JP2004138620A - Pattern inspection method - Google Patents

Abstract

【Task】
Provided is an exposure apparatus that forms a pattern on a mask on a substrate at a resolution higher than the resolution of an optical system using a black-and-white mask, and a method thereof.
[Solution]
An exposure apparatus, light source means for emitting a laser, scanning means for scanning the laser light emitted from the light source means, irradiation means for irradiating the mask with the laser scanned by the scanning means, and laser irradiation by the irradiation means An imaging optical system means for forming an image of the light from the mask irradiated on the wafer, and a table means for mounting the wafer, the laser emitted from the light source using a galvanometer mirror The laser is scanned, the mask is irradiated with the scanned laser, and light from the mask irradiated with the laser is focused on a wafer through an imaging optical system to expose the wafer.
[Selection] Fig. 8

Description

The present invention eliminates the influence of interference light generated in an extremely fine circuit pattern formed on a mask, and forms an image with high resolution on a substrate through a projection lens using an excimer laser beam or the like for exposure. The present invention relates to an excimer exposure method and apparatus, an excimer exposure method, and a mask circuit pattern inspection method.

In LSI manufacturing, a circuit pattern on a mask is exposed and transferred onto a wafer to form a fine circuit pattern on the wafer. However, in order to meet the needs for high integration of LSIs, circuit patterns transferred onto wafers have become extremely fine and have reached the resolution limit of the imaging optical system.

Therefore, various techniques have been developed to transfer an extremely fine circuit pattern.

For example, there is a method of exposing using X-rays such as SOR (Synchrotron Organized Resonance) light.

方法 There is also a method using EB (electron beam, electron beam exposure machine).

In addition, since the exposure throughput is fast and the handling is relatively easy, "Excimer Laser Stepper for Sub-Half-micron Lithography, Akikazu Tanimoto, SPII No. 1088 Optical Laser Microlithography 2 (1989)""Excimer Laser Stepper for Sub-half
Micron Lithography, Akikazu Tanimoto, SPIE Vol. 1088 Optical Laser
Microlithography 2 (1989) "or a method using an excimer laser disclosed in JP-A-57-198631.

位相 Furthermore, Japanese Patent Publication No. Sho 62-50811 discloses a phase shifter method for improving the resolution by devising a mask. This phase shifter method increases the resolution by interfering light from an adjacent pattern. By providing a film (phase shifter) whose phase is alternately shifted by π so that the phase of an adjacent pattern is inverted. Realize.

References introducing theoretical analysis of partial coherent imaging include “Optics of Stepper (1), (2), (3), (4)” (Optical Technology Contact, Vol. 27, No. 12, pp. 76
2-771, Vol. 28, No. 1, pp. 59-67, Vol. 28, No. 2. pp. 108-119, Vol. 28, No. 3,
pp. 165-175).

例 Further, an example in which the resolution is improved by using a spatial filter is described in JP-A-3-27516.

JP-A-57-198631

"Excimer Laser Stepper for Sub-Half Micron Lithography, Akikazu Tanimoto, SPII No. 1088 Optical Laser Microlithography 2 (1989)", "Excimer Laser Stepper for Foresight, Sub-half, Russia, Toronto, Canada, London, Canada, Canada, USA. 2 (1989) " "Stepper Optics (1), (2), (3), (4)" (Optical Technology Contact, Vol. 27, No. 12, pp. 762-771, Vol. 28, No. 1, pp. 59) -67, Vol.28, No.2, pp.108-119, @ Vol.28, No.3, pp.165-175)

The prior art disclosed in Japanese Patent Publication No. Sho 62-50811 has a problem that it is difficult to arrange a phase shifter and it is difficult to manufacture a mask provided with the phase shifter.

An object of the present invention is to solve the above-mentioned problems of the prior art and expose an excimer or the like which can transfer a black and white ultrafine circuit pattern formed on a mask onto a substrate with a resolution equal to or higher than that of a phase shifter. It is an object of the present invention to provide a method and an apparatus therefor.

Another object of the present invention is to provide an exposure method such as an excimer, which can simulate and confirm data of a transfer pattern onto a substrate which is actually transferred by an exposure apparatus by a calculation process.

Further, an object of the present invention is to enable a very fine circuit pattern formed on a mask to be inspected with high accuracy even when a very fine circuit pattern formed on a mask is different from a pattern transferred on a substrate. It is another object of the present invention to provide a mask circuit pattern inspection method.

The present invention provides an illumination unit that emits light such as an excimer laser beam that maintains coherence to some extent, an imaging unit that forms an image on a wafer by illuminating a mask (including a reticle) with the illumination unit, and a mask. An exposure apparatus or an exposure method, such as an excimer, provided with a light shielding unit that shields at least a part of the zero-order diffracted light among the transmitted or reflected light, can solve the above-described problem.

The present invention is also an exposure apparatus such as an excimer, wherein the spatial filter is formed of a device that shields a region corresponding to NA of the illumination unit from light.

The present invention is also an exposure apparatus such as an excimer, which has an integrator and a spatial filter as the illumination means.

The present invention is also an excimer exposure apparatus, wherein the mask has a circuit pattern formed with a line width of approximately 1/2 of an image resolution.

Further, the present invention provides an illumination unit that applies annular diffused illumination formed from a number of virtual point light sources to a mask almost uniformly in an exposure region, and a diffuse illumination that is almost uniformly provided by the illumination unit. An optical pupil that blocks at least a part of the zero-order diffracted light or the low-order diffracted light of the light transmitted through the mask, and forms an image of a circuit pattern formed on the mask on the substrate in the exposure region. And a method of step-and-repeat and sequentially removing a circuit pattern formed on a mask on a substrate.

Further, the present invention illuminates a mask on which a circuit pattern is formed, and forms an image by an imaging unit by shielding at least a part of the 0th-order diffracted light from light transmitted or reflected by the illuminated mask circuit pattern. And exposing it to a substrate.

The present invention is also an exposure method for excimer or the like, wherein the minimum line width of the circuit pattern on the mask is formed in conformity with the imaging resolution of the imaging means.

The present invention is also an excimer or other exposure method, characterized in that the mask has a circuit pattern formed with a line width of approximately の of the imaging resolution.

