AVS1997 Session MS-TuP: MS Poster Session
Tuesday, October 21, 1997 5:30 PM in Room Exhibit Hall 1
Tuesday Afternoon
Time Period TuP Sessions | Topic MS Sessions | Time Periods | Topics | AVS1997 Schedule
|
MS-TuP-1 Simulation of Negative Ion Trapping Regions in an Inductively Coupled HBr/BCl3/Cl2 Plasma
J. Shon, P. Vitello (Lawrence Livermore National Laboratory) Using a detailed plasma mechanism for HBr/BCl3/Cl2 gas mixtures developed in a well mixed model,[1] generation, transport, recombination and trapping of negative ions can be realistically modeled. We are using a two dimensional transport code[2] which solves a set of time dependent fluid equations for electrons and ions self-consistently with Poisson's equation for the electric potential. Parametric studies have been performed with respect to pressure (2-100 mTorr), input power (50-2000kW), gas mixtures(HBr/BCl3/Cl2), and chamber size. Preliminary results show trapping regions of Cl- and Br- in a spatially resolved sheath. [1] E. Meeks and J. W. Shon, IEEE Trans. on Plasma Sci., 23, 539, 1995. [2] P. Vitello, J. N. Bardsley, G. Dipeso, and G. J. Parker, IEEE Trans. on Plasma Sci., 24, 123, 1996. |
|
MS-TuP-2 Nano-Metrology for the Data Storage Industry
D.A. Chernoff (Advanced Surface Microscopy, Inc.) The optical disc industry is in the midst of a revolutionary change in manufacturing requirements. The transition from Compact Disc (CD) to DVD (Digital Video Disc) is equivalent to a 2 or 3 generation change in the integrated circuit industry. DVDs use nanometer technology: the smallest features are about 400 nm long, 320 nm wide, with a track pitch of 740 nm. These dimensions are a factor of 2-2.5 smaller than for CDs. Furthermore, the dimensional tolerances have been tightened by an even larger factor. One of the specific manufacturing challenges is controlling track pitch, a problem analogous to overlay registration for IC's. We have developed a high accuracy process for making the needed measurements. The process uses a general purpose Scanning Probe Microscope with the following added features: - the use of uniform and accurate 0.3 mm calibration gratings - the calculation of feature positions with sub-pixel precision - replacement of the nonlinear raw length scale with a corrected, linear length scale. We also present applications in magnetic recording and IC fabrication. See also www.a1.com/asm |
|
MS-TuP-3 Role of Spectroscopic Ellipsometry in Thin Film Metrology
J.A. Woollam (University of Nebraska, Lincoln); J.N. Hilfiker (J. A. Woollam Co., Inc.); D. Thompson, T.E. Tiwald (University of Nebraska, Lincoln); C. Herzinger (J. A. Woollam Co., Inc.) There is growing industrial need for accurate thin film standards for optical instrument calibration. NIST and private companies sell "traceable thickness standards", and "round-robins" are periodically conducted. We propose that several factors must be simultaneously involved in establishing standards for accurate film thickness measurement: 1) Instrumentation: light beam collimation, angle of incident accuracy, polarizer quality, stability of light source and electronics, detector or source polarization sensitivity, method and accuracy of azimuth angle calibration; 2) Data Analysis: positional accuracy of sample mounts, material structure model assumed, regression parameters chosen, regression algorithm used, "goodness of fit" choice, extent of parameter correlation, confidence limits on fits (and how determined); and 3) Samples: optical constants must be known or accurately measured, accounting for interfaces, roughness, and oxides. Samples must be uniform and cleaning procedures consistent. Atmospheric condensation, or surface oxidation can make sample properties time-dependent. Certain materials, structures and compositions can be better choices than others. In summary, samples alone do not make a standard. Rather, the sample, instrumentation factors, and analysis procedures must simultaneously be accurately defined. |
|
MS-TuP-4 A Throttle Cycle Cryocooler with Mixed Gas Refrigerant for a Water Trap
