JP2001507870A - Method and apparatus for integrating optical and interference lithography to generate complex patterns - Google Patents

Abstract

SUMMARY OF THE INVENTION The present invention provides a method and apparatus for defining a single structure (11) on a semiconductor wafer by spatial frequency components, wherein some of the spatial frequency components (12-16) are optical Obtained by lithography, some are obtained by interference lithography techniques. Interference lithography images high frequency components and optical lithography images low frequency components. The optics collects many spatial frequencies and the interference shifts the spatial frequencies to higher spatial frequencies. Thus, since the mask does not need to provide high spatial frequencies, the mask is configured to generate only low frequency components, which allows for the manufacture of simpler masks with larger structures. These methods and apparatus facilitate the writing of more complex repetitive and non-repetitive patterns with a single exposure at higher resolutions than currently available when using only known optical lithography. I do.

Description

DETAILED DESCRIPTION OF THE INVENTION                   Integrate optical and interference lithography                  Method and apparatus for generating complex patterns   This invention relates to NAVY's DARPA / Department dated May 29, 1996 for two years. This was done in the course of research pursuant to Grant No. N66001-96-C8617.Technical field   The present invention generally relates to the use of interference techniques to create repeating structures during semiconductor manufacturing. More specifically, an interference lithography for making arbitrarily complex patterns on a wafer. On integration with optical lithography.Background technology and technical issues   Advances in the production of integrated circuits using large scale integration (VLSI) are becoming smaller and smaller. It has been characterized by shrinking functional dimensions. Transverse direction of transistor function Dimensions range from ５5 micrometers (4K DRAM) in 1970 to today's 0.35 micrometers (64M DRAM). For functional dimensions The continuous improvement in this is achieved by reducing linear dimensions by 30% every three years. "Moore's law" to estimate the decrease in exponential function size characterized by Is an essential part of This “law” is described in the “National Technology Roadmap for Semi conductors "(Semiconductor Industry Association, 1994) As such, it is the basis of the semiconductor industry plan.   Throughout this advance, optical lithography has become the dominant resource for manufacturing applications. It remained a sography technique. To enable this dramatic scale reduction, light Many advances have been made in chemical lithography. Lithography of the current state of the art The wavelength of light used in Fitzur is from mercury G-line (436 nm) to mercury I-line (365 nm) to 248 DUV (KrF laser). Current , 193 nm ArF laser based steppers have been developed, History Historical trends continue. At the same time, the optical system has a numerical aperture (NA) of 0.2 From -0.6 to 0.7.   Further important improvements along these directions are unlikely, and Suggests industry needs to undergo significant changes in lithography technology There are several factors. The main one of these factors is the functional dimension Is shorter than the available light wavelength. Further below the wavelength Even when the reduction in scale makes line width control more difficult, It is increasingly important to control the line width. 193nm ArF wavelength For wavelengths below, it is believed that transmissive optical materials are not available and Easily, a transition to a total internal reflection system will be needed. Current multilayer reflectors and And aspheric optics have not been sufficiently developed to meet these requirements. This is a problem. The transition to reflective optics results in a significant reduction in possible NA This is most likely to reduce the benefits of shorter wavelengths.   A light source with sufficient average power for high throughput manufacturing is For shorter wavelengths, it is another major problem. EUV lithography Plasma source generated by the laser and 5x reduced aspheric total reflection with multilayer reflector A promising approach based on optics. However, this plan is Cost-effective to meet industry demands for graphics capabilities in a timely manner Whether it will lead to a new lithography tool is not yet known.   Another factor suggesting a substantial change in lithography technology is the future ULSI It is related to the complexity of the mask required for the generation. This complexity is by definition Increase by a factor of four per generation (ie, four times the number on the chip) Transistor). In addition, optics, collectively known as resolution enhancement techniques Many of the potential solutions to lithography problems are due to increased mask complexity (Eg, serifs, helper bar, and other sub-resolution features Or replace traditional 2D chrome masks on glass Requires a three-dimensional mask (phase shift technology). These trends are Increases the difficulty and cost of fabricating structures at high yields.   Many other lithographic techniques are currently being studied. These lithography technologies Include X-ray technology, electron beam technology, ion beam technology, and probe tip technology. There is art. Each of the above techniques has advantages and disadvantages, but nonetheless, In the meantime, none of the above techniques are satisfactory alternatives to optical lithography There is no doubt that it is not.   In recent years, interference lithography, or the use of two or more coherent light beams, has Exposure of the photoresist layer using the generated standing wave pattern Provides a very simple technique to create the required scale for several ULSI generations Proven as Optical exposure for low frequencies with spatial frequencies ranging up to NA / 1 Certain techniques exist that combine with interference lithography exposure for high frequencies You. However, these methods limit the high spatial frequency spatial content to a small number of points, As a result, only a periodic pattern having a period corresponding to the high-frequency exposure can be obtained. No. For example, see the following document. "Fine" issued on May 16, 1995 -Line Interferometric Lithography ", Steven R.J. Brueck, Salee m US Patent 5,415,835 to Zaidi and An-Shyang Chu; issued June 1, 1993 Titled "0verlay of Submicron Lithographic Features"  R.J. Brueck and Saleem H.C. Zaidi, U.S. Patent No. 5,216,257; August 30, 1994. `` Method and Apparatus for Alignment of Submicron Lithogra phic Features ”, Steven R.J. Brueck and Saleem H.C. Zaidi USA Patent No. 5,343,292; "Use of Diffracted Light from February 2, 1991" Kenneth, titled `` Latent Images in Photoresist for Exposure Control ''   P. Bishop, Steven R.J. Brueck, Susan M. Gaspar, Kirt C. Hickman, John R. McNeil, S.M. Sohail H. Naqvi, Brian R. Stallard and Gary D. Tipton rice National Patent Application Serial No. 07 / 662,676; U.S. Patent No. issued June 2, 1998 U.S. Patent Application Serial No. 08 / 614,991, filed July 15, 1998, Issue, `` Methods and Apparatuses for Lithography of Sparse Arrays o f Sub-Micrometer Features ”, Steven R.J. Brueck, Xiaolan Chen , Saleem Zaidi and Daniel J .; Devine; "Arra," published September 21, 1993. ngement for Producing Large Second-Order Optical Nonlinearities in aWave guide Structure Including Amorphous SIO2. Myers, Nand ini Mukherjee and Steven R.J. Brueck US Patent No. 5,247,601; 1993 Published on August 24, "Methods and Apparatus for Large Second-Order Non  linearities in Fused Silica, "Steven R.J. Brueck, Richard A . U.S. Patent No. 5,239,407 to Myers, Anadi Muskerjee and Adam Wu; January 1991. `` High Position Resolution Sensor with Rectifying Cont, issued on the 22nd acts ", Steven R.J. Brueck, S.M. Schubert Kristin McArdle and And Bill W. Mullins, U.S. Pat. No. 4,987,461; issued Nov. 14, 1989; Wavelength-Resonant Surface-Emitting Semiconductor Laser '' even R.J. Brueck, Christian F. Schauss, Marek A. Osinski, John G. McIner ney, M .; Yasin A. Raja, Thomas M. Brennan and Burrell E. Hammons US Special No. 4,881,236; "Method for Fine-Line Interferom" filed on September 16, 1992 etric Lithography ", Steven R.J. Brueck, Saleem Zaidi and An-S US Patent Application Serial No. 08 / 635,565 to hyang Chu; filed on March 16, 1995 , `` Methods and Apparatus for Lithography of Sparse Arrays of Sub-Microm eter Features ”, Steven R.J. Brueck, Xiaolan Chen, Saleem Zaid i and Daniel J. Devine US CIP Patent Application Serial No. 08 / 407,067; 1 `` Method for Manufacture of Quantum Sized Periodic, '' filed on September 20, 993 Structures in Si Materials ", Steven R.J. Brueck, An-Shyang Ch u, ruce L. Draper and Saleem H. Zaidi U.S. Patent Application Serial No. 08/123 No. 543, filed on June 6, 1995, entitled "Method for Manufacture of Quantum Sized". Periodic Structures in Si Materials ", Steven R.J. Brueck, An- Shyang Chu, Bruce L. U.S. Patent No. 5,705,321 to Draper and Saleem Zaidi; 19 “Manufacture of Quantum Sized Periodic Structures I, filed on September 25, 1996 n Si Materials ", Steven R.J. Brueck, An-Shyang Chu, Bruce L. Draper and Saleem H. Zaidi US DIV Patent Application Serial No. 08 / 719,896 Issue: “Use of Diffracted Light from Latent Images i, filed March 5, 1992 n Photoresist for Optimizing Image Contrast. Bish op, Lisa M. Milner, S.M. Sohail H. Naqvi, John R. McNeil and Bruce L. Dra U.S. Patent Application Serial No. 07 / 847,618, filed Sep. 8, 1995, entitled "Tec hnique for Fabrication of a Poled Electrooptic Fiber Segment '' , Steven R.J. Brueck and Xiang-Cun Long U.S. Patent Application Serial No. 08/5 No. 25,960; filed on June 20, 1995, entitled "Method and Apparatus for Real-time Spe ckle Interferometry for Strain or Displacement of an Object Surface '' Entitled, Steven R.J. Brueck, David B. Burckel, Andrew Frauenglass and Saleem Zaidi US Patent No. 5,426,498; SIA, National Technology Roadmap fo r Semiconductors (1994); J.W. Goodman, Introduction to Fourier Optics, 2nd Ed. (McGraw Hill, NY 1996); J.W. Goodman, Statistical Optics (John  Wiley, NY 1985); Xiaolan Chen, S.H. Zaidi, S.R.J. Brueck and D.J. Dev ine, Interferometric Lithography of Sub-Micrometer Sparse Hole Arrays fo r Field-Emission Display Applications (Jour. Vac. Sci. Tech B14, 3339-33 49, 1996); S.H. Zaidi and S.R.J. Brueck, Multiple-Exposure Interferome tric Lithography (Jour. Vac. Sci. Tech. B11, 658, 1992); Yariv, Intr oduction to Optical Electronics (Holt, Reinhard and Winston, NY 1971) .   