Further, according to the present invention, the circuit pattern on the mask is formed with a line width of about 1/2 of the imaging resolution, and for a wide circuit pattern transferred onto the substrate, a transmission portion is formed on the mask. An excimer or other exposure method characterized by being formed in a line-and-space or lattice pattern with a pitch of about 1/2 to 1/3 of the image resolution.

The present invention also relates to an excimer or other exposure method, which comprises dividing a circuit pattern formed on a mask into a fine pattern portion and a large pattern portion, forming an image by an imaging unit, and transferring the image onto a substrate. is there.

Further, the present invention provides a mask data converting means for converting and generating mask data from wiring data, and using excimer laser light for a circuit pattern formed on a mask for mask data obtained from the mask data converting means. Calculation means for performing calculation processing based on a transfer function substantially equivalent to imaging means for transferring the image onto the substrate and calculating data of a transfer pattern onto the substrate. is there.

Further, the present invention provides a mask data converting means for converting and generating mask data from wiring data, and a circuit pattern formed on a mask, such as an excimer laser beam, for mask data obtained from the mask data converting means. Calculation means for performing arithmetic processing based on a transfer function substantially equivalent to imaging means for transferring light onto a substrate using light to calculate data of a transfer pattern onto a wafer, and illumination for illuminating a mask with excimer laser light or the like Means for imaging light transmitted or reflected by the mask illuminated by the illuminating means at a detection position and having a transfer function substantially equivalent to that of the imaging means, and an image formed at the detection position. An inspection apparatus having a light receiving means for receiving an image circuit pattern to obtain an image signal; an image signal obtained from the light receiving means of the inspection apparatus; A mask circuit pattern inspection system further comprising a comparison means for comparing the the data transfer pattern onto the wafer.

(4) In an exposure apparatus using light such as excimer laser light, the cause of lowering the contrast of a circuit pattern transferred onto a substrate is that diffracted light cannot be sufficiently captured in the aperture (pupil) of the imaging means. Meanwhile, light is diffracted from the circuit pattern on the mask according to the wavelength used and the dimensions of the circuit pattern. At this time, when the circuit pattern becomes extremely fine with respect to the exposure wavelength, the diffraction angle increases, and the intensity of the diffracted light also increases. As a result, light does not enter the opening of the imaging means (projection lens) used for transfer, which causes a reduction in resolution.

Therefore, in order to reduce the diffracted light as much as possible, as in an excimer laser stepper, to reduce the diffracted component, or to increase the NA of the image forming means (projection lens) so as not to miss the diffracted light as much as possible. There is a thing to take in a lot.

On the other hand, according to the present invention, while the diffracted light from the circuit pattern on the mask has a small amount of components captured by the imaging means (projection lens), the component not diffracted from the circuit pattern on the mask (zero-order diffracted light) Of the light required for image formation, only the 0th-order diffracted light is largely taken into the image forming means (projection lens), and is relatively diffracted by the image forming means (projection lens). Focusing on the fact that a small amount of light component is taken in, at least a part of the 0th-order diffracted light is shielded so that the balance between the amounts of the diffracted light and the 0th-order diffracted light emitted from the imaging means (projection lens) is relatively controlled. In order to realize high-resolution exposure, the contrast in which an extremely fine circuit pattern formed on a mask is image-transferred onto a substrate through an imaging means (projection lens) is improved. It is intended.

Further, in order to efficiently realize the shielding of the 0th-order diffracted light, it is necessary to increase the coherency of the illumination light of the mask. Further, in the projection exposure apparatus, the spatial coherence degree (sigma, σ) of the illumination system is increased in order to provide the remodeling ability to a high spatial frequency. These two conditions seemingly contradict each other, but can be simultaneously achieved by using an annular illumination light source used in a phase contrast microscope or the like.

Also, since the annular light source can increase the coherence, the depth of focus of the optical system can be increased.

In particular, the present invention provides an illumination unit that applies an annular diffused illumination formed from a large number of virtual point light sources to a mask in an exposure region almost uniformly, and the illumination unit performs a diffuse illumination in a substantially uniform manner. Has an optical pupil that blocks at least a part of the zero-order diffracted light or the low-order diffracted light of the light transmitted through the mask, and forms a circuit pattern formed on the mask on the substrate in the exposure region on the substrate. With the provision of the reduction projection lens, it is possible to improve the contrast in which an extremely fine circuit pattern formed on the mask is image-transferred onto the substrate through the reduction projection lens, thereby realizing high-resolution exposure.

It is apparent that the present invention is not necessarily limited to the projection type exposure method using excimer laser light.

The present invention can avoid imbalance between the 0th-order diffracted light and the light intensity of the diffracted light when transferring the mask pattern. Therefore, the conventional black-and-white mask pattern can be used with a resolution equal to or higher than that using a phase shifter. It has the effect that it can be exposed.

First, the principle of the present invention will be described with reference to FIGS.
That is, in the present invention, a circuit pattern on a mask is transferred with an improved contrast rather than faithfully transferred onto a substrate by an imaging means (projection lens). That is, in the projection exposure, it is not always necessary to “transfer the circuit pattern accurately”, but it is based on a new technical idea that “the circuit pattern desired to be obtained on the substrate (wafer) should be transferred with high contrast”.

FIG. 1A is a sectional view of a mask 100 in which a mask circuit pattern 104 is formed on a glass substrate 101 by chrome 102. A waveform 301 in FIG. 1B is a strong signal intensity distribution of an image forming pattern formed on the substrate (wafer) 200 by the projection exposure apparatus 3000 shown in FIGS. This waveform 301 is considered separately into a waveform 302 due to the zero-order diffracted light and a waveform 303 due to the high-order diffracted light. When the mask circuit pattern 104 is a fine circuit pattern 105 as shown in FIG. 1A, the waveform 301 detected by the higher-order diffracted light is smaller than the waveform 302 generated by the 0th-order diffracted light. AV becomes smaller. Here, since the component of the waveform 302 can be removed by blocking the 0th-order diffracted light, the detected waveform has a high contrast like the waveform 302.