A. Khatri, M. Boiarski (APD Cryogenics, Inc.); S.B. Nesterov (Moscow Power Engineering Institute, Russia) Performance of a turbomolecular vacuum pump can be enhanced greatly during a pumpdown process when combined with a water trap. The challenge has always been in designing a water trap so that it will have low profile and will have minimum effect on the system configuration. It must also provide the optimal ratio between the conductance of the non-condensing component such as nitrogen and the trapping coefficient of the condensing component such as water vapor in the temperature range of 115 K to 130 K. A closed cycle mixed gas refrigerant cryocooler has been developed to cool the condensing array of a water trap. This water trap (AquaTrapTM) operates in the temperature range of 115 K to 130 K, has no moving parts in the cold end, requires no controller to maintain the temperature, provides optimum ratio between the nitrogen conductance and the trapping coefficient for the water vapor, uses room temperature gas supply lines, needs no maintenance, has high reliability and can be separated from the compressor unit by the gas lines. It has been tested in the range of 100 mm, 160 mm, 200 mm and 250 mm ID sizes. The throttle cycle cryocooler uses an oil lubricated compressor to provide temperature in the narrow range of 115 K to 130 K depending on the heat load condition on the water condensing surface connected to a cold finger. The AquaTrapTM design optimization was conducted using the test particle Monte-Carlo method. Calculation results show good agreement with the experimental data. It is shown that the refrigeration capacity depends on the composition of the mixed gas refrigerant and can be customized to operate over the desired temperature range for a given system. Extensive ability in predicting characteristics of mixed gas refrigerants and the experience in designing reliable systems will make it possible to use this cryocooler in many other applications. |
|
MS-TuP-5 Semiconductor Manufacturing Education and Training: Graphic User Interface Issues
M.O. Ovessi (University of Maryland, College Park); D.E. Eckard (North Carolina State University); A.R. Rose, G.W. Rubloff (University of Maryland, College Park) We have previously exploited physically-based dynamic simulation of equipment, process, sensor, and control system behavior for engineering analysis, design, and optimization. In addition, we are packaging such simulator structures to accomplish effective manufacturing education and training for semiconductor processes. In order to make the training environment learner-friendly, graphic user interface (GUI) design and module architecture must address both ergonomic and cognitive issues. First, dynamically changing variables must readily reflect moderate changes (5-10%) as well as dynamic range (e.g. pressure from 10E-9 torr to 1 atm); we have built new, quite different elements which accomplish this. Second, effective feedback of error and warning messages must be provided which the user is operating the virtual equipment; we have constructed and demonstrated an architecture which permits coding of conditions in the simulator itself, while providing readily edited error messages for the instructor. In addition, we are working to understand the tradeoffs and benefits between 2-D and 3-D graphical representation, to develop software tools for capture and playback of prior event sequences, and to design education, experiments which measure the cognitive value of these approaches to effective learning modules. This work is supported in part by the National Science Foundation under grant NO. EEc-9526147, and benefits from cooperation with CVC products Inc., and Visual Solutions, Inc. |
|
MS-TuP-6 Correlation of Al/Cu.5 in-film Particle Performance with Non-metallic Contaminants in Al/Cu.5 Targets
S. Annavarapu, P. Gilman (Materials Research Corporation) In-film particles in Al/Cu.5 metallization layers have been identified as an important source of process yield loss in devices incorporating sub-0.5µm Al metal lines. The defects are reportedly spherical or tear-drop shaped ranging from 0.5µm to 2.0µm and are assumed to derive from the impaction of molten metal micro-droplets on substrates during sputtering. It is generally believed that such "droplets" are a by-product of micro-arcing during sputtering at non-conducting particles in the Al/Cu.5 targets. In this study, the qualitative "droplet" performance of 10 different Al/Cu.5 targets was correlated with the level of non-metallic contaminants and inclusions contained in the targets. The sputtered targets were analyzed by glow discharge mass spectroscopy, LECO gas analysis, optical and electron microscopy, scanning acoustic microscopy, auger electron spectroscopy and ultrasonic C-scan. The analyses showed no correlation with metallic impurities, microstructure or machined surface roughness of the target. Examination of the target sputtered surface by electron microscopy, auger analysis and scanning acoustic microscopy revealed the prevalence of 5µm to 50µm carbon and oxide particles in the worst-performing targets, and limited evidence of arc-damage at exposed inclusions. Ultrasonic C-scan analysis of the targets, confirmed by microscopy, revealed 5µm to 80µm inclusions dispersed through the target bulk with cumulative counts ranging from 10 to 1500 per target. Though TEM analysis only detected sub-2µm oxide particles, the number density of oxide inclusions was consistent with the ultrasonic data. The "droplet" performance of Al/Cu.5 targets showed a good correlation with the frequency of oxide or carbon inclusions larger than about 25µm and the interstitial impurity (O, C, H) content measured by the LECO method; O > 20ppm, C > 10ppm, and bulk H > 0.1ppm appear to be threshold levels for arcing and particle generation. |
