[15] T. R. Kramer, Process Planning for a Milling Machine from a Feature-Based Design, in Proceedings of Manufacturing International '88, Atlanta, Georgia, April 1988, ASME Vol. III, American Society of Mechanical Engineers, New York (1988) pp. 179-189.

[16] T. R. Kramer and W. T. Strayer, Error Prevention and Detection in Data Preparation for a Numerically Controlled Milling Machine, in Proceedings of 1987 ASME Annual Meeting, (ASME PED Vol. 25), American Society of Mechanical Engineers, New York (1987) pp. 195-213.

[17] Alkan M. Donmez, Robert J. Gavin, Lew Greenspan, Kang B. Lee, Vincent J. Lee, James P. Peris, Eric J. Reisenauer, Charles O. Shoemaker, and Charles W. Yang, The Turning Workstation in the AMRF, NBSIR 88-3749, National Bureau of Standards, Gaithersburg, MD (1988).

[18] Howard T. Moncarz, Architecture and Principles of the Inspection Workstation, NBSIR 88-3802, National Bureau of Standards, Gaithersburg, MD (1988).

[19] H. G. McCain, R. D. Kilmer, and K. N. Murphy, Development of a Cleaning and Deburring Workstation for the AMRF, in Deburring and Surface Conditioning '85: Conference Proceedings, September 23-26, 1985, Chicago, Illinois, Society of Manufacturing Engineers, Dearborn, Michigan (1985). [20] F. M. Proctor, K. N. Murphy, and R. J. Norcross, Automating Robot Programming in the Cleaning and Deburring Workstation of the AMRF, in Deburring and Surface Conditioning '89, San Diego, CA, February 1989, Society of Manufacturing Engineers, Dearborn, Michigan (1989).

[21] S. Szabo, M. Juberts, and R. D. Kilmer, Automated Guided Vehicles at the National Bureau of Standards Automated Manufacturing Research Facility, Journal of Unmanned Vehicle

Systems, pp. 38-47, Summer 1984.

[22] D. Libes and E. Barkmeyer, The Integrated Manufacturing Data Administration System (IMDAS)—An Overview, Int. J. Comput. Integr. Manuf. 1, 44-49 (1988).

[23] Siegfried Rybczynski, Edward J. Barkmeyer, Evan K. Wallace, Michael L. Strawbridge, Don E. Libes and Carol V. Young, AMRF Network Communications, NBSIR 88-3816, National Bureau of Standards, Gaithersburg, MD (1988).

[24] Charles McLean, Mary Mitchell, and Edward Barkmeyer, A Computer Architecture for Small-Batch Manufacturing, IEEE Spectrum 20 (5), 59-64 (1983).

[25] Howard M. Bloom, Cita M. Furlani, and Anthony J. Barbera, Emulation as a Design Tool in the Development of Real-Time Control Systems, in 1984 Winter Simulation Conference Proceedings, Dallas, Texas, November 28-30, 1984, Institute of Electrical and Electronics Engineers, New York (1984) pp. 627636.

[26] W. P. Meade, The National Bureau of Standards' Automated Manufacturing Research Facility (AMRF)—A Manufacturing Technologies Resource Center, Management Collaborative Group, Chapel Hill, NC (1988).

[27] K. B. Lee, Computer-Controlled Fastener Manufacturing Workstation, Technical Paper Series 941719, Society of Automotive Engineers, Warrendale, PA (1994).

[28] J. S. Albus, The NIST Real-time Control System (RCS): An Application Survey, in Proceedings of the AAAI 1995 Spring Symposium Series, Stanford University, Stanford, CA, March 27-29, 1995.

[29] An Assessment of the National Institute of Standards and Technology Programs, Fiscal Year 1990, National Academy Press, Washington, DC (1991).