Further, another principle of the present invention will be described below. That is, in the present invention, the phase of an adjacent circuit pattern on the mask is intended to be inverted (that is, to be shifted by π) without using a phase film. As shown in FIGS. 2 and 3, when the light-shielding portion between the adjacent circuit patterns 321 (A) and 323 (B) of the mask (reticle) 100 is narrow, the light-shielding portion becomes a completely complementary figure. If the parts have a finite width, they will be approximately complementary. Circuit patterns 321 (A) and 323 (B) adjacent to the mask (reticle) 100 are applied to a diffraction image plane 3203 by an imaging optical system (projection lens) 3201 on a Fraunhofer diffraction image according to the principle of Babinet. In this case, the light intensity is equal and the phase is shifted by π except for the central point (0-order diffracted light). In this diffraction image plane 3203, in the diffraction patterns other than the zero-order diffraction light, the light from the adjacent circuit patterns 321 (A) and 323 (B) is inverted in phase (shifted by π), and the diffraction pattern (diffraction image plane) ) Above, the light shielding plate 324 (image formation spatial filter 3302) shields at least a part of the 0th-order diffracted light, so that the light imaged on the substrate 200 surface (image formation surface) is inverted in phase. This is equivalent to the image of light from an adjacent pattern (shifted by π), and an ultra-fine circuit pattern of high contrast on a substrate 200 (an ultra-fine circuit pattern of about 0.1 μm or less on a wafer). Circuit pattern) can be transferred. That is, as shown in FIG. 2 or FIG. 3, at least a part of the zero-order diffracted light is shielded by the light shielding plate 324 (imaging spatial filter 3302) on the diffracted image plane, so that only the diffracted light whose phase is inverted is formed on the image plane. From the imaging plane, it looks as if a phase film is formed on the mask. As a result, the same circuit pattern as in the phase shifter method is formed on the wafer, and the intensity distribution on the image forming surface is higher on the wafer 200 than in the case of the conventional reduced projection exposure shown in FIG. Is obtained. FIG. 7 is a diagram for explaining the case of the conventional reduced projection exposure. The images of the adjacent fine circuit patterns 321 and 323 have low contrast on the image forming surface. In short, in the case of the reduced projection exposure according to the present invention, as shown in FIGS. 2 and 3, an extremely fine circuit pattern having a higher contrast on the imaging surface as compared with the case of the conventional reduced projection exposure shown in FIG. (An extremely fine circuit pattern of about 0.1 μm or less on the wafer) is transferred and exposed.

Hereinafter, the example shown in FIG. 4 will be described using mathematical expressions. In the example of FIG. 4, the light emitted from the mercury lamp 3101 is condensed on the light source spatial filter 3301 by the condenser lens 3103, and the mask 100 is illuminated by the condenser lens 3106.
A part of the light transmitted through the mask is shielded by the imaging spatial filter 3302, and is imaged on the wafer 200 by the imaging lens 3201. Assuming that the shape of the light source spatial filter 3301 is l (u, v), the shape of the pattern on the mask 100 is f (x, y), and the shape of the imaging spatial filter 3302 is a (u, v), The image intensity gp (x, y) is calculated by the following (Equation 1).

In the expression (1), since the lights emitted from the respective (u, v) on the light source spatial filter 3103 do not interfere with each other, the light is integrated after calculating the intensity on the imaging plane. Here, according to Kubota, Wave Optics (Iwanami Shoten), generally, the resolution of an optical system can be considered using the response function of the optical system or the optical transfer function (OTF, Optical Transfer Function). The response function H (u, v) in the example of FIG. 4 is expressed by the following equation (2) using the pattern f (x, y) on the object surface and the intensity gp (x, y) of the image. Is calculated.

FIG. 32 shows a calculated response function of the present optical system in a curve 351. The horizontal axis indicates the spatial frequency s, and for reference, the corresponding numerical aperture of the imaging lens (when NA = 0.38). The vertical axis indicates the response function normalized by the zero-order component. Here, the position of the point 355 indicates the size of the aperture of the imaging lens. A curve 352 indicates a response function when the conventional optical system, that is, the light source spatial filter 3301 and the imaging spatial filter 3302 are not used, a curve 353 indicates an apparent response function by the phase shift method, and a curve 354 indicates coherent light such as a laser. The response function when using light is shown for reference. The response function 352 of the conventional method extends to a position s1 indicated by the following equation (3).

The method according to the present invention is the same as the conventional method in that the response function extends to the position shown in the above equation (Equation 3), but the curve has a stable shape from the vicinity of NA 0.2 to the position of 0.6. Has become. Further, in the present invention, the response function of the low-frequency component is reduced by devising the shape of the mask pattern as described later, so that the response function of the entire system of the present invention has a curve 356. A curve 357 is shown by being normalized, and has a stable response function over a wide band from low frequency components to high frequency components. This indicates that a fine pattern can be formed with high contrast according to the present invention. Specifically, for example, a line and space of 0.3 μm is located at a point 358, and the contrast is C1 in the conventional method, but is as high as C2 in the present invention. In the present invention, the shapes of the light source spatial filter 3301 and the imaging spatial filter 3302 are determined so as to optimize the response function calculated using the above (Equation 1). As described above, according to the present invention, the value of the response function of the high frequency component can be relatively increased by reducing the response function of the low frequency component.
Further, by using the annular light source spatial filter, the response function can be extended to the band according to the above equation (3). Both the size and the width of the light source spatial filter-3301 and the imaging spatial filter-3302 can be optimized by calculating a response function using the above-described equation (2).

A method of calculating the image formation state on the wafer 200 by simulation and confirming the mask shape will be described later. However, since the equation (1) cannot be solved analytically, the equation (1) is based on the equation (1). Calculated by numerical calculation. In addition, when a response function that can be calculated by the above equation (2) is obtained, and the response function is calculated by the following equation (4), the calculation time can be reduced.

33 to 38, a method for determining the shapes of the light source spatial filter 3301 and the imaging spatial filter 3302 will be described. The optical system of the present invention is a so-called partial coherence imaging optical system, and cannot be sufficiently explained by a so-called response function. The optical system of the partial coherence imaging is described in “Optics of Stepper” (Optical Technology Contact, Vol. 27, No. 12, pp. 762-771). Using this concept, the imaging characteristics of the optical system using the annular light source and the annular spatial filter of the present invention are calculated.

