniformity in a planar ICP with a side-pumped chamber asymmetry, by modeling the 3-D neutral transport with a commercial fluid code (with slip boundary conditions) at typical processing conditions (5-50 mTorr, 20-500 sccm). We compare these results to experimental measurements of the wafer pressure distribution, and to etch uniformity measurements for both ion-assisted and chemical etching processes. The results show that, for typical operation of a high conductance plasma chamber, the side-to-side wafer pressure gradient is less than ~2%, even for the worst case high flow/low pressure conditions. Hence, at typical polysilicon etching conditions, the minor pressure and flow nonuniformities associated with asymmetric pumping have no significant effect on the resulting etch uniformity. At some neutral depleted chemical etch conditions, however, etch nonuniformities can become apparent in the absence of well-designed gas injection schemes. __________________ [1] M. J. Kushner et al., J. Appl. Phys., submitted March, 1996.
MS-MoP-18 Linking a Detailed Chemistry Model to Topographical Simulation in Plasma CVD of Silicon Dioxide
E. Meeks, R. Larson, P. Ho (Sandia National Laboratories); J. Rey (Technology Modeling Associates); S. Huang (Lam Research Corporation)
The predictive abiliities of topography simulators are limited by the ability of the model user to accurately supply ion and radical fluxes at a gas/surface interface. In the results reported here, we have used a well mixed reactor model of a high density plasma to predict and characterize fluxes, deposition rates, and ion-milling rates on a flat surface. These calculated rates become input to the TERRAIN topographical simulator, which predicts CVD gap-fill characteristics for inter-metal-layer dielectric deposition. The gas-phase chemistry in the plasma reactor model is comprised of electron impact reactions with silane, oxygen, hydrogen, and argon, as well as neutral radical recombination, abstraction, and oxidation reactions. The surface reaction mechanism contains four classes of reactions: silicon-containing radical deposition, radical abstraction, ion-induced desorption, and physical ion sputtering. This is consistent with the proposed reaction paths and experimental observations of S. M. Han and E. Aydil at UCSB.\super 1\ We include relative thermochemistry of the surface and gas species to allow reversible reaction dynamics. The plasma model results show good agreement with measured ion densities, as well as with measured net deposition rates. The TERRAIN model converts the flat-surface deposition and sputter rate to local angular-dependent rates. This is achieved through Monte Carlo determination of the ion angular distribution that results from the ions traversing an rf sheath. Results of linked simulations are presented and compare to SEMs of oxide deposition in trenches. \super 1\S. M. Han and E. S. Aydil, to appear in Thin Solid Films, 1996.
MS-MoP-19 Comparing Continuum and Kinetic(DSMC) Reactor Scale Models for Low Pressure Plasma Systems
T. Bartel, J. Johannes (Sandia National Laboratories); D. Economou (University of Houston); M. Riley (Sandia National Laboratories)
Low pressure plasma systems (< 25 mtorr) are currently being modelled with both continuum and kinetic (Direct Simulation Monte Carlo - DSMC) techniques. Continuum codes are typically used for their computational advantage even though the kinetic codes correctly describe the low pressure nonequilibrium phenomena. This paper will focus on contrasting the assumptions of both methods and compare results with GEC Plasma Chlorine data taken at 20 mtorr. Differences between continuum and kinetic descriptions or implementations in a bulk plasma/sheath model, neutral and ion boundary conditions, ion energy equation, and plasma chemistries will be addressed. We will present results, both numerical and data, of spatially resolved distributions for Cl-, Cl2+, and Cl (there was no etch in the experimental system) at 20 mtorr and 185 watts. Also, surface angular and energy distributions for incident ions will be compared. The continuum code used was MPRES (University of Houston) and the kinetic code was Icarus (Sandia 2D DSMC code). Finally, we will present results for a C2F6 chemistry on the AMAT Omega reactor using both methods and compare with available data.
Time Period MoP Sessions | Topic MS Sessions | Time Periods | Topics | AVS1996 Schedule