The limiting spatial frequency in interference lithography is generally considered to be ~ λ / 2. available. Here, λ is a laser wavelength, and a 1: 1 line and space Critical dimension (CD) is λ / 4. This is usually k₁say λ / NA It should be contrasted with the limiting CD of the imaging optics system used. Where k₁Is Is a function of manufacturing tolerances and the optical system, where λ is the center wavelength of the exposure system , And NA is the numerical aperture of the imaging optical system. k₁Is typically 1.0 To ~ 0.6. This is an oversimplified theory of restrictive scale Clearly, it plays a key role. 193 wavelength optical lithography In the case of a microtool, the estimate is 0.6, which is ~ 0.19 micron It leads to a limited CD of meters. In contrast, for the I line (365 nm) Interference lithography has a limiting resolution of ~ 0.09 micrometers You. With 193 wavelengths, the limiting resolution of interference lithography is ~ 0.05 mask. It is a micrometer. This is already a current EUV lithography estimate ( Wavelength 13 nm and NA 0.1, k₁0.08 micrometer when Better than a CD).   The major obstacles associated with interference lithography are the VSLI and ULSI contexts. Develop sufficient pattern flexibility to generate useful circuit patterns in the quist is connected with. Two-beam interference exposure simply involves line-an It only produces a periodic pattern of space. Multi-beam (4 or Exposure produces a two-dimensional structure, but is relatively simple, such as holes or columns. It also generates the structure of the repeating pattern. More complex structures, for example, originated on May 16, 1995 `` Method and Apparatus for Fine-Line Interferometric Lithograp hy ”, S.R.J. Brueck and Saleem H.C. Zaidi U.S. Patent 5,415,835 No. and Jour. Vac. Sci. Tech. As described in B11 658 (1992), It can be made by using several interference exposures. For additional flexibility, the patents mentioned above Combining interference lithography and optical lithography as described in Achieved by So far, demonstrations have included, for example, Defining an array of lines and delimiting the field by a second optical exposure, etc. There are relatively simple examples. Multiple exposures are more complex structures, but However, it has been demonstrated to produce repetitive structures.   In addition to the limited pattern flexibility, currently known interference lithography techniques Do not have well-defined synthetic procedures to obtain the desired structure.   Accordingly, there is a need for a technique that specifically overcomes the above disadvantages associated with the prior art. .Summary of the Invention   The present invention provides optical lithography and drying in a manner that overcomes many of the disadvantages of the prior art. A method and apparatus for integrating interference lithography is provided.   The preferred embodiment of the present invention relates to optical lithography technology and interference lithography technology. Method and apparatus for analyzing a lithography task during Particularly suitable fruit According to the embodiment, the integration of interference lithography and optical lithography is facilitated, Optical systems that create complex structures on the same workpiece, for example, a semiconductor wafer System is provided. In a preferred exemplary embodiment, the two masks are To block the two parts of the uniform parallel light beam imaged on the wafer Be composed. The interferometric optical system is integrated into the apparatus and the two mask images are At substantially equal angles and opposite angles with respect to a plane perpendicular to c Bring on wafer. According to this preferred embodiment, after passing through the interference optics, To create an image plane that coincides with the Eha plane, the mask should be properly adjusted with respect to the optical axis. Can be tilted.   According to another aspect of the present invention, any desired pattern can be transferred to a number of specific interference and A formal analysis procedure is proposed to divide exposure and optical lithography exposure.   According to another aspect of the invention, multi-beam interference lithography comprises a number of discrete Spatial frequency, or alternatively, extended to include a continuous range of spatial frequencies This allows more complex repetitive patterns and non-repetitive patterns in a single exposure. To make writing easier.   According to another aspect of the present invention, an enhanced multi-beam interference lithography What you get now using only known optical lithography in combination with lithography Methods and apparatus are provided for making arbitrary structures at higher resolutions than can be provided. You.   According to another aspect of the invention, a mask used in interference exposure is optically defined. Methods and apparatus are provided.   According to yet another aspect of the invention, a combined optical and interference lithography To reduce the complexity of the mask used in the optical lithography part of the exposure Methods and apparatus are provided.BRIEF DESCRIPTION OF THE FIGURES   Hereinafter, the present invention will be described with reference to the accompanying drawings. In the figure, the same reference numerals indicate the same elements. Shows the event.   FIG. 1 is a schematic of a decomposition of a particular structure into rectangles to facilitate a Fourier transform. In the figure, the rectangles are shown slightly offset for clarity.   FIG. 2 shows the optical transfer function for coherent and non-coherent illumination. 5 is an exemplary graph.   FIG. 3 shows an optical lithography tool with limited diffraction at a given wavelength and NA. Prior art VLSI pattern of 0.18 micrometer CD written by It is a diagram showing an example of the pattern, the left column shows the results of non-coherent irradiation, the right column Shows the results of coherent irradiation.   FIG. 4 shows a 0.18 mask written using a combination of optical exposure and interference exposure. 7 shows an embodiment of a VLSI pattern of an micrometer CD.   FIG. 5 illustrates the context of integrated optical and interference lithography technology. 3 shows a simplified mask structure.   FIG. 6 shows a field stop (fi) on a wafer according to a preferred embodiment of the present invention. FIG. 2 is a schematic diagram of an exemplary optical system useful in imaging eld stops.   FIG. 7 extends the optical system of FIG. 6 to include interferometric lithography technology. It is a schematic diagram of a system.   FIGS. 8A and 8B illustrate the generation of a mask image biased to a high spatial frequency. FIG. 3 is a schematic diagram of another interferometric optical system useful for the present invention.   9A and 9B show a rectangular aperture imaged using the optical system of FIG. 8A. SEM micrographs of edge areas, each in focus and 5 shows an SEM micrograph when the focus is out of focus.   FIG. 10 shows a spatial frequency spatial representation of a combination of imaging optical exposure and interference exposure. FIG.   FIG. 11 shows the spatial frequency content (spatial fr.) Of an image according to a preferred embodiment of the present invention. light to bias the frequency (equency content) away from zero center frequency It is a schematic diagram of a learning system.   FIG. 11B is another embodiment of the schematic shown in FIG.   FIG. 12 illustrates the use of a complete pattern mask with prisms to combine sub-masks. Optical system for imaging and biasing frequency components away from low frequencies FIG.   FIG. 12B is another embodiment of the schematic shown in FIG.   FIG. 13 shows interference lithography using the configuration of FIG. (IIL) and 0.18 mm written at 365 nm wavelength 1 shows a schematic example of a VLSI pattern of an Chromometer CD.   FIG. 14 shows an implementation using the method and apparatus described above.Detailed Description of Preferred Exemplary Embodiments   Before describing the present invention in detail, a discussion of Fourier optics is provided. J.W. Goodman, In troduction to Fourier Optics, 2nd Ed. (John Wiley Press, NY 1996) I want to be. In this specification, the entire contents of the above documents are incorporated by reference.   For simplicity, the following description assumes the use of non-coherent illumination. , Each Fourier component of the image can be processed and evaluated independently. Then the explanation is Pushes the limits of coherent irradiation. For coherent illumination, each component is an electric field And the result of squaring to determine the intensity on the exposed surface. In practice, actually Optical lithography tools typically use partially coherent illumination You. For a more detailed description of partial coherence, see, for example, J.W. Goodman , Statistical Optics (John Wiley, NY 1985). Partially coffee The use of lent radiation increases the computational complexity of mathematical analysis, but is not discussed here. As will be apparent, it does not significantly affect the invention in other respects.   Optical lithography systems include an illumination subsystem, a mask, and an imaging optics subsystem. It can be divided into a stem, and a photoresist optical response. Eye of irradiation subsystem The goal is to provide uniform illumination of the mask. Light beam passes through the mask Then, the light is diffracted into a number of plane waves corresponding to the Fourier components of the mask pattern. This Each of these plane wave components is k of the wave number vector._xComponent and k_yCharacterized by component Propagation in different spatial directions. Mathematically, this is in the plane of the mask Thus, it is shown as follows. Here, the addition is performed at the allowed spatial frequency k._x= N / P_xN = 0, ± 1, ± 2,. .; K_y= M / P_y; M = 0, ± 1, ± 2,._xAnd P_yHaso These are the repetition periods of the pattern in the x and y directions, respectively. P_xAnd P_yIs For non-repeating patterns, it can be as large as the exposure die size. Black on glass The system mask has a binary transmission function (each pixel is either 1 or zero). Therefore, the Fourier transform of (1) simply corresponds to the decomposition of the desired pattern into rectangles. Given by the sum over the sin (x) / x function with the appropriate phase shift You. That is, it becomes as follows. Where a_i(B_i) Is the extent of each rectangle in x (y), c_i(D_i) Is x This is the offset of the center of the rectangle from the coordinate origin in (y).   