"Projecting the course of future development in lithography is perilous, however, because we are approaching serious limitations in both device design and optical lithography." Four years ago Phillip IBlais gave this warning in his guest editorial for the special issue of Optical Engineering dedicated to Submicron Lithography. That special issue had no papers on optical tooling but reported on electron-beam, x-ray, and ion-beam lithographies, with obvious implications for the future of optical lithography. Since then, the numerical apertures of lenses have in creased, the exposing wavelengths have shortened, and one-half micrometer optical lithography has been demonstrated. With the aid of new resists, excimer lasers, and exotic deep-UV optics, production of integrated circuits with even smaller critical dimensions is being seriously considered. Optical lithography is entering an exciting new era, driven by optical engineers pushing the evolution of this technology into realms previously assumed to be beyond its limits. With operating bandwidths greater than 1 x 109 pixels/s and rapidly rising, optical lithography tooling is continuing to be the workhorse for patterning wafers, even as the critical dimensions continue to decrease. The authors of this special issue were invited to show us how this is happening.
Annular field projection optical systems have been used in production for the manufacture of circuits with feature sizes about 75% of the size of those produced by stepper lenses with the same diffraction limits. One factor contributing to this is the uniformity of imagery over the field inherent in these systems, both in and out of focus. The limits of optical microlithography will be determined by the smallest feature sizes that can be recorded with adequate focal range. For partially coherent illumination, the intensity/defocus (l-D) diagram of an optical system for a particular pattern can be used to give both the depth of focus for a given recording window and the recordability of a pattern as measured by the slope of the intensity distribution of the aerial image at the recording point. The inherent focal range advantage of annular field optical systems, when combined with their other advantages, makes them prime candidates for extending the range of optical lithography to feature sizes below 0.5 µm.
The general properties of telescopic optical systems consisting of a set of concentric surfaces are discussed. It is shown that if a monocentric system existed for which an incident collimated beam emerged as exactly collimated, then the system would give geometrically perfect imaging of any point in space. It is further shown that a real image of a real object is possible only for monocentric systems containing an odd number of reflecting surfaces. Such perfect collimation is not attainable, but it can be approached, yielding very highly corrected unit-magnification systems. Two such systems, one with a single reflection and one with three, have come into general use for microlithography. Their structure and properties are discussed.
To investigate the way to achieve 0.5 µm photolithography, experiments have been performed with a high-numerical-aperture lens, with multilayer and contrast-enhancement-layer resist processes, and with an excimer laser deep-UV stepper. The 0.6 N.A. lens is for the g-line and has a 5 mm x 5 mm field size. Single-layer resist exposures show good profiles at 0.6 µm line/space (L/S), with no effects from highly oblique illumination, and a depth of focus of 1.25 µm. Multilayer resists using spin-on-glass and contrast-enhancement layers improve the resolution to 0.375 µm with a large N.A. lens and to 0.5 µm with an i-line lens of N.A. = 0.35. Because the large numerical aperture alone cannot reach 0.5 µm and because a large field, large N.A. lens is difficult to manufacture, an i-line lens with moderate N.A. and large field appears to be a better approach. As a more advanced experiment, a KrF excimer laser stepper with an achromatic quartz/fluorite lens of N.A. = 0.37 shows no effect of speckle and has produced a 0.35 µm L/S in PMMA for a k of 0.5. The resolution with MP2400 photoresist is only 0.4 µm because of deep-UV absorption. This shows the need for more work on practical deep-UV photoresists.
A deep-ultraviolet step-and-repeat system, operating at 248.4 nm, has been developed by retrofitting a commercial lithography tool with a KrF excimer laser and custom designed, fused silica condenser and projection optics. It is a reduction system with a field size of 14.5 mm and a variable numerical aperture of 0.20 to 0.38. Resolution of 0.5 µm features over the full field is obtained with routine use, and 0.35 µm resolution is attainable under more limited conditions. This paper describes the development and testing of the system, including recent modifications, and discusses some of the technical issues associated with short wavelength excimer laser based lithography.
Practical resolution limits of submicrometer patterning by optical lithography are estimated by computer simulation. As a result, the optimum numerical aperture NA that gives the highest resolution is determined. Assuming that the permitted defocus value is ± 1 µm, a litho-graphic resolution of about 0.7 µm is obtained with 0.5 NA and near-UV exposure. A 0.5 µm resolution is obtained with 0.35 NA and deep-UV exposure. It is suggested that resolution of less than 0.4 µm will be realized by the use of resist system technologies, optical exposure tools that optimize NA as well as exposure wavelength, and precise focusing control of less than ± 0.5 µm. Furthermore, it is predicted that the goal of 0.4 to 0.5 µm resolution will be achieved industrially before the end of the 1980s by the use of optical lithography.