In 1987, the semiconductor industry was undergoing a technological transition into the submicrometer range of device dimensions. Small dimensions that are very important to device performance or yield are called critical dimensions (CD). Optical metrology technology was adequate to measure the critical dimensions above 1 μm, but as these dimensions shrunk into the submicrometer regime, the industry felt that the development of a new technology would be necessary. Scanning electron microscopy began to be employed as the new "tool" to measure submicrometer structures. In an effort to assist the industry in this transition, two papers were published in the same issue of the NBS Journal of Research summarizing the knowledge on optical metrology at that time. These papers discussed the capabilities for extension into the submicrometer regime and reported on the promising scanning electron microscopy and its potential to take over from optical metrology. The two pioneering review papers, Submicrometer Linewidth Metrology in the Optical Microscope [1] by Diana Nyyssonen and Robert Larrabee, and Submicrometer Microelectronics Dimensional Metrology: Scanning Electron Microscopy [2] by Michael T. Postek and David C. Joy, helped to reorient the metrology direction of the semiconductor industry, with impacts being felt even today.

By the year 1987 optical microscopes had been used for looking at small things for several centuries and had been optimized for this purpose. However, they were not optimized for accurate dimensional metrology in the submicrometer regime. Scanning electron microscopes had also been used for looking at small things, but only for decades instead of centuries. They also were not optimized for submicrometer dimensional metrology. Accurate measurements of submicrometer dimensions in both kinds of microscopes were more difficult to make and interpret than was generally recognized at that time.

These two back-to-back papers [1,2] served to clarify a number of misconceptions by those in industry who were actually manufacturing the microscopes or using them to make critical submicrometer dimensional measurements. Both papers were aimed directly at submicrometer measurements for quality control purposes in the semiconductor and magnetic-storage tape-head industries. An unusually large number of requests for reprints were received from readers in the United States, and the oral feedback revealed that the

papers were extensively faxed between colleagues in foreign countries. It is impossible to document the savings to industry due to the resulting improvements in quality control attributable to these papers because quality control information is often considered proprietary. Anecdotal feedback at subsequent technical meetings and during the authors' visits to industry clearly indicated that material in the papers was important and that the savings were substantial.

As the dimensions of interest continued to shrink in the years following publication, these papers helped set the stage: 1) for the improvements in the basic instrumentation used for optical and SEM metrology, 2) for the motivation to develop theoretical models for interpretation of such measurements, and 3) for the more intelligent use of the resulting measurement data. The information in these papers is still relevant to submicrometer metrology even though much progress has occurred since their publication. They should still provide useful background information on micrometer and submicrometer measurements for new metrologists and for new, or more demanding, applications in the new millennium (e.g., for linewidth and overlay measurements in the semiconductor industry and for critical dimensions in tape-head, microfabrication, micromachining industries).

The submicrometer optical metrology paper [1] assessed the capabilities and limitations of optical submicrometer dimensional metrology and how well it would be able to meet the measurement needs of future semiconductor processing technologies (e.g., linewidth measurements). The fact that the wavelength of the commonly used visible light in the optical measuring tools was comparable to the feature sizes of interest led to serious limitations. The paper discussed the need to model mathematically the effects of diffraction in the image and thereby develop a meaningful criterion for deciding which point on the image corresponds to the edge of the feature whose dimensions were being measured. Nyyssonen and Kirk developed such a model [4] and Nyyssonen used that model for the calibration of NIST's first photomask linewidth standards [5]. The modeling (and the measurement) is much more difficult for opaque specimens (e.g., silicon wafers) and becomes increasingly difficult as the feature heights. become larger than about a quarter wavelength and as the aspect ratio (feature-height/width) approaches unity. These factors, plus the non-vertical edge shapes of the

features, severely compromised the accuracy of dimensional measurements in the submicrometer regime.