Referring briefly to FIG. 1, the above decomposition of the desired pattern into rectangles is a typical VLS Shown schematically for an I-gate structure. In particular, typical in the illustrated embodiment The exemplary pattern shown as a typical gate 11 is divided into a plurality of rectangles 12-16. Understood. For clarity, these rectangles are shown offset in the figure You.   The imaging optics subsystem modulates the propagation of the Fourier component to the image (wafer) plane. Imposes a transfer function (MTF). In the case of diffraction-limited optics with circular symmetry, transmission The function is the spatial frequency k_opt= NA / λ, where NA is the numerical aperture And λ is the center wavelength. In the limitation of non-coherent illumination, the MTF is Go odman, Introduction to Fourier Optics, 2nd Ed. (McGraw Hill, NY 1996) Given by reference. That is, it is as follows. = 0. For a coherent optical system, the transfer function is simply become. (4) k ≦ k_optIn the case of T_coh(K_x, K_y) = 1, and                 k> k_optIn the case of T_coh(K_x, K_y) = 0 Here, as described above, this transfer function is suitably adjusted before evaluating the aerial image intensity. Applies to optical electric fields.   Referring now to FIG. 2, these two transfer functions are described in more detail below. Along with the transfer function in the case of interference lithography. The scale is k_opt= Set by NA / λ, where NA is the numerical aperture of the optical subsystem, λ is the center wavelength of the illumination system. In the case of non-coherent illumination 21, transmission Function T_incohIs applied to each Fourier component of the intensity pattern at the mask. Kohi Transfer function T_cohIs the Fourier component of the electric field at the mask And the strength is evaluated on the wafer surface. Interference lithography up to 2 / λ The optical transfer function of the filter (23) is also shown. Used to plot this figure The parameters set are λ = 365 nm and NA = 0.65. Fig. 2 H, k / k_optRemoves any obvious dependence on λ by normalizing to It is.   Finally, the image intensity at the wafer surface is transferred to the resist. Resist is typical Shows a non-linear response. In this regard, the industry typically uses positive and negative Both tone resists are used. Shown here for clarity The calculations performed are for a negative tone resist (ie, resist irradiation). Areas are retained during development and non-irradiated areas are removed). Calculate simple mask By reversing the irradiation area and the non-irradiation area, it can be used as a positive tone resist. It is understood that the above can be applied properly. Photoresist The answer is approximated as a step function. That is, when the local intensity is smaller than the threshold The photoresist is assumed to be completely removed during development and is larger than the threshold. For high local strengths, the resist thickness is not affected by the development process. actually Resist has finite contrast and non-local intensity affects development . The effect of these practical considerations is that high spatial frequency variation in the results presented below is Eliminating some of the motion and the steep sidewalls predicted by this simple model Rather than producing a finite sidewall slope. However, these effects are Does not affect the scope or content of   Referring now to FIG. 3, examples of the prior art are 365 nm, 248 nm, and 1 nm. 0.1 with a 93 nm optical lithography tool (each from bottom to top in the figure) This known technique is used to print a typical VLSI gate pattern on an 8 micrometer CD. The result of applying Fourier analysis is shown. The desired pattern is indicated by a dotted line in each cell. You. Only the perimeter of the pattern printed by the optical tool is shown. Around these Inside, the resist is exposed and remains intact during development. Outside the surrounding The resist is not exposed and is removed during development (negative resist). As above In the case of a positive resist, reversing this response is a simple modification to the analysis. is there. The pattern is periodic. Therefore, in actual exposure, each cell is repeated many times. Returned and, of course, the frame that outlines each cell is not printed. In FIG. Different exposure tools and exposure conditions to provide more information Neighboring cells with the results are shown. NA is 365 nm (I line) Set to 0.65, 248 nm (KrF laser source) and 193 nm (ArF (Laser source) was set to 0.6. The left column shows the results of non-coherent irradiation. The right column shows the results of coherent irradiation. Simple solution shown in background chapter As expected from the analysis, I-line optical lithography tools can Cannot write the structure of the title CD easily (for example, ~ k₁λ / NA ~ 0.8 ×. 365 /. 65-0.45 micrometers minimum resolution). Printed Shape can result in severe distortion due to limited available frequency components Is shown. In fact, an image of coherent illumination is not even two separate features But combine into a single structure. Although not shown in FIG. 3, the pattern is also Are very sensitive to process variations and are sensitive to small variations in optical exposure levels. Change Significant patterns can be obtained from 248nm optical tools However, these patterns still have significant roundness at the edges. And shows the deviation from the desired structure. Can be obtained from 193nm tool Even patterns are far from ideal.   The above procedure describes the lithography tasks between optical lithography and interference lithography. Can be used to parse the query. This analysis technology is at the heart of the present invention. is there.   Before describing the analysis technique of the present invention in detail, a review of the MTF for interference lithography. The value is shown first.   More specifically, the IL is, in its simplest form, in a plane perpendicular to the wafer. Incident on the substrate at equal and opposite azimuth angles (θ) Use two coherent light beams. The strength at the wafer is given by the following equation: available. Here, the intensity of each beam is individually I₀And k = 1 / λ, where λ is the coherent Is the wavelength of the transparent beam and the phase factor φ is the Indicates the position of the Multiply this phase factor to form the desired final pattern It is necessary to make appropriate adjustments in exposure, that is, Can get. This means that the MTF is 1. That is, the intensity is at least zero. Space circumference The wave number is given by the following equation: (6) k_x= 2 ksin (θ) The maximum spatial frequency is k_x|_max≡k_IL= 2 / λ. This is an imaging optical system The maximum spatial frequency in the case of_max= 2k_opt= 2NA / λ It is. In addition, the modulation transfer function for interference exposure is Unlike a sharp drop in the case, typically all k <k_ILIt is 1 in the case of. This includes optical imaging systems for coherent illumination and non-coherent illumination. It is shown in FIG. 2 with the corresponding MTF in both cases of the optical imaging system. Ko The MTF of the helical irradiation is not the Fourier component of the intensity, but the Fourier component of the electric field. It is important to remember that it applies to minutes. Involved in taking strength The nonlinear square operation is 2 × k_optTo generate frequency components up to.   Thus, the defined procedure for exposing a desired pattern according to the present invention is optical Related to optical lithography and interference lithography. Particularly preferred embodiment bear If optical lithography is mainly used to provide lower frequency components , Interference lithography is mainly used to provide higher spatial frequency components It is. The threshold is determined by the frequency of the interference exposure (ie, the maximum and minimum spatial frequencies) and Amplitude (any frequency component whose Fourier amplitude is lower than a preset level Can be set for both. This is used in the prior art example (FIG. 3). The same VLSI pattern as shown is shown in FIG.   Continuing to refer to FIG. 4, the left column shows an embodiment for setting a frequency range. Upper left For panels, the full frequency space available through interference lithography is used and Thus, in this case, no optical lithography step is required. Resulting Pattern more closely represents the desired pattern as exemplified by any prior art. The desired pattern, even at substantially shorter wavelengths. This is shown in FIG. However, in this example, 51 exposures were required. In the left column The two panels underneath are first the low frequency limiting embodiment, then the low and high An example is shown that limits both frequencies. In each case, the low frequency components are Thus obtained. The right column sets a threshold value for the intensity of the interference lithography exposure. The results are shown. As the threshold gradually increases (from top to bottom), the interference lithography exposure The number of times of light gradually decreases, and the approximation to the desired structure gradually becomes less ideal. It becomes. This phenomenon is due to the number of exposures, Emphasis on the trade-off between on-line fidelity. This tradeoff is preferably Can be optimized in each level of specific context according to the teachings of the present invention .   