The imaging characteristic of this partial coherence imaging is represented by the following (Equation 5) using the concept of Transmission Cross-Coefficient, T (x1, x2) indicating the relationship between the shape of the light source and the shape of the pupil plane of the detection optical system. It is calculated by the formula. Further, according to the above-mentioned "optical system of the stepper", the imaging characteristic (OTF, Optical Transfer Function) of this optical system is approximately determined by the lowest-order Transmission Cross-Coefficient, T (x, 0). T (x, 0) is represented by a correlation function between the shape of the light source and the shape of the pupil plane.

That is, the characteristic of the partially coherent imaging represented by the complicated equation (5) is a geometrical problem of a correlation function between the shape of the light source and the shape of the pupil plane. FIG. 33 shows a light transmitting portion 3305 of the light source spatial filter 3301 and a light shielding portion 3306 of the imaging spatial filter 3302. The correlation function between the light source and the pupil plane at the coordinate x is indicated by the area of the hatched portion 364 in FIG. Similarly, the correlation function between the light source and the pupil plane according to the related art is shown by a hatched portion 365 in FIG.

The curve 367 in FIG. 37 shows the correlation function in the case of FIG. 33, in other words, Transmission.
Shows the calculated value of Cross-Coefficient, T (x, 0). The value is larger in the high frequency region than the correlation function and the curve 366 in the conventional case shown in FIG. That is, the contrast increases. This FIG. A. = 0.38, σ = 0.9. Further, the horizontal axis of FIG. 37 indicates that each N.D. A. Are shown as the minimum pattern dimensions. (For example,
0.3 means line and space 0.3 microns. It can be seen that the 0.3 micron OTF is about twice that of the conventional example.

FIGS. 36 and 37 are diagrams for deepening intuitive understanding of the shape setting of the light source spatial filter 3301 and the imaging spatial filter 3302, and assisting the setting. The shaded area A in FIG. 35A shows a correlation function between the outer diameter of the light source and the light shield 3306, and the shaded area B in FIG. 35B shows a correlation function between the inner diameter of the light source and the light shield 3306. Further, a hatched portion C in FIG. 35C indicates a correlation function between the transmission portion 3305 of the light source and the maximum pupil of the optical system. The final correlation function between the shape of the light source and the spatial filter is indicated by a hatched portion 367 in FIG. 36, which is obtained by the above CA-B. By determining the correlation function in this manner, the effect of the spatial filter or the annular illumination can be intuitively understood, and on the contrary, it is helpful in determining these shapes. Specifically, the value of the high frequency region 382 becomes larger than that of the medium frequency region 381 due to the annular illumination. Further, while the low-frequency region 383 is too large, the effect of the spatial filter A and B is printed on the sea route to further reduce the value in the low-frequency region. Thus, the effects of the annular illumination and the spatial filter were intuitively described.

Of course, the shape of the light source and the shape of the spatial filter should be evaluated based on the equation (5), and approximately should be evaluated by the correlation function between the light source and the spatial filter.

輪 Moreover, since the annular light source can increase the coherence, the depth of the optical system can be increased. Here, the narrower the band width of the annular light source, the higher the coherence of the light source, so the depth of focus increases, and the larger the diameter of the annular light source of the annular light source, the higher the degree of spatial coherence, and the higher the resolution. .

FIG. 39 illustrates the effects of the annular light source and the annular spatial filter used in the present invention. FIG. 39A shows a pupil 3301 of the imaging lens, a light source image 3305a (zero-order diffracted light) formed on the pupil, and a light source formed by a pattern (circuit pattern) formed on the mask 100 in the y direction. 3305b and 3305c are shown.
FIG. 39A shows a case where the light source has a ring shape, and FIG. 39B shows a case where the light source is circular.
A filter for blocking the 0th-order diffracted light is indicated by a hatched portion 371. The hatched portion 371 also blocks light 372 that is part of the diffracted light 3305b and 3305c of the light source at the same time. In the case of (a), only 10% to 20% is blocked, but (b) In the case of, 40% or more is shielded from light. That is, the purpose of efficiently shielding only the 0th-order diffracted light is more effectively achieved by the annular light source shown in FIG. That is, the performance of the annular light source is higher. Here, the smaller the width of the annular light source, the smaller the ratio of the diffracted light that is blocked.

Further, in the present invention, a spatial filter 3306 having a ring width of about 30% as narrow as the ring width of the light source as shown in FIG. However, since it is only necessary to block a part of the zero-order diffracted light, FIG.
As shown in (b), the width of the annular zone is almost the same as that of the light source, and a spatial filter having a transmittance of about 70% may be used. Of course, the ring width may be smaller than the ring width of the image of the light source, and the transmittance may be lower than 70%. Further, although the case where the light-shielding ratio of the 0th-order diffracted light is set to approximately 30% is shown here, the light-shielding ratio is not limited to 30% as described later. Further, in order to block a part of the zero-order diffracted light, a phase plate may be used in which the spatial filter 3306 has a transmittance of 100% and shifts the phase at the used wavelength by π. Such a filter is also a filter that substantially blocks a part of the zero-order diffracted light. Further, light may be shielded by using polarized light and a polarizing plate.

Further, as shown in FIG. 50, the annular width of the filter having the transmittance of about 70% may be larger than the annular width of the light source. With such a configuration, not only the zero-order diffracted light but also a part of the low-order diffracted light can be shielded, and the MTF curve can be further improved. Also, the example shown in FIG. 40A is configured to block a part of the low-order diffracted light as in the example of FIG.

As described above, in order to efficiently block the 0th-order diffracted light, a ring-shaped light source is effective, and the effect increases when the width of the ring is reduced. Here, the annular light source can be considered as small light sources arranged in an annular shape. That is, it can be considered as an aggregate of coherent point light sources. Therefore, as shown in FIG. 41, a light-shielding plate 376 smaller than the light source 375 is installed at a position corresponding to the set of the light sources 375 having a spatial coherence degree of about 0.1 to 0.3 similar to a point light source. Even so, the object of the present invention is achieved. It goes without saying that a light-shielding plate 376 having a reduced transmittance may be used. FIG. 41A shows an example in which the elements are arranged in a ring shape. Further, if this concept is advanced, the shielding of the 0th-order diffracted light can be achieved even if it does not have an annular shape as shown in FIG. At the same time, as shown in FIG. 41 (c), such a light source and a light-shielding plate may be arranged as a plurality of annular zones. Further, as shown in FIG. 41D, in order to block a part of the zero-order diffracted light, a light-shielding plate may be provided only for a part of the corresponding light source.