The general problem of optical linewidth metrology was discussed with emphasis on: 1) definition of linewidth for non-ideal features, 2) precision and accuracy (now referred to as Type 1 and Type 2 errors), 3) effects of measurement errors on process control, 4) instrument design, 5) resolution of the measuring microscope, 6) optical-based linewidth standards, and 7) alternative linewidth measurement techniques. The factors affecting measurements of small feature dimensions were discussed and illustrated by calculated image waveforms for a typical patterned polysilicon line on a silicon dioxide layer upon a silicon substrate. In these calculations the waveform changed as the silicon dioxide layer thickness was varied and the edge geometry of the line deviated from vertical. In addition, the different kinds of microscopes used in optical metrology were discussed and illustrative image profiles under various illumination conditions were presented.

Perhaps the main message of this paper was that submicrometer optical metrology was more difficult than commonly envisioned at the time and that many factors came into play that were often overlooked, ignored, or inadequately treated in practical applications. With the ongoing impetus of the semiconductor industry toward ever-smaller submicrometer dimensions at that time, this attitude had to change if the anticipated future needs for decreased measurement uncertainty and increased accuracy were to be met. This paper helped set the stage for the change that did, in fact,

occur.

The scanning electron microscope used in low accelerating voltage mode was initially felt to be the panacea for the problems encountered by optical submicrometer metrology. The paper by Postek and Joy [2] demonstrated that, although the SEM was capable of precise measurements, accuracy was another issue altogether. It also pointed out a number of pitfalls associated with the instrument, making use of a simple micrograph of a dime (Fig. 2). This micrograph drove home the point that just because an image came from an SEM did not mean that it was an accurate representation. As important as it was to understanding the instrumental problems, this paper also pointed out that the main limitation of the SEM for accurate submicrometer metrology is the electron beam/sample interaction, which affects the generation and collection of the measured signals. This was the first paper to stress the need for understanding the electron beam/sample interaction as a requirement for accurate metrology with the SEM.

Following the publication of this paper, a heightened awareness of the issues associated with SEM metrology prompted significant improvements in the instrumenta

Fig. 1. Comparison of calculated image profiles of the edge of an opaque vertical-wall small-height (e.g., photomask) line. The ordinate is relative transmitted light intensity and the abscissa is the distance from the edge of the line in micrometers. The step-function rise to full transmission shown by the straight lines in the figure represents what one would ideally expect. The bright-field image (solid curve) and the confocal image (dashed curve) show that the edge is not located at the point of 50% of full transmission.

Fig. 2. Scanning electron micrograph of a dime. This image and the discussion in the original paper clearly demonstrated to the reader that one cannot assume that just because the micrograph was taken with an SEM and that the magnification and linescale are displayed that they are accurate. The proper magnification should be 4.6x. This simple demonstration showed that the read-out of the commercial SEM should not necessarily be trusted at face value and thus led many users to scrutinize their SEM measurements more carefully, setting the stage for many new improvements in SEM metrology.

tion as described in later publications [7,10]. Today, fully automated CD-SEM instruments are routinely be ing utilized in semiconductor production applications throughout the world.

Diana N. Nyyssonen joined NBS as a physicist in 1969 and quickly developed the first photomask linewidth standards. Her work at NBS showed how optical image simulation modeling could be used as a tool for applying optical microscopes in submicrometer metrology and, by so doing, exceed the classical resolution limits of imaging microscopes. She won the Department of Commerce Silver Metal for her work in this area. In 1985, she left NBS to form her own R&D company specializing in optical dimensional metrology. She later joined IBM Corporation and specialized in scanning probe microscopes.

Robert D. Larrabee joined NBS in 1976 as a physicist specializing in the electrical characterization of bulk silicon. In 1985 he became the Group Leader of the Microelectronics Dimensional Metrology Group, replacing Diana Nyyssonen. Under his leadership, the group continued the existing photomask linewidth projects and initiated new SEM metrology programs with Michael Postek and other members of his group [3]. He held the position of Group Leader until his retirement in 1994. In 1999 he participated in the award of the Department of Commerce Bronze metal Team Award for his post-retirement work in developing a new optical overlay metrology tool and a novel standard for use in its alignment.