As mentioned above, one of the major difficulties facing optical lithography is the need This is an increase in the complexity of the mask. The analysis procedure outlined herein is Preferably, optical lithography is used for the low spatial frequency components of the image to reduce interference interference. The complexity of the mask is dramatically increased by providing high spatial frequency components through lithography. Can be reduced.   Referring now to FIG. 5, an exemplary VLSI similar to the pattern shown in FIG. The pattern is again shown using the above paradigm. That is, optical lithography Is used for the low spatial frequency components of the mask (leveraged) and Luffy is used for high spatial frequency components. More specifically, in the upper left of FIG. The panel of the low spatial frequency component (k_opt= NA / λ (In dotted line) shows the resulting image (dotted line) ). The upper right panel shows the result of adding 51 interference exposures (solid line) (Fig. 4 upper left panel) and 7 using interference amplitude lithography with simple amplitude threshold The result (broken line) of limiting the exposure to only one exposure is shown. These results are shown in FIG. Substantially equal to the result. The middle panel on the left in Figure 5 is a much simpler mask (Referred to as simplified mask A and shown in dashed lines) and Without limiting spatial frequencies beyond the spatial frequency, this Shows the resulting intensity profile when passing the image from the disc . This is a much simpler mask than the complete pattern, It produces very similar results in a limited frequency space. Mask pattern A is Passing the pattern of A through the imaging simulation described in the previous paragraph Was obtained by simple trial and error from the low frequency results of the upper panel. Repetition For patterning, pattern A is simply one rectangular opening per repeating unit. Note that   With continued reference to FIG. 5, the middle panel on the right shows the simplified pattern A. Shows the result of using optical exposure and adding interference lithography exposure (no threshold) The complete range of available spatial frequencies), this panel is about a complete mask Very close to the panel. The dashed curve is for seven interference lithography exposures. You. The low frequency mask is further added to the simple straight line segment shown in the lower left panel. It can be simplified. Due to the repeating pattern, this simply extends to the full height of the die It is a wide line.   The lower right panel of FIG. 5 shows 51 (solid line) and 7 (dashed line) interference lithographs. 3 shows the result of adding a high frequency component by using the exposure. The result is a complete mask Is very close to the result obtained in this case, again with a 193 nm optical exposure tool. Much better than the results you can get. Between the resulting exposures The difference is very slight, because the simplified mask Showing that it can be used for tep, while still maintaining excellent pattern fidelity Relieves the difficulty of fabricating masks. As described above, for each level, Careful trade-off between number (manufacturing throughput) and pattern fidelity Need to be   By combining multiple (two or more) beams in one exposure, It is also possible to reduce the number of cross exposures. A simple example of this is "Meth ods and Apparatuses for Lithography of Sparse Arrays of Sub-Micrometer F eature ", Steven R.J. Brueck, Xiaolan Chen, Saleem Zaidi and Daniel J. Related US Patent Application No. 08 / 399,3 filed February 24, 1995 by Devine Issue 81 provides some details. This earlier application writes a specific two-dimensional exposure Specifically, a sparse hole array (> 1) is formed in the positive photoresist layer. : 3 hole diameter: distance between holes). the above The application provides specific examples of three, four and five beam exposures.   It is often desirable to demarcate the field of interference lithography exposure And it is often necessary. Typical die size for one exposure (Today, 20 × 30mm^TwoIs the wafer size (diameter 200mm ~ 300 mm). In order to obtain uniform exposure over the die, The beam is expanded and turned to a uniform intensity over the field size. Only And the beam edges are necessarily non-uniform, allowing the wafer to be exposed Even so, the pattern becomes substantially non-uniform. One technique that addresses this is c. Add a field stop (mask) opening directly above Eha to define the boundary of the exposure area It is to determine. However, a significant problem with this approach is the diffraction effect of the aperture. Where L is the distance from the field stop to the wafer. L is about 1mm For practical separation distances, this diffraction ringing is Extends about 0.3 mm into the pattern. In many applications, this is an unacceptable result It is. One technique to eliminate this ringing is to scatter the scattered edges. The wavelength scale so that the field is not added coherently away from the edge. Is to roughen the edge of the opening randomly.   In addition, it offers significantly increased flexibility without leaving traditional lithography Another alternative when moving the field stop away from the wafer is to perform two functions: Is to add an optical system that simultaneously provides This function is 1) Imaging the field stop on the wafer, and 2) the field stop The purpose is to make the incident parallel beam a parallel beam at the position of the wafer .   Referring now to FIG. 6, field stop 31 is imaged on wafer 32. A preferred exemplary embodiment of the optical system used for₁And f_Two And the mask 31 includes the first lens 3 Distance f from 3₁And the spacing between the lenses is appropriately f₁+ F_TwoIs , The wafer 32 is at a distance f from the second lens 34._TwoJust placed behind. Double the statue The rate is -f_Two/ F₁Is indicated by The field stop is suitably the first level Placed in front of the lens 33 and collimated (ie, the wavefront has a very large radius of curvature). Laser source that is substantially uniform over the area of the field stop. Irradiated. In this configuration, the curvature of the wavefront effects the effect of this optical system. Not qualitatively. Diffraction limited edge of field stop image on wafer The edge definition is appropriately proportional to the wavelength of the optical system, and , Is inversely proportional to the numerical aperture. This transfers the mask image and at the same time the entire wavefront This is only one class of optical systems that serve to maintain flatness. Suitable A general condition for specifying a clear optical system is the ABCD ray that indicates the optical system. The B and C terms of the entire transfer matrix are zero (A . Yariv, Introduction to Optical Electronics (Holt, Reinhart and Winston , NY 1971), the description of the ABCD ray tracing transfer matrix).   The mathematical description of this imaging system is based on the Fourier optics concept introduced above. Is easy. Since mask irradiation uses a coherent uniform beam, Analysis of a healing imaging is appropriate. The electric field immediately after the mask is written as Can be Here, M (k_x, K_y) Is the Fourier transform of the mask transmission function. It is assumed to be sporadic. For an aperiodic mask pattern, k_x, K_ygarden Barrel addition is replaced by integration in the usual manner. Through the optical system , The following modulation transfer function is imposed: Here, the subscript E indicates that this transfer function applies to an electric field instead of an intensity. It is shown. The electric field at the wafer is then given by: The strength is given by the following equation. Here, I is the Fourier transform of the intensity, which is a real number in the case of a simple transmission mask. k_x’And k_y′ Is the square of the electric field Fourier transform k_xAnd k_y2xk_optTo reach   The low frequency pattern defined by the mask splits the optical path and guides the interference optics. By entering, a higher spatial frequency may be shifted. Referring now to FIG. Thus, a preferred exemplary embodiment of the present invention implements the optical system of FIG. Fig. 2 shows an extension to include a device for integration into a graphics system. Than Specifically, each of the masks 41 and 42 is appropriately positioned above or below the figure. Are introduced into two parts (e.g., half) of the light beam. Mask 41 And 42 are not necessarily the same. In-phase wave by last optical element To compensate for any tilt introduced into the plane, these masks are Can be arranged, preferably at an angle, such that is in a plane perpendicular to the wafer. Optical cis The system comprises lenses 33 and 34 arranged as described with reference to FIG. Become. Finally, interferometric optics are introduced to provide a high frequency bias.   The interferometric optical system shown in FIG. 7 suitably has a plurality (eg, four) mirrors. (45, 46, 47 and 48), which mirror the light beam Each segment (for example, two) is divided into Interference is caused on the plane of Eha15. In the illustrated embodiment, an interferometric optical system Suitably at substantially equal and vertical angles to a plane perpendicular to the wafer. , To provide a mask image on the wafer. The advantages of this system include: The two beams have equal central optical path lengths and no induced astigmatism. I can do it.   Referring now to FIG. 8, a mask image biased to a high spatial frequency is generated. Various other embodiments may be used for this. In particular, FIG. For example, a lens 34 and a lens 34 configured to impart a mask image to the wafer 32 A simple Fresnel configuration including a mirror 51 is used. The configuration of FIG. It is attractive in that it is a much simpler configuration involving only one of 1), Two central optical path lengths are not equal, requiring different mask planes for the two masks And FIG. 8B shows a prism (52) configuration where the center optical path lengths are equal. But the prism requires different mask planes for x-rays and y-rays, Introduce aberrations.   