Key to the current and future success of optical microlithography is the excellence of mask making, more specifically the excellence of electron-beam mask making. Probe-forming e-beam mask makers have become the industry standard and provide the speed, flexibility, and accuracy required for today's pattern generation. Projected requirements for 0.5-µm lithography, however, present challenges in the form of pixel resolution, linewidth tolerance, design grid increments, overlay accuracy, throughput, and data management that require new approaches. Innovative new electron optics techniques such as the variable shaped beam and variable axis immersion lens, together with novel stage concepts and data compaction, that can meet those challenges are discussed.
This paper reviews the impact of advances in photoresists and processing on submicrometer imaging using optical lithography. Among the topics discussed for the extension of single-layer resists into the submicrometer regime are the use of dyes, the development of materials for the deep-UV region, image reversal, and contrast-enhancement layers. Approaches to dealing with substrate topographical effects, such as multilayer resists and antireflection coatings, are reviewed. Future directions to extend optical lithography further and address current problems are discussed.
Unconventional use is made of a statistical model for the photographic process to explain (in a limited way) the mechanism of infrared presensitization photography (IRPP). The model is used to explore the influence of thermal expansion of silver halide grains and their quantum sensitivity. The latter is shown to be the critical parameter.
A method has been developed to measure the free-water content and grain size of snow by use of near-infrared methods. It requires no direct physical sampling and uses reflectances at three optimally selected wavelengths. The method is based on two basic phenomena: first, snow reflection as a multiple-scattering phenomenon is strongly dependent on the grain size and the spectral variations in the ice absorption coefficient, and second, owing to differences between the absorption coefficients of ice and water, the reflection spectrum of dry snow differs from that of wet snow. Hemispherically integrated spectral reflectance has been measured in the range between 600 and 2000 nm for several types of snow including metamorphic states from newly fallen to melted-refrozen and densities from 0.10 to 0.45 g/cm3. The effect on reflectance of up to 20 vol % free water has been studied. The reflectance of snow has also been analyzed theoretically using an approximate solution of the radiative transfer equation. The experiments and theoretical calculations show that the free-water content and average grain diameter can be derived from reflectance measurements made at 1030, 1260, and 1370 nm. From measurements made so far it is estimated that an accuracy of Ã‚Â± 1.5 vol % can be achieved in the measurement of the free-water content and of ± 0.2 mm can be achieved in the measurement of the average grain size.
The potential advantages of applying coherent optical fiber techniques to broadband networks are reviewed, and a number of system types are described to illustrate the operating principles. Particular consideration is given to networks that rely on the use of optically tunable lasers either to define a source frequency or to select a channel from a frequency-multiplexed transmission using the heterodyne receiver technique.
A typical geographical map consists of a large number of lines and symbols representing various physical entities and their spatial relationships. The design of a computer-based system to automatically extract information from a map and answer queries is a challenging task that requires an intelligent query processor and a sophisticated image analysis system. One of the tasks of the image analysis system is recognition of symbols and lines. In this paper we describe simple line tracking and symbol identification routines. The routines are capable of tracking various types of lines (continuous, broken, intersecting) and symbols having simple geometrical shapes.
Schemes for obtaining lateral, radial, and combined lateral and radial shear are investigated. The shear interferometers utilize the properties of holographic lenses. It is shown that the magnitude of lateral shear can be adjusted during experimentation. Both theory and experimental results are presented.
Detail in the spot diagrams of Fig. 12 was lost in reproduction of the published paper. The figure is reproduced here in its entirety with improved detail in the spot diagrams. The following discussion further explains the figure.
I was deeply saddened to learn that George O. Reynolds, a long-time member of SPIE and active participant in its activities, had died suddenly on February 17, 1987, of a pulmonary thrombosis at the age of 49. On behalf of the officers, governors, members, and staff of the Society, I wish to express our heartfelt sympathy to his mother Marjorie Reynolds, his wife Anne, his sons William, Gordon, and Andrew, and the rest of his family and friends. This editorial is dedicated to his memory.
Progress in the fields of integrated optics and fiber optics is continuing at a rapid pace. Recognizing this trend, the goal of the author is to provide an introductory textbook on time-harmonic electromagnetic theory, with an emphasis on optical rather than microwave technologies. The book is appropriate for an upper-level undergraduate or graduate course. Each chapter includes examples of problems. The book focuses on several areas of prime importance to intergrated optics. These include dielectric waveguide analysis, couple-mode thoery, Bragg scattering, and prism coupling There is very little coverage of active components such as electro-optic modulators and switches. The author assumes the reader has a working knowledge of vector calculus and is familiar with Maxwell's equations.