David C. Joy is currently a Distinguished Scientist, Director of the EM Facility, and Professor at the University of Tennessee. He also holds a joint appointment with Oak Ridge National Laboratory, where he is a member of the Staff in the Materials and Ceramics Division. Since publication of the subject paper, he has contributed to the improvements in SEM and SEM modeling. He has recently published two books, Monte Carlo Modeling for Microscopy and Microanalysis and Semiconductor Characterization by Scanning Electron Microscopy. He has contributed to the evolution of the scanning electron microscope as a viable production tool through his research in low accelerating voltage electron microscopy, modeling, electron holography, and nano-tip development.

Michael T. Postek is currently the Leader of the Nano-scale Metrology Group at NIST. Since this paper appeared, he has worked closely with International SEMATECH and its member companies in the development of scanning electron microscopy as a tool for

semiconductor production. He has been awarded a 1998 R&D 100 award for the development of SEM Monitor (a tool used to test the performance level of automated production SEMs) and two Department of Commerce Silver Medals for his work in metrology with the scanning electron microscope. He is currently completing development of an accurate low accelerating voltage SEM magnification standard (SRM 2090), a sharpness standard RM 8091, and a production-critical SEM width standard.

Prepared by Michael T. Postek and Robert D. Larrabee.

Bibliography

[1] Diana Nyyssonen and Robert D. Larrabee, Submicrometer Linewidth Metrology in the Optical Microscope, J. Res. Natl. Bur. Stand. 92, 187-204 (1987).

[2] Michael T. Postek and David C. Joy, Submicrometer Microelectronics Dimensional Metrology: Scanning Electron Microscopy. J. Res. Natl. Bur. Stand. 92, 205-228 (1987). [3] Robert D. Larrabee and Michael T. Postek, Parameters Characterizing the Measurement of a Critical Dimension, in Handbook of Critical Dimension Metrology and Process Control, (SPIE Critical Review, Vol. CR52), Kevin M. Monahan (ed.), SPIE Optical Engineering Press, Bellingham, WA (1994) pp. 2-24. [4] Michael T. Postek, Scanning Electron Microscope-based Metrological Electron Microscope System and New Prototype Scanning Electron Microscope Magnification Standard, Scanning Micros. 3, 1087-1099 (1989).

[5] R. D. Larrabee, L. Linholm, and M. T. Postek, Microlithography Metrology: Scanning Electron Microscope Metrology, in Handbook of VLSI Microlithography: Principles, Technology and Applications, William B. Glendinning and John N. Helbert (eds.), Noyes Publications, Park Ridge, NJ (1991) pp. 148-238. [6] Michael T. Postek, Jeremiah R. Lowney, Andras E. Vladar, William J. Keery, Egon Marx, and Robert D. Larrabee, X-ray Lithography Mask Metrology: Use of Transmitted Electrons in an SEM for Linewidth Measurement, J. Res. Natl. Inst. Stand. Technol. 98, 415-445 (1993).

[7] Michael T. Postek, Critical Issues in Scanning Electron Microscope Metrology, J. Res. Natl. Inst. Stand. Technol. 99, 641-671 (1994).

[8] Michael T. Postek, Andras E. Vladar, Samuel N. Jones and William J. Keery, Interlaboratory study on the lithographically produced scanning electron microscope magnification standard prototype, J. Res. Natl. Inst. Stand. Technol. 98, 447-467 (1993). [9] J. R. Lowney, M. T. Postek, and A. E. Vladar, A Monte Carlo Model for SEM Linewidth Metrology, in Integrated Circuit Metrology, Inspection, and Process Control VII, (Proceedings SPIE 2196), SPIE, Bellingham, WA (1994) pp. 85-96. [10] M. T. Postek, Scanning Electron Microscope Metrology, in Handbook of Critical Dimension Metrology and Process Control, (SPIE Critical Review, Vol. CR52), Kevin M. Monahan (ed.), SPIE Optical Engineering Press, Bellingham, WA (1994) pp. 46-91.