Here, a mathematical description of these systems is obtained with respect to FIG. This theory In order for the light to be compatible with the alternative optics of FIG. 8, it does not affect the essential results. Small changes are needed. Applying equation 9 to the optical system of FIG. Is given.Here, the subscripts u and l have the same mask Fourier transform. Has been added to the mask Fourier transform to show that it is not limited. w = 2Π sin θ / λ is the spatial frequency bias added to each beam by interference optics You. Taking the intensity gives an image to be printed on the photoresist layer. That is, It looks like below. And in the special case that the two masks are the same, this becomes . Equation (13) corresponds to the mask image modulated by the high frequency interference pattern . For example, if the mask pattern is simply a field stop, the resulting The printed pattern is a high Spatial frequency line: Space pattern. As described above, The edge of the loop is typically defined within a distance of ~ λ / NA due to optical system limitations. Is determined. Clearly more complex mask patterns are possible. In fact, the optical system Any mask pattern within the spatial frequency range of the It can be reproduced with an additional high spatial frequency modulation introduced by the stem.   Referring now to FIG. 9, a uniform within a finite field using the optical configuration of FIG. Line: Multiple SEM micrographs showing first experimental evidence of printing a space pattern A mirror photograph is shown. The light source was a 364 nm Ar ion laser. this is, Very low NA optical system using only a single element uncorrected spherical lens (.About.0.06) and diffraction limited edge resolution is only .about.6 micron. Meters. Probably a significant contribution from lens aberrations . The top two SEMs (FIG. 9A) show the case in focus. Fields G Vertical (parallel to the grating line) and horizontal (dry) Both perpendicular to the grating line) are defined within -10 micrometers. This Is 2 times the theoretical diffraction limit (λ1 / NA〜4 micrometers). Within a factor of ~ 3. Note that the grid period of 0.9 micrometer is Qualitatively, a built-in measuring device is provided. In contrast, the lower SEM (FIG. 9B) Similar results are shown for a defocused state. The intensity fringing at the edge It is due to the occasion. The distance to the first stripe of ~ 30 micrometers is Can be used to calibrate the distance from the point plane (ie, defocus) at about 3.5 mm You.   Equation (13) rewrites the distribution of spatial frequency contributions associated with this image Can be rewritten. That is, it becomes as follows. A closer examination of equation (14) reveals three frequency space domains with frequency content. It is suggested that there is. That is, it is modulated by the lens system and A low-frequency region represented by two different copies of this intensity pattern. These duplicates One is shifted by + 2wx as a result of interference optics and one is -2wx Shift. This situation is shown in FIG. 10A, which is shown in FIG. 8A. Using an exemplary optical system such as that covered by the exposure shown in Equation 14, The complete spread of the frequency space is modeled appropriately. In particular, available laps The complete spread of wavenumber space is given by the radius k_ILModeled as a large circle 104 of = 2 / λ The three filled circles 106, 108 and 110 have significant frequency content. Represents three frequency space regions having The optical system used to demonstrate FIG. In this case, the low-frequency region has a radius of 2 k with NA = 0.06._opt= In the circle of 2NA / λ Included, two regions created by interference are off by 2w = 2 sin (θ) / λ Set, for a 0.9 micrometer pitch grating, sin (θ) = 0. 2. The radius of each offset region is equal to the radius of the low frequency region. equation( From 3), the MTF of the optical system has a radius in each of these frequency ranges. It decreases monotonically along the direction and is zero at the edge of the circle shown in FIG. 10A. still In the lens optical system alone (FIG. 6), a grating with a period of 0.9 micrometer is used. Cannot be made and the interference system combined with the lens system (FIG. 8A) Generates a grid while preserving the frequency content of the image and the low frequency content that defines the area of the image. To the higher frequencies needed to achieve this. Importantly, The combined optical system covers the frequency space region where the frequency content is continuous. A barred image is obtained. This is because a pattern with a small period (x is 12 × CD, y Only the points in the frequency space that are relatively widely spaced about 5 × CD) This should be contrasted with the above description of the periodic structure needed to reproduce the pattern. (See, for example, the above description of FIGS. 1, 3, 4, and 5).   For this first demonstration, a very conservative optical system was used. FIG. 10A As can be seen from the figure, the resulting pattern is Covers only a very small area. FIG. 10B shows a small number of exposures and duplicates. For example, in the case of a lens NA of 0.33, Shows the potential to cover many. The offset region in the frequency space is k_xDirection and K_yShown in both directions. These are the optical systems shown in FIGS. In two exposures using an optical system similar to the d Apparatuses for Lithography of Sparse Arrays of Sub-Micrometer Feature s ", Steven R.J. Brueck, Xiaolan Chen, Saleem Zaidi and Danie l J. Devine, U.S. Patent Application Serial No. 08/399, filed February 24, 1995, As described in 381 for plane waves (eg, a single point in frequency space) In one exposure using a suitable four-beam interference system. k_xShaft And k_ySince the relative phase of the frequency components along the axis can vary from pattern to pattern, It is likely that two exposures will be required. NA of λ / 3 is the diameter between three circles Along with the smallest NA at which continuous coverage of the frequency space is achieved , This NA of λ / 3 was selected. The frequency response of the optical system is at the edge of the circle Because of inadequacy, in order to achieve full coverage, additional exposure or Yo Any of the larger NA optical systems is required. Frequency of linear pattern Most of the content is k_xAxis and k_yFocused on values close to the axis. Therefore, this frequency Number space coverage can be satisfactory. If not, additional dew Light or additional beam paths, or both, may be used as needed.   Sufficient exposure at a NA large enough to ensure complete coverage of the frequency space Even with light, this configuration simply does not allow imaging of any pattern. This is because each intensity pattern represents the actual mask image, which D) Some restrictions are imposed on the components. Specifically, the simpler equation 10 In the case of a transmissive mask, the intensity after the mask must be real and positive Thus, the following equation is required. The same relationship must apply to the final image. However, interference optical systems The system imposes a more restrictive relationship, for example: Again, for the final image, the pairs shown in each row must apply. No, but due to the requirement that a relationship be applied between these pairs , The final image is overly constrained.   According to the present invention, at least the following three approaches for arbitrating the above constraints are provided. There is a solution. (1) Breaking this symmetry by repeating a number of exposures, (2) Using a phase mask or other three-dimensional mask, the overall Breaking symmetry, and (3) using different masks for the upper and lower arms And by introducing additional optics to center the frequency response by k_x = K_y= 0 away from the positive and negative frequencies due to the lens MTF Modify the optical system by giving different weights to the terms That is.   In the special case of practical importance (3) (above), one of the interference systems Arm, for example, the upper arm, only the field stop opening is arranged, and This can be obtained by placing a more complex mask on the lower arm. Then Equation 12 can be written as: Equation 17 states that the field stop is large enough that M_uRelated to Fourier The component is M₁Under the assumption that the frequency is much lower than Was derived. Specifically, each frequency of addition in Equation 17 represents a pattern, Replaced by an appropriate function to limit the area defined by the field stop obtain. That is, it becomes as follows. Where a and b are the linear dimensions of the field stop and the Fourier transform pair It is used that the multiplication and convolution of are equal.   This result still completely solves the constraint problem presented with respect to equation 16. May not be resolved.   In particular, Fourier coefficients, for example, Ml (k_x, K_y) = Ml^*(-K_x, -K_y) Has the constraint that it does not match any final image. As mentioned above, this is optical By shifting the center frequency response of the system away from | k | = 0 Can be avoided. One possible optical mechanism to accomplish this is shown in FIG.   Referring now to FIG. 11, effective masks 41 and 42 are used Suitably include a simple open field stop opening. Prism 71 Suitably positioned behind mask 42 to provide an angular offset of frequency components I do. Center rays 72 and 73 are shown to provide additional information. Prism 71 preferably has a zero frequency component emerging from mask 42 It is selected to be directed to the edge of the 33 opening. In this configuration, the prism 7 1 gives the overall slope or bar to the frequency components accepted by the lens system. Ias w_tiltEffectively. In the embodiment shown, the prism angle is The light beam of zero spatial frequency emerging from the lens 42 is tilted and this light beam is It is chosen to just avoid the opening of 3). Use this result as an upper mask Limited to a simple field stop aperture of D coefficient is much lower than the mask frequency and can be ignored at first And the sum field in the mask is: The strength is as follows. W_tiltTo, for example, a positive k_xOnly adjusted to be accepted by the lens system By doing so, the final constraint on the formation of any image is removed. FIG. There are many possible variations in the use of a prism as shown in FIG. For example, Priz The prism may be placed in front of the mask as in FIG. May be replaced with a suitable diffraction grating used in any of the reflections, A Fourier plane filter is used to select only the appropriate subset of the mask transmission. (E.g., the aperture of the focal point of the first lens (33) of the optical system). .   