42A and 42B show the light intensity distribution of the light source and the transmittance of the spatial filter in the radial direction, respectively. Here, FIG. 42 shows a rectangular distribution for both the light intensity distribution and the transmittance distribution. However, as shown in FIG. 43, there is no problem if a gentle distribution is shown. This can be understood by considering that the OTF is represented by these correlation functions. That is, since the correlation function is obtained by integrating weights with respect to the overlapped portion, the value of the correlation function does not change significantly even if a gentle distribution is shown. In each case, the distributions are equal in the radial direction and concentric distributions, and this is important.

The fact that the light source described here may have a distribution as shown in FIG. 43 (a) means that the intensity of the light source shown in FIGS. 41 (b), (c) and (d) decreases as the center increases. Indicates that the shape may be used. Such an embodiment has the effect that a light source having less coherency can be produced and at the same time the low frequency component can be further reduced.

As described above, in the present invention, to effectively block a part of the 0th-order diffracted light with an exposure apparatus or another image forming optical system is a means for solving the objects of improving resolution and depth of focus. However, in order to efficiently shield the 0th-order diffracted light, the 0th-order diffracted light and the diffracted light need to be separated and not overlapped on the Fourier transform plane where the spatial filter is arranged. Coherency must be high. That is, it is desirable to be close to a point light source. On the other hand, in order to improve the resolution of the imaging optical system, it is desirable to increase the spatial coherence degree of the light source, that is, the σ value. That is, a large light source is desirable. In other words, it is necessary to balance the contradiction of a point light source and a large light source.
It is the annular light source and the spatial filter of the present invention that make these conflicts compatible.
In order to satisfy this efficiently, it is one measure to use an aggregate of small light sources. Further, if this assembly is arranged in a large annular shape, the condition of a large light source is satisfied. That is, if the width of the orbicular zone is reduced, coherency is increased and the depth of focus is improved, and if the radius of the orbicular zone is increased, the degree of spatial coherence is increased and the resolution is improved.

Therefore, if the radius of the orbicular zone is increased while blocking a part of the above zero-order diffracted light, there is a condition that the diameter of the spatial filter 3306 becomes as large as the pupil of the optical system. This condition is a condition in which a part of the zero-order diffracted light is shielded and the size of the annular light source is maximized. That is, the condition for obtaining the maximum resolution is obtained for a certain reduction projection lens. FIGS. 44 (a), (b) and (c) show this embodiment. In each case, the size of the light source is larger than the lens pupil. In general, it has been considered that when the size of the light source is increased, the depth of focus becomes shallower and cannot be used for lithography. However, since the depth of focus can be increased by using the annular light source as described above, a light source larger than the pupil as shown in FIG. 44 can be used. Each of the embodiments also has a configuration in which a part 377 of the zero-order diffracted light is shielded. The OTF having this configuration is also represented by a correlation function. Therefore, there is an effect that the cutoff frequency of the OTF is longer than when the size of the light source is smaller than the pupil. Another advantage of this configuration is that a large projection lens is used without using a reduction projection lens that requires high accuracy. A. There is a point that the resolution can be improved only by the improvement of the illumination system which is easy to realize. In this embodiment, the N.N. A. A pattern of approximately 0.2 μm can be transferred by i-line using a 0.4 lens.

What is important in these 0-order diffracted light shielding image forming systems is that the OTF curve gradually decreases monotonically. FIG. 45 shows an OTF 378 of the present invention. Here, when the OTF is not gentle and undulates like 379, the contrast becomes low near the minimum point of the wave, and the pattern is not correctly transferred in an actual LSI pattern having various spatial frequency components. . However, a special pattern formed only of a specific spatial frequency component is not limited to this, and the contrast may be increased only for a specific spatial frequency. That is, the MTF curve needs to be within the range of the specific width Wb. This Wb is to be calculated from a transfer result calculated by a transfer simulator described later.

Therefore, even when the light source shown in FIG. 44 is large, the OTF that gradually decreases monotonously is required. From this, when an actual LSI pattern is transferred, as shown in FIG. 44, the difference between the inner diameter of the light source and the pupil diameter of the reduction projection lens and the difference between the outer diameter of the light source and the pupil diameter of the reduction projection lens are almost equal. Desirably equal. In order to obtain a realistic depth of focus, it is desirable that the ratio between the pupil diameter and the inner diameter of the reduction projection lens be 0.6 or more.

In the example of FIG. 44A, a filter having an appropriate transmittance may be disposed in a portion 380 corresponding to the zero-order diffracted light of the light source in the pupil of the reduction projection lens. In this case, the width of the portion 377 outside the light source pupil can be reduced, and thus the size of the light source can be reduced. Conversely, the width of the portion 380 in the pupil of the reduction projection lens can be made larger than 377. In this case, effects such as easy maintenance of stable contrast in a wide band in which the light intensity is easy to increase are produced.

Further, the object of the present invention can be achieved to some extent even if the outer diameter of the light source of the annular zone is the same as the outer diameter of the pupil of the reduction projection lens. However, the configuration shown in FIG. 44A is advantageous in that the configuration is simple and easy to implement because a part of the zero-order diffracted light can be cut without inserting a spatial filter in the pupil of the reduction projection lens. Having.

Further, as shown in FIG. 47, the resolution is further improved by installing the annular light source 378 further outside the light source of FIG. FIG. 48 shows an example in which the light source shown in FIG. 47 is implemented. Thus, N. A. Since it is difficult to use a light source having a large value in a lens system, a laser light source 3121, beam scanning means 3122, and ring-shaped mirrors 3123 and 3124 are used. In the case of this embodiment, the N.I. A. 0
. 0.15 μm can be resolved using the lens No. 4.

The present embodiment also has no difference from the basic idea of the present invention in that a part of the zero-order diffracted light is shielded.