This 1989 paper [1] reported a breakthrough which led to a Nobel Prize for William D. Phillips-the first Nobel to be awarded to a NBS/NIST staff scientist. The experiment described in the paper demonstrated that light from a laser could be used to cool atoms to a much lower temperature than was previously thought possible. The fact that light carries momentum and can exert a force on objects was realized by James Clerk Maxwell in his theory of electromagnetism, developed in the 19th century. At the turn of the century, experiments by Lebedev [2] and Nichols and Hull [3] for the first time measured these forces in the laboratory. This concept of radiation pressure helped explain why comet tails point away from the sun and was important in understanding the stability of certain types of stars, but it had little laboratory relevance until the advent of the laser. In 1975, two groups proposed the counter-intuitive idea that radiation pressure from a laser could be used to cool atoms [4,5]. By carefully choosing the frequency of the laser, it appeared possible to cause the atoms to emit light at a slightly higher frequency (energy) than they absorbed, carrying away the thermal energy of the atom. This frequency difference derived from the Doppler shift due to the motion of the atoms. Doppler cooling was first demonstrated [6] with trapped ions in 1978 (at NBS by the Wineland group).

William Phillips joined the Electricity Division at NBS in 1978 to work on the gyromagnetic ratio of the proton and the SI ampere experiments, with the understanding that he could devote some of his time to developing laser cooling ideas for neutral atoms and atomic beams. He was joined in his efforts by a long-term visitor to NBS, Harold Metcalf from the State University of New York at Stony Brook. Several of the key achievements in neutral-atom laser cooling were produced by what would become the Laser Cooling Group. These included the demonstration of efficient ways of decelerating atomic beams with laser light: Zeeman cooling [7], in which the changing Doppler shift of a decelerating atomic beam is compensated by a spatially-varying magnetic field, and "chirped" cooling [8], where the Doppler shift is compensated by a changing laser frequency. These two methods are still the only methods used today to decelerate atomic beams. Another significant accomplishment of the Laser Cooling Group was the first trapping of neutral atoms with magnetic fields [9] in 1985. Magnetic

trapping is now widely used in dozens of experiments studying Bose-Einstein condensation of dilute gases. By the late 1980s a few groups around the world were investigating the properties of "optical molasses," the name given to a "sticky" configuration of laser beams that could cool and hold on to atoms for as long as a few seconds. The action of the light upon the atoms created a viscous environment for the atoms, hence the "molasses" appellation. A group at Bell Laboratories, headed by Steven Chu, had measured a temperature of 240 μK [10] for a sample of sodium optical molasses, in accord with the Doppler cooling theory that had been developed a few years earlier. The NBS group had made a number of measurements of the properties of a sodium molasses, such as the lifetime of the atoms in the molasses. Each measurement they made had disturbing discrepancies between the results and the Doppler theory. At the urging of one of the members of the team, Paul Lett, then a recently arrived postdoctoral fellow, they set out to measure the temperature of the atoms. This was something that they had shied away from earlier because its measurement was rather difficult and it had already been done at Bell Labs. They crafted a new, sensitive technique to measure the velocity distribution of the cold atoms, and thus extract the temperature. The time-of-flight technique they developed looked for fluorescence from atoms that traversed a probe laser beam after being released from the molasses. The duration of the pulse of fluorescence would be inversely proportional to the atomic velocity, which would allow the extraction of the temperature. For reasons of convenience, they placed the probe above the molasses (at 240 μK the atoms would have plenty of thermal velocity to overcome gravity to reach the probe). After a number of puzzling days with no signals, they moved the probe under the molasses and immediately saw a strong pulse of fluorescence (Fig. 1). To their great surprise, the temperature that they found was 40 μK (the atoms were so cold that gravity turned them around before they could reach the probe placed above the molasses). This result was six times lower than what the theory had predicted was the ultimate limit, the so-called Doppler limit, as well as contradicting the Bell Labs results [11]. To assure that some unknown feature of their new measurement technique was not deceiving them, they measured the temperature with three other methods, all of which were