The calculation of a pattern that can be generated using this concept is shown in FIG. Upper left The panel (indicated as “all frequencies”) shows all reasonably available interference exposures Shows the results of using. This panel was included for comparison and was 4 shows the best pattern obtained, and is essentially identical to the upper left panel of FIG. You. The middle panel on the left is the double exposure, i.e. imaging interference exposure and non-coherent exposure. 4 shows the results of optical exposure applied to the paint. The middle panel on the left is shown in FIG. With one configuration, a single connection offset (biased) in the y direction, that is, the vertical direction, is used. Image interference lithography (IIL) exposure and traditional non-coherently illuminated light Result of the combination of the chemical exposure and the 0.4 A lens NA was used. In the example shown in the middle panel on the left, The cross exposure is preferably performed, for example, by a prism 71 of FIG. It can be biased at the wave number. According to a preferred embodiment, this bias is then It is effectively canceled by the optics (eg, mirrors 45-48 in FIG. 7) and ultimately , Generate the appropriate frequency distribution for the pattern. Mass with desired pattern Is suitably located on one arm of the optical system to outline the field The open opening to be drawn is suitably located on the other arm. The fruit shown in FIG. In the case of the embodiment, the pattern is P in the horizontal (x) direction._x= 12CD, in vertical (y) direction P_y= 5CD. The Fourier series representation of the pattern is Number component k_{n, m}= 2Π (nx / P_x+ My / P_y). Pair normalized to CD The corresponding spatial frequency is CD / P_x= 0.083 and CD / P_y= 0.2, (N, m) current practically feasible maximum supported by interference lithography The value is (11,5). That is, n_max= Int (2P_xCD / λl) = Int (P_xk_IL) Where the Int function returns the integer part of the argument. Prism 7 1 means that all of the frequencies_tite= 0.5 was chosen to tilt. Therefore , The high spatial frequency Fourier component, eg, m = 3,4, Shifted to a cone. Introduced by interference optics, for example, mirrors 45-48 in FIG. Bias W_ILIs 2W_IL= W_tiltAnd the offset tilt is interference optics Was chosen to just cancel the added bias.   Continuing to refer to FIG. 13, the lower left panel shows a case where the lens NA is 0.7. 2 shows a similar calculation. Interestingly, the left and right edges of the pattern The bulges more when the NA is larger, but in the center of the lower feature. The knob is less well defined for larger NAs. This is mainly due to this feature X is not optimally provided by either interference exposure or imaging optical exposure, x This is because a higher spatial frequency component in the direction is required. This is the upper right panel Can be improved. This top right panel makes this exposure more satisfying Once the same W to provide the spatial frequency needed to define_tilt= 0.5 Two imaging interference exposures, one biased to y and one biased to x by the same amount Shows the results of using with non-coherently illuminated standard optical exposure. This All of these exposures had a lens NA of 0.4. Obviously, the knob is The case of exposure is better defined. The horizontal bar undulations are lower This occurs because the wavenumber components are not exactly at the correct amplitude and phase. these The components result from each of the imaging interference exposures (the second term of Equation 19) and the optical exposure You. Most practical microelectronic patterns always have both directions And it is highly likely that two interference exposures are required. Middle right Panel replaces the traditional optical exposure with a single low frequency interference exposure (λ / P_xof Frequency). This result is very similar to the previous panel, The required exposure is much simpler. Finally, the lower right panel shows λ / P_x Imaging exposure using a lens with NA = 0.7, with one interference exposure at Fig. 4 shows the calculation of the image of the field table using light. This result is the best possible Similar to a pattern, but requires only three exposures Equivalent to the full complexity of box processing. These examples are based on imaging interference exposure and traditional It plays a role in showing the flexibility enabled by the combination with the proper imaging exposure. Desired There are a number of ways to reach the results.   Very importantly, imaging interference exposure is not limited to periodic arrays of structures . This is evident from earlier examples of imaging non-repetitive field stops. Proven. Any arbitrary pattern can be shown as a Fourier series, where , The repetition period is the exposure field. For a typical ULSI scale, this Has a very large number of Fourier components that must be included (eg, potentially Means on the order of 100 million). Obviously, this means that the individual Fourier components Although impractical when exposed separately, imaging optics that handle many frequency components Imaging interference exposure combining the ability with the ability of interferometry to enable high spatial frequencies It doesn't matter if   Several possibilities for optimization are further explored in the context of the present invention Can be done. For example, if the entire exposure is divided into several sub-exposures, these exposures Adjusting the relative strength of the is a simple procedure. Further, the imaging interference exposure requires 2 Need one (or possibly more) light paths, and again, these light paths Can be adjusted. In addition, the last two described Adjusting the lens NA between the various sub-exposures as shown by the panel (Figure 13) May be advantageous. That is, the lens NA for non-coherent optical exposure Reduces to the point where only a single frequency component used in simple interference exposure is passed Have been. The two beam intensities of the interference exposure are considered equal and the The intensity was also equal for all of the imaging exposures. A simple interference exposure (ie, (Only two beams and a field stop aperture on each arm) The intensity is adjusted to provide the best qualitative fit to the desired pattern. It was arranged.   Hands to identify and manufacture the masks needed to define any image The order remains to be determined. Note that different frequency space regions form a complete image In this case this is a series of sub-masks, since different sub-masks are needed to It is a mask. The total number of sub-masks is used for the spatial frequency content of the image, especially for exposure And the required image fidelity. The mathematics outlined above gives a functional design device. But the actual UL This becomes a difficult task due to the complexity of the SI pattern. In principle, reproduction The number of spatial frequencies to be performed can be the same as the number of pixels in the image, 2 × 3 cm^TwoStatue of And 2 × 10 for a 0.18 micrometer CD^Ten! Can be the number of things.   Referring now to FIG. 12, another approach uses optics to complete the pattern. Is to generate a mask from the original mask specified in (1). Lines, helper bar Or without any increase in resolution, such as phase shift Complete pattern written by current mask fabrication procedure of write patterning The irradiation mask (81) is irradiated with non-coherent light. Behind the mask An existing wavefront tilting prism (82) similar in function to the prism of FIG. is there. Alternatively, as shown in FIG. 11B, the prism (82) is arranged in front of the mask 81. Is placed. This is the sky imaged by the optical system (lenses 33 and 34). Bias W at inter-frequency_biasIs introduced. The intensity pattern on the wafer surface is It is represented by the following equation.   Instead of imaging the wafer, this intensity pattern creates a mask blank (83). Can be used to expose. The mask blank (83) will later form the sub-mask. Developed and patterned to form. This is exactly what the required foot 11 can be used in the arrangement of FIG. Sub-mask. If the following relationship is satisfied, the appropriate spatial frequency on the wafer surface It is easy to show that generation is facilitated. (22) 2W = W_bias   FIG. 14 shows that modeling is valid by showing excellent agreement between experiment and modeling. An example is shown using the above-described method and apparatus. The result is The above invention leads to a dramatic increase in resolution for a given lens system. Are clearly shown. With regard to these experiments, the inventors have developed an inexpensive achromatic doublet. It was used. To ensure a relatively aberration-free image, the lens should be only 0.04 Of NA. The pattern contains five nested "L" s, and the central feature is Elongated to provide separated lines, 10 × 10 CD^TwoLarge box shape is dense Immediately adjacent to the line that was made. Repeated putters to provide large exposed images Were separated by 40 CDs, providing a rigorous test for the effects of scattered light. From this viewgraph, the feature is a 2 μm CD.   The upper left panel shows the result of a conventional image (coherent illumination). On the above CD Small features (high spatial frequencies) are not resolved. The lower left panel shows the model The results are shown. Very well matched.   The middle panel on the top says that one exposure is in the x direction and the second exposure is in the y direction. The results of two consecutive offset exposures are shown. Dense line: space putter The high spatial frequency information corresponding to the signal is captured, but no low frequency information. in this case However, the modeling results are in good agreement. Especially, the corner of "L" is consistent with the experiment Note that it is not printed in the manner described.   Finally, the left panel shows three consecutive exposures of imaging interference lithography. (IIL) shows the results. The complete pattern is printed. Again, the model Very well with Le.