Next, an embodiment of a pattern transfer system (reduction projection exposure optical system) 3000 according to the present invention will be described with reference to FIGS. That is, in the pattern transfer system (reduction projection exposure optical system) 3000, of the light from the Hg lamp 3101, the i-line having a wavelength of 365 nm is selectively transmitted by the color filter 3102, and the surface of the integrator 3104 is condensed by the condenser lens 3103. Is collected. Light incident on each element 3107 (FIG. 9) in the integrator 3104 is individually emitted as an emission angle α, and illuminates the mask 100 with a condenser lens 3106. Here, the integrator 3104 will be described later. The light source spatial filter 3301 is formed in an annular shape in which a large number of virtual point light sources are arranged, and installed near the output end of the integrator 3104.

Then, a mask circuit pattern 104 on the mask 100 (for example, shown in FIG. 20).
) Has high contrast on the wafer (substrate) 200200 through an imaging lens (reduction projection lens) 3201 and an imaging spatial filter 3302 installed near the pupil of the imaging lens 3201. The image is transferred as a wafer circuit pattern 204 (for example, as shown in FIG. 18).

By the way, the image of the light source spatial filter 3301 formed in an annular shape in which a large number of virtual point light sources are arranged is formed in an annular space by the condenser lens 3106 and the imaging lens (reduction projection lens) 3201. The relationship forms an image at the position of the filter 3302. 13 and 14 show the image forming relationship of the light source spatial filter 3301 and the image forming spatial filter 3302 in this embodiment. Both the light source spatial filter 3301 and the imaging spatial filter 3302 have an annular shape.
The light source spatial filter 3301 forms the light source of the orbicular zone 3305 having the outer diameter DLO and the inner diameter DLI, and both the inside and the outside of the orbicular zone 3305 are shielded from light.
The imaging spatial filter 3302 has a structure in which an annular portion 3306 having an outer diameter DLO and an inner diameter DLI is shielded from light, and both the inside and the outside of the annular portion 3306 transmit light. Note that the entrance portion of the imaging lens (reduction projection lens) 3201 is indicated by 3205, and the exit portion is indicated by 3206.

Here, the imaging spatial filter 3302 is located at the front position 3202 of the imaging lens 3201, at the rear position 3204 of the imaging lens, and at the pupil position 3203 within the imaging lens. Is also good. The position 3203 has the best effect on the design, and the position 3204 has the lowest cost and the sufficient effect. FIG. 11 is a perspective view of an imaging lens 3201 having an imaging spatial filter 3302 mounted at a position 3204. Since the imaging spatial filter 3302 is formed of a metal plate, it is supported by the support bar 3311. It goes without saying that the smaller and thinner the support rod 3311 is, the better. FIG. 12 shows an example of an imaging spatial filter 3302 formed by forming a light-shielding film 3313 on a glass substrate 3312. In this case, the support rod 3311 is unnecessary, but it is necessary to design the imaging lens 3201 in consideration of the aberration due to the glass substrate 3312.

Here, assuming that the imaging magnification between the light source spatial filter 3301 and the imaging spatial filter 3302 is M, the outer diameter of the light source spatial filter 3301 and the outer diameter of the imaging spatial filter 3302 as shown in FIG. The relationship between DIO and DII which is the inner diameter of the imaging spatial filter 3302 will be described later. In short, if a part or all of the zero-order diffracted light is shielded, a fine circuit pattern on the mask can be formed on the wafer with high contrast.

As described above, in addition to the DLO and DLI of the light source spatial filter 3301, the DIO and DII of the imaging spatial filter 3302 may be configured as a variable spatial filter such as a liquid crystal display element or may be provided with a plurality of spatial filters having different dimensions. It is possible to control the annular dimension of the spatial filter by exchanging.

Next, an embodiment of the entire projection exposure system of the present invention will be described with reference to FIGS.

First, the pattern data generation system 1000 will be described.

In the pattern data generation system 1000, the wiring data creation unit 1102 forms wafer pattern shape data 1103 to be formed on the substrate (wafer) 200 based on wiring drawing data 1101 such as design data. The pattern conversion unit 1104 converts the wafer pattern shape data 1103 into mask pattern shape data 1105 to be formed on the mask (reticle) 100. At this time, the pattern transfer simulator 1108 determines that the mask pattern on the mask 100 converted by the pattern conversion unit 1104 is the same as the wafer pattern shape data 1103 formed by the wiring data creation unit 1102 and the annular zone 3305 of the light source spatial filter 3301. On the substrate 200 by the actual pattern transfer optical system 3000 based on the setting conditions of the outer diameter DLO, the inner diameter DLI, etc., and the setting conditions of the outer diameter DIO, the inner diameter DII, etc. of the annular portion 3306 of the imaging spatial filter 3302. It is checked whether or not the pattern substantially matches the wafer pattern shape data 1103 at the time of exposure, and is fed back to the pattern conversion unit 1104 to be corrected. Outer diameter DLO, inner diameter DLI, etc. of part 3305, The outer diameter DIO, inner diameter DII, etc. of the band 3306) are obtained, and the result is led to the light source spatial filter controller (adjuster) 3303 and the imaging spatial filter controller (adjuster) 3304 via the spatial filter control system 3305. , Control (adjust) the shapes of the light source spatial filter 3301 and the imaging spatial filter 3302. The pattern generation unit 1106 converts the mask pattern shape data 1105 converted by the pattern conversion unit 1104 into EB data 1107 suitable for the electron beam drawing apparatus 2103.