|
MS-TuP-7 The Determination of Volatile Organic Compounds in Vacuum System Components by Short Path Thermal Desorption
E.D. Butrym (Rutgers University); C.W. Baker, S.M. Colby (Scientific Instrument Services, Inc.) Techniques for the detection and identification of volatile organic compounds are presented. Specific attention is given to materials commonly found in vacuum systems. Examples are chosen to illustrate methods for determining sources of contamination. These include: synthetic polymers (O-Rings), lubricants (vacuum grease), and oils (pump oil). Temperature dependencies are shown. A Short Path Thermal Desorption System (Scientific Instrument Services, Inc.) is used to introduce volatile compounds in to an HP 5890 Gas Chromatograph and an HP 5971 MSD mass spectrometer. The thermal desorption system includes an glass lined sample cartridge with a needle tip attached to one end. This needle is inserted directly into the GC Injection Port. A long transfer line and cold spots are thus avoided. Analytes are cryo-trapped at the head of the GC column to narrow the temporal width of the injection slug. The sample cartridge is heated at the start of the analysis process. Solid samples are placed directly into the sample cartridge for analysis. For liquid samples, an absorbent such as Tenax(TM) is placed |
|
MS-TuP-8 CFD Modeling for Control and Optimization.
J.L. Ebert, G.W. van der Linden, R.L. Kosut, A. Emami-Naeini (SC Solutions, Inc.) Computer models of single wafer processes can be used for the design and evaluation of the chamber and related components such as control laws. To allow for quick design iterations it is highly desirable to have a computationally efficient, yet accurate, description of the system dynamics. Unfortunately, straightforward application of finite element methods quickly leads to large and computationally intensive models, especially when the dynamics of solids as well as gases are simultaneously taken into account. This paper will develop an approximate modeling approach, by considering the separation of the dynamics of the solids and the gases. The idea is to use the results from finite element CFD models in the low order finite volume models of the solids. This separation is motivated by the fact that the thermal dynamics of the solids are significantly slower and not strongly coupled to the fluid dynamics, because of the difference in heat capacity. As a result, the fluid dynamics can be approximated as a static map at the boundaries, which can be used in the thermal dynamics model of the solids. It will be shown that the proposed approach leads to an efficient, yet reasonably accurate, model of the total system dynamics. This model is espcecally useful for preliminary equipment and control design. |
|
MS-TuP-9 Numerical Simulation of Microvacuum Magnetosensor
V.A. Fedirko (Moscow State University of Technology "Stankin", Russia); N.G. Belova (FTIRAN, Russia) We report on the numerical modelling of a magnetosensitive element based on a microvacuum field emission array. A cylindrical multielectrode microcell with a circular emitter and a split plate anode is suggested as a magnetosensitive element. Circular grid electrodes to control the electron beam and a plate target electrode serving as secondary electron emission source are included in the model. Electron beam transport in the self-consistent electric field and in an external magnetic field have been simulated using modified cylindrical particle mesh code with the original Monte Carlo procedure for the emission events treatment. We have discovered that the split-anode difference current increases markedly with the intensity of a magnetic field parallel to the cell axis. It thus can serve as a magnetoresponse signal resulting in a high enough sensitivity. Static characteristics and dynamic response of the element have been studied for different electrodes position. The sensitivity enhancement by secondary emission from the target electrode have been investigated. The results enable to optimize the cell structure for the maximum sensitivity. |