[Procedure for Amendment] Article 184-8, Paragraph 1 of the Patent Act [Date of Submission] August 18, 1998 (August 18, 1998) [Contents of Amendment] [Figure 7] FIG. 11 FIG. FIG. 14

────────────────────────────────────────────────── ─── Continuation of front page    (81) Designated countries EP (AT, BE, CH, DE, DK, ES, FI, FR, GB, GR, IE, IT, L U, MC, NL, PT, SE), OA (BF, BJ, CF) , CG, CI, CM, GA, GN, ML, MR, NE, SN, TD, TG), AP (GH, GM, KE, LS, M W, SD, SZ, UG, ZW), EA (AM, AZ, BY) , KG, KZ, MD, RU, TJ, TM), AL, AM , AT, AU, AZ, BA, BB, BG, BR, BY, CA, CH, CN, CU, CZ, DE, DK, EE, E S, FI, GB, GE, GH, GW, HU, ID, IL , IS, JP, KE, KG, KP, KR, KZ, LC, LK, LR, LS, LT, LU, LV, MD, MG, M K, MN, MW, MX, NO, NZ, PL, PT, RO , RU, SD, SE, SG, SI, SK, SL, TJ, TM, TR, TT, UA, UG, UZ, VN, YU, Z W (72) Inventor Chen, Xiaolan             United States New Mexico 87100,             Albuquerque, A 205, Buena Bis             TA Drive S. E. 929 (72) Inventor Flowengrass, Andrew             United States New Mexico 87106,             Albuquerque, Girard S. E.             417 (72) Inventors Zaidi, Salem H.             United States New Mexico 87111,             Albuquerque, Fostoria Road             N. E. 9813

Claims (1)

[Claims] 1. A two-dimensional spatial pattern in a photosensitive material on a substrate,   A first mask for providing illumination of a first mask featuring the first mask pattern; And an imaging system for imaging the first mask pattern on the photosensitive material on a substrate. Using a first optical arrangement including a first imaging system A first exposure of the reactive material, the first exposure having a first intensity pattern;   A second exposure of said photosensitive material using a second optical arrangement, said second exposure comprising: Exposure 2 has a second intensity pattern and the spatial frequency of the two-dimensional spatial pattern Combining the first and second intensity patterns to provide first and second subsets. Together, this defines a second spatial pattern in the photosensitive material, Exposure and   Processing the photosensitive material, thereby illustrating the two-dimensional spatial pattern; ,   Two-dimensional spatial pattern in photosensitive material on a substrate resulting from the above process. 2. The process of claim 1, wherein said photosensitive material is a photoresist layer. 3. The process according to claim 1, wherein the substrate is a wafer. 4. Due to the processing, the photosensitive material is converted into the substrate according to the two-dimensional spatial pattern. The photosensitive material is acted upon as a mask to modify the proper properties of the plate. The process of claim 1 wherein the physical change occurs. 5. The second optical arrangement comprises a second mask pattern. A second illumination system for providing illumination of the mask; A second imaging system for imaging the second mask pattern on the conductive material. The process of claim 1. 6. The third optical arrangement is characterized in that the second optical arrangement is characterized by a third mask pattern. A third illumination system for providing illumination of the mask, the second mask pattern And the electric field corresponding to the third mask pattern interferes coherently, and Thereby, an intensity pattern is provided on the photosensitive material on the substrate. A third imaging system for imaging the third mask pattern. The process according to claim 5. 7. 7. The method of claim 6, wherein the third mask is substantially identical to the second mask. On-boarding process. 8. The first and second exposure steps are performed continuously in time. The process of claim 1. 9. The step of combining, using an illumination source having orthogonal polarization, Adding the first and second intensity patterns substantially simultaneously. The process of claim 1. 10. The combining step uses mutually non-coherent illumination sources, Adding the first and second intensity patterns of the exposure substantially simultaneously. The process of claim 1, comprising. 11. The first exposing step comprises a first wavelength λ₁And a first opening NA₁Using a first optical lithographic exposure system, characterized in that A step of exposing the photosensitive material to a low spatial frequency component of a two-dimensional spatial pattern. And the spatial frequency magnitude of the low spatial frequency component is about NA₁/ Λ₁Yo The process of claim 1, wherein the process is smaller. 12. Wherein said second exposure exposes using a multi-beam interference exposure, whereby Providing a high spatial frequency component of the two-dimensional spatial pattern to the photosensitive material; The spatial frequency magnitude of the high spatial frequency component is approximately NA₁ / Λ₁Greater, the second exposure   Established by the angle of incidence of the multi-beam on the photosensitive material on the substrate An intensity pattern on the substrate at a spatial frequency;   In the photosensitive material, established by the exposure dose used in the second exposure The amplitude of the intensity pattern;   A phase of the intensity pattern with respect to a reference point on the substrate,   The process according to claim 1, characterized in that: 13. The second exposure is performed using an image-forming interference exposure, and Process for providing a high spatial frequency component of the two-dimensional spatial pattern to a conductive material. And the magnitude of the spatial frequency of the high spatial frequency component is approximately NA₁/ Λ₁Than Large, the second exposure is   A center spatial frequency component at the center spatial frequency;   Of the central spatial frequency component established by the exposure used in the second exposure Amplitude and   A phase of the central spatial frequency component;   An adjustment is made to define the two-dimensional spatial pattern in the photosensitive material on the substrate. At least one of the circles centered on the central spatial frequency having adjusted amplitude and phase. A range of said high spatial frequency component having a spatial frequency within the part;   The process according to claim 1, characterized in that: 14． The imaging interference exposure is formed by an imaging interference optical system, the imaging interference exposure comprising: The optics system   Wavelength λ_TwoA spatially coherent illumination source,   Numerical aperture NA_TwoAnd the magnification M_TwoA second optical imaging system having:   Polar angles (-arc) with respect to a coordinate system fixed to the second imaging optical system sin [M_Twosin (θ_Two)], Φ_TwoMeans for attaching a second mask to   Providing the second mask with a substantially uniform plane wave at a polar angle with respect to the coordinate system; (-Arcsin [sin (θ_Three) / M_Two], Φ_TwoOptical means for irradiating   The polar angle (θ_Two, φ_TwoMeans for attaching said substrate to ,   Direct a reference plane wave through a third mask to define an exposure area on the substrate And the light produced by the illumination of the second mask and the second optical imaging system Coherently with the school, the polar angle (θ_Three, φ_Two) On the substrate To image the third mask on the substrate at the zero order spatial frequency of the incident image Optical means of   Adjusting the substrate position or relative optical path length, thereby adjusting the circumference of said first exposure. Matching to ensure an appropriate phase relationship between the wave number component and the frequency component of the second exposure Means,   Whereby the second exposure is performed on the substrate, the second exposure comprising:   φ_Two+ [Sin (θ in the direction represented by_Two) + Sin (θ_Three−θ_Two)] / Λ_Two And-[sin (θ_Two) + Sin (θ_Three−θ_Two)] / Λ_TwoTwo offset centers Spatial frequency,   In the spatial frequency space centered on each of the offset center spatial frequencies, Diameter NA_Two/ Λ_TwoA spatial frequency component within at least said portion of the circle of   Characterized in that the relative amplitude and the spatial frequency component within at least a part of the circle And phase substantially reproduce the relative amplitude and phase of the desired pattern on the substrate 14. The process according to claim 13, wherein 15. The angle θ_TwoIs θ_Three/ 2, which allows the A photosensitive material is symmetrically illuminated and the center line of the second optical imaging system is -Θ_Three/ 2, and the center line of the imaging optical means is + θ_Three/ 2, and the central spatial frequency is φ_TwoIn the direction represented by And +2 sin (θ_Three/ 2) / λ_TwoAnd -2 sin (θ_Three/ 2) / λ_Two The process of claim 14, wherein 16. The angle θ_TwoIs fixed to 0, whereby the second mask and And the substrate is perpendicular to a centerline of the second optical imaging system, and , The center spatial frequency is φ_TwoIn the direction represented by_Three) / Λ_TwoAnd -sin (θ_Three) / Λ_TwoThe process of claim 14, wherein 17． The imaging interference exposure is performed by an imaging interference optical system, and the imaging interference light The science system   Wavelength is λ_TwoAnd coherent is σ_TwoPartially spatially coherent illumination source When,   Numerical aperture NA_TwoAnd magnification M_TwoA second optical imaging system having   Polar angles (-arc) with respect to a coordinate system fixed to the second imaging optical system sin [M_Twosin (θ_Two)], Φ_TwoMeans for attaching a second mask to   For the coordinate system, (-arcsin [sin (θ_Three) / M_Two], Φ_Two) Inside Optical means for irradiating the second mask at a polar