Next, the mask manufacturing system 2000 will be described. That is, in the mask manufacturing system 2000, the film forming apparatus 2101 forms the metal chromium, the chromium oxide, or the plurality of stacked films 202 of the metal chromium and the chromium oxide on the mask substrate 101.
The coating device 2101 applies the resist film 203 on the mask substrate 101 formed by the film forming device 2101. Then, the electron beam drawing apparatus 2103 draws and forms the same circuit pattern as the mask pattern shape data 1105 according to the EB data 1107 generated from the pattern generation unit 1106. Then, the circuit pattern on the mask substrate 201 is developed by the developing device 2104 to complete the mask 100. The completed mask 100 is subjected to pattern inspection by comparing the image data detected by the pattern inspection device 2106 with the mask pattern data 1103 or the wafer pattern data 1105 or the data from the transfer simulator 1108. If there is a defect, an ion beam processing device or the like is used. Is corrected by the pattern correction device 2105 composed of the above. Finally, the foreign material inspection device 2107 inspects the foreign material on the mask 100. If there is a foreign substance, it is cleaned by the cleaning device 2108. Here, since the mask 100 according to the present invention can be formed by a single layer film 102 as shown in FIG. 20, for example, it has a feature that it is cheaper to clean than a phase shifter mask. Further, the mask 100 can be manufactured more easily than a mask of a phase shifter. In the pattern inspection apparatus 2105, the detected image data may be compared with any of the mask pattern data 1103, the wafer pattern data 1105, and the data from the transfer simulator 1108. However, it is most effective to make the light source and the imaging optical system of the pattern inspection apparatus 2105 equivalent to the pattern transfer system (reduced projection exposure system) 3000 according to the present invention, and to compare and inspect the wafer pattern data 1105. That is, the pattern inspection apparatus 2105 is configured by an optical system equivalent to the pattern transfer system (reduction projection exposure system) 3000, the mask 100 to be inspected is installed on the mask stage 3401, and the wafer (substrate) 200 is installed. What is necessary is just to arrange | position a light receiving element in a position, and to detect the image formed on a light receiving element. By configuring the pattern inspection apparatus 2105 in this manner, the same circuit pattern as the very fine circuit pattern actually projected and exposed on the wafer (the very fine circuit pattern of about 0.1 μm or less on the wafer) is exposed to light. The image signal can be detected as a high-contrast image signal from the light-receiving element without being affected by the interference of light, and as a result, even a fine circuit pattern can be accurately inspected by comparison inspection with the wafer pattern data 1105. .

Next, a pattern transfer system (reduced projection exposure optical system) 3000 which is an important configuration of the present invention will be described. That is, in the pattern transfer system 3000, of the light from the Hg lamp 3101, the i-line having a wavelength of 365 nm is selectively transmitted by the color filter 3102, and is condensed on the surface of the integrator 3104 by the condenser lens 3103. Light incident on each element 3107 (FIG. 9) in the integrator 3104 is individually emitted as an emission angle α, and illuminates the mask 100 with a condenser lens 3106. Here, the configurations of different embodiments of the integrator 3104 are shown in FIGS. 9 and 10, respectively. FIG. 9 shows a case where the cross section of the integrator 3104 is annular. FIG. 10 shows an embodiment of an integrator 3104 in which an annular shape is formed by a light shielding plate 3105. In short, it is clear that other configurations may be used as long as they serve as a spatial filter, that is, as long as they have a light shielding function in an annular shape. The DLI and DLO of the light source spatial filter 3301 configured as described above may be configured as a variable spatial filter or may be controlled by providing a plurality of spatial filters of different dimensions and replacing them with each other. It is desirable that control or adjustment can be performed by a command from the spatial filter control unit (adjustment unit) 3303. Otherwise, you will be very restricted.

Here, in this embodiment, when a light-shielding plate is placed on the light source surface, the total exposure amount is reduced. Therefore, it is necessary to increase the light intensity of the light source. However, it has been difficult for conventional lamps to increase the light intensity. A strobe light source for fiber illumination is disclosed in "Applied Physics Society of Japan Fall Meeting 1991, 11p-ZH-8", Yamamoto et al., "Development research of a strobe light source for fiber illumination". A strobe light source as disclosed herein has not been used in an exposure apparatus. However, it is effective to use such a lamp in the present invention in which it is necessary to obtain a sufficient light intensity.
Furthermore, this light source is suitable for the present invention because of its large diameter.

In the embodiment shown in FIG. 44 of the present invention, an intagrator using an optical fiber as shown in FIG. 49 is effective. The integrator using this optical fiber is obtained by bundling a large number of optical fibers 380. The light introduction surface 3131 is circular so as to condense light from the light source 3101 and is cheap, and the light emission surface is a bundle of optical fibers re-bunched so as to have a ring shape. By using an optical fiber in this manner, a ring-shaped light source can be efficiently produced using a light source having a circular light source shape such as a mercury lamp. Further, since the integrator 3104 can be made flexible by using an optical fiber, there is an effect that the light source, which is a heat source, can be installed away from the apparatus main body that needs temperature control.

Here, it is preferable that the bundle of fibers shown in FIG. 49 be separated without being fixed, so that the outer diameter and the inner diameter can be changed. Such a variable mechanism is controlled by the light source spatial filter control mechanism 3303.

Then, the light transmitted and diffracted through the mask pattern 104 (FIG. 15) on the mask 100 passes through the imaging lens 3201 and the imaging spatial filter 3302 to be formed on the wafer 200 as the wafer pattern 204 (for example, shown in FIG. 18). Image.

Here, the imaging spatial filter 3302 is located at the front position 3202 of the imaging lens 3201, at the rear position 3204 of the imaging lens, and at the pupil position 3203 within the imaging lens. Is also good. The position 3203 has the best effect on the design and the position 3204 has the lowest cost and the sufficient effect. FIG. 11 is a perspective view of an imaging lens 3201 having an imaging spatial filter 3302 mounted at a position 3204. Since the imaging spatial filter 3302 is formed of a metal plate, it is supported by the support bar 3311. It goes without saying that the smaller and thinner the support rod 3311 is, the better. FIG. 12 shows an example of an imaging spatial filter 3302 formed by forming a light-shielding film 3313 on a glass substrate 3312. In this case, the support rod 3311 is unnecessary, but it is necessary to design the imaging lens 3201 in consideration of the aberration due to the glass substrate 3312.

Here, the image of the light source spatial filter 3301 is formed by the condenser lens 3106 and the imaging lens 3201 at the position of the image spatial filter 3302. 13 and 14 show the image forming relationship between the light source spatial filter 3301 and the image forming spatial filter 3302 in this embodiment. Both the light source spatial filter and the imaging spatial filter have an annular shape. The light source spatial filter 3301 has an annular portion 3305 having an outer diameter DLO and an inner diameter DLI, and both the inside and the outside of the annular portion 3305 are shielded from light. You. The imaging spatial filter 3302 has a structure in which an annular zone 3306 having an outer diameter DIO and an inner diameter DII is shielded from light, and light is transmitted both inside and outside the annular zone 3306. Assuming that the imaging magnification between the light source spatial filter 3301 and the imaging spatial filter 3302 is M, DLO, DLI, DIO, and DII have the following equation (Equation 6).