angle with respect to the center;   The polar angle (θ_Two, φ_TwoMeans for mounting the substrate on   Directing a reference wave through a third mask to define an exposure area on the substrate; Optical means for imaging said third mask on a substrate, said zero-order spatial frequency of said imaging being The number is the polar angle (θ_Three, φ_Two) Is incident on the substrate and the second Field and interference caused by the illumination of the mask and the second optical imaging system Optical means;   Adjusting the substrate position or the relative optical path length, thereby adjusting the frequency of the first exposure. A matching method that ensures an appropriate phase relationship between several components and the frequency components of the second exposure. Steps and   Whereby the second exposure is performed on the substrate, the second exposure comprising:   φ_Two+ [Sin (θ in the direction represented by_Two) + Sin (θ_Three−θ_Two)] / Λ_Two And-[sin (θ_Two) + Sin (θ_Three−θ_Two)] / Λ_TwoTwo offset centers Spatial frequency,   In the spatial frequency space centered on each of the offset center spatial frequencies, Diameter (1 + σ_Two) NA_Two/ Λ_TwoA spatial frequency component within at least a portion of the circle of Characterized by   Relative amplitude and phase of a spatial frequency component within the at least the portion of the circle Provide a substantial reproduction of the relative amplitude and phase of the desired pattern on the substrate. The process of claim 13. 18. The angle θ_TwoIs θ_Three/ 2, which allows the A photosensitive material is symmetrically illuminated and the center line of the second optical imaging system is -Θ_Three/ 2, and the center line of the imaging optical means is + θ_Three/ 2, and the central spatial frequency is φ_TwoIn the direction represented by And +2 sin (θ_Three/ 2) / λ_TwoAnd -2 sin (θ_Three/ 2) / λ_Two The process of claim 17, wherein 19. The angle θ_TwoIs fixed to 0, whereby the second mask and And the substrate is perpendicular to a centerline of the second optical imaging system, and , The central spatial frequency is φ_TwoIn the direction represented by_Three ) / Λ_TwoAnd -sin (θ_Three) / Λ_TwoThe process of claim 17, wherein 20. The second mask and the third mask have substantially the same spatial pattern. Further comprising the second optical imaging system and the optical means are substantially equal. Providing a numerical aperture, the two-dimensional pattern on the photosensitive material on the substrate, Spatial frequency 2 sin (θ_Two/ 2) the mask convolved with the high spatial frequency component of The process according to claim 14, characterized by a reduced spatial pattern of: 21. The second mask and the third mask have substantially the same spatial pattern. Further comprising the second optical imaging system and the optical means are substantially equal. Providing a numerical aperture, the two-dimensional pattern on the photosensitive material on the substrate, Spatial frequency 2 sin (θ_Two/ 2) the mask convolved with the high spatial frequency component of 18. The process according to claim 17, characterized by a reduced spatial pattern of: 22. Amplitude and position of the spatial frequency component caused by the illumination of the mask A phase is not collected by the second optical system and the optical means. There are no restrictions on the amplitude and phase of the spatial frequency component generated in the wave number space domain. In addition, the amplitude and phase desired for the final imaging in the frequency space domain are substantially the same. As described above, the case where the second mask and the third mask are configured 15. The process of claim 14, further comprising: 23. Amplitude and position of the spatial frequency component caused by the illumination of the mask A phase is not collected by the second optical system and the optical means. No restrictions on the amplitude and phase of spatial frequency components generated in the wavenumber space domain Substantially the same as the amplitude and phase desired for final imaging in the frequency space domain. In the case where the second mask and the third mask are configured, 18. The process of claim 17, further comprising: 24. Amplitude and position of the spatial frequency component caused by the illumination of the mask The phase is substantially the same as the amplitude and phase desired for final imaging, but the spatial frequency The second mask and the third mask are configured to be shifted with respect to , The optical imaging system may include the spatial frequency within a desired frequency spatial region on the substrate. Adjusted to print the wave number, to spatial frequency components that are not in the frequency spatial domain 15. The process of claim 14, further comprising the case where there are no restrictions. 25. Amplitude and position of the spatial frequency component caused by the illumination of the mask The phase is substantially the same as the amplitude and phase desired for final imaging, but the spatial frequency The second mask and the third mask are configured to be shifted with respect to , The optical imaging system may include the spatial frequency within a desired frequency spatial region on the substrate. Adjusted to print the wave number, to spatial frequency components that are not in the frequency spatial domain 18. The process of claim 17, further comprising the case where there are no restrictions. 26. Complete mask and coherently illuminated imaging interference optical system A process of optically generating the mask using   The desired frequency space region is at least partially located in the plane of the substrate. Process on a blank mask blank coated with photoresist And   Thereafter, the mask blank is processed, thereby developing the photoresist. Transferring the optically generated pattern to the mask;   14. The method of claim 13, further comprising an optical generation step comprising the process of: On-boarding process. 27. The two-dimensional pattern of the second mask is mainly arranged in orthogonal directions x and y. Including a structure with oriented edges, thereby providing a small Said polar angle φ to encompass the high spatial frequency caused by the dimensional structure_TwoIs the x direction In the third exposure, and in the third exposure to the corresponding y-axis. Φ to encompass the high spatial frequency caused by the small structures of small dimensions in_Two Are aligned with the orthogonal y-axis, which allows the 15. The process of claim 14, wherein the inter-frequency coverage is maximized. 28. The two-dimensional pattern of the second mask is mainly arranged in orthogonal directions x and y. Including a structure with oriented edges, thereby providing a small Said polar angle φ to encompass the high spatial frequency caused by the dimensional structure_TwoIs the x direction In the third exposure, and in the third exposure to the corresponding y-axis. Φ to encompass the high spatial frequency caused by the small structures of small dimensions in_Two Are aligned with the orthogonal y-axis, which allows the The process of claim 17, wherein the inter-frequency coverage is maximized. 29. The height caused by small structures with small dimensions along the corresponding x-space axis The offset center spatial frequency of the second exposure to include Matching the numbers along the x frequency axis;   High sky created by small structures with small dimensions in the corresponding y-space axis Intersect the center spatial frequency of the third exposure so as to include the inter-frequency Aligned along the frequency axis, thereby providing spatial Maximizing the frequency coverage; and   23. The process of claim 22, further comprising the process of: 30. The height caused by small structures with small dimensions along the corresponding x-space axis The offset center spatial frequency of the second exposure to include Matching the numbers along the x frequency axis;   High sky created by small structures with small dimensions in the corresponding y-space axis Intersect the center spatial frequency of the third exposure so as to include the inter-frequency Aligned along the frequency axis, thereby providing spatial Maximizing the frequency coverage; and   24. The process of claim 23, further comprising: 31. The height caused by small structures with small dimensions along the corresponding x-space axis The offset center spatial frequency of the second exposure to include Matching the numbers along the x frequency axis;   High sky created by small structures with small dimensions in the corresponding y-space axis The offset center spatial frequency of the third exposure to include the inter-frequency. Aligned along the orthogonal y-frequency axis, thereby creating a linear pattern of spatial patterns Maximizing the spatial frequency coverage of the   25. The process of claim 24, further comprising: 32. The height caused by small structures with small dimensions along the corresponding x-space axis The offset center spatial frequency of the second exposure to include Matching the numbers along the x frequency axis;   High sky created by small structures with small dimensions in the corresponding y-space axis The offset center spatial frequency of the third exposure to include the inter-frequency. Aligned along the orthogonal y-frequency axis, which creates a spatial pattern of linear Maximizing the spatial frequency coverage of the   26. The process of claim 25, further comprising the process of: 33. The first exposure and the second exposure from a single coherent radiation source The process of claim 1 wherein withdrawing.