DIO = M · δ · DLO
DII = M · ε · DLI
M · DLI ≦ DII ≦ DIO ≦ M · DLO (Equation 6)
Here, δ and ε are coefficients and satisfy the following equation (Equation 6).

0.7 ≦ δ ≦ 1.0
1.0 ≦ ε ≦ 1.3 (Equation 7)
When these coefficients satisfy the above equation (Equation 7), the effect of the present invention appears most remarkably. However, it is not always necessary to satisfy these expressions, and it is only necessary that part or all of the zero-order diffracted light be shielded.

In setting δ and ε, the pattern transfer simulator 1108 selects the values of δ and ε of the spatial filter having the highest contrast.

FIG. 23 shows the contrast when the values of δ and ε are changed. According to this figure, the best contrast is obtained when δ is 0.8ε and 1.1. However, it is clear from FIG. 23 that good contrast cannot be obtained only at this value.

Here, the numerical aperture of the exit side 3204 of the imaging lens (reduction projection optical system) 3200 is NAO, and the numerical aperture of the light source image projected at the same position of 3204 is NAL. Here, NAL / NAO is defined as spatial coherence degree σ. FIG. 24 shows the relationship between σ and contrast. The best contrast is obtained when σ is about 0.9. However, high contrast can be obtained even if σ is slightly shifted from 0.9.

The object of the present invention is most remarkably achieved when the light source spatial filter 3301 and the imaging spatial filter 3302 shown in FIG. 13 are used. Is shielded from light, so that a high-contrast circuit pattern can be imaged on a wafer even if the spatial filters shown in FIGS. 14 to 17 are used. In the spatial filters shown in FIGS. 14 to 17, the shaded portions shown in the figures are the light shielding portions.

The pattern transfer system 3000 includes a light source unit 3100 including an Hg lamp 3101, a color filter 3102, a condenser lens 3103, an integrator 3104, and a condenser lens 3106, and an imaging optical system 3200 including an imaging lens 3201. , A light source spatial filter 3301, an imaging spatial filter 3302, a light source spatial filter control unit (adjustment unit) 3303 for controlling the light source spatial filter 3301, an imaging spatial filter control unit (adjustment unit) 3304 for controlling the imaging spatial filter 3302, and Based on command signals such as DLO, DLI, DIO, and DII obtained from the pattern transfer simulator 1108, the light source spatial filter control unit (adjustment unit) 3303, the imaging spatial filter control unit (adjustment unit) 3304, and the positioning mark detection unit 3403 Control signal A spatial filter control system (adjustment system) 3300 composed of an overall control unit 3305 for controlling the whole operation, a mask stage 3401 for mounting the mask 100, a wafer stage 3402 for mounting the wafer 200, and a positioning mark on the wafer Mark detection unit 3403 for detecting the position, mask stage control system 3404 for controlling mask stage 3304 in accordance with a command from positioning mark detection unit 3403, and wafer stage control system 3405 for controlling wafer stage 3402 in accordance with a command from positioning mark detection unit 3403. And a positioning unit 3400.

The following operation is performed by the above configuration. That is, the mask 100 manufactured by the mask manufacturing system 2000 is mounted on the mask stage 3401 and illuminated by the light source unit 3100. Part of the 0th-order diffracted light transmitted from the mask 100 and from the light source spatial filter 3301 in the light source unit 3100 is shielded by the imaging spatial filter 3302, and the higher-order diffracted light and a part of the 0th-order diffracted light are imaged by an imaging optical system (reduced A circuit pattern is formed on the wafer 200 through a projection lens 3200.

Here, as shown in FIG. 13, as the spatial filter, a ring-shaped one formed by arranging a large number of virtual point light sources is used in order to make the light source have coherence. There are two types of coherence, temporal and spatial. The temporal coherence is the wavelength band of the light source, and the shorter the band, the higher the coherence. The spatial coherence is the size of the light source, and corresponds to the size of the light source spatial filter 3301 in the present invention. However, when the size of the light source is reduced in order to increase the coherence, the light intensity of the light source is reduced, so that the exposure time becomes longer, and the exposure throughput is reduced.

Therefore, when the annular light source spatial filter 3301 is used, the image of the light source spatial filter 3301 formed at the image forming position is the zero-order diffracted light. That is, by using the annular filter 3301, it is possible to realize a light source with high intensity and coherence. This is the same as using a ring-shaped spatial filter to obtain coherent light from white light in a phase-contrast microscope, and is shown by Kubota, Wave Optics (Iwanami Shoten).

There is another reason that the light source spatial filter 3301 has an annular shape. As described above, there is a relationship as shown in the figure between the dimension of the pattern on the wafer to be transferred and the degree of spatial coherence of the light source that transfers the pattern of the dimension with the highest contrast. Therefore, when the spatial coherence degree is determined according to the dimension (pitch) of the pattern to be transferred, the effect of the present invention is remarkably exhibited. By forming the light source spatial filter 3301 and the imaging spatial filter 3302 on the orbicular zone, the degree of spatial coherence, that is, the size of the orbicular zone is controlled and the cost is reduced. However, it goes without saying that the greater the radius of the annular zone (degree of spatial coherence) between the annular zone light source and the spatial filter, the better the resolution.

Also, in the present invention, both the annular light source and the spatial filter are concentric circles. The concentric shape makes it possible to make the circuit pattern to be transferred by the MTF have no directionality. When transferring an LSI circuit pattern having circuit patterns in various directions, it is important that the MTF has no directionality. Further, the concentric filter has an effect that it is difficult for a lens to have complicated aberrations as compared with a non-concentric filter as shown in FIG.

Since the above effects can be achieved if the intensity of the 0th-order diffracted light and the higher-order diffracted light are balanced, the imaging spatial filter 3302 can be omitted and the light source spatial filter 3301 alone can achieve a slightly lower effect. Conversely, the above-described effect can be achieved slightly lower by omitting the light source spatial filter 3301 and using only the imaging spatial filter 3302.

Next, a method of forming a pattern will be described more specifically with reference to FIGS. That is, as described above, the object of the present invention can be achieved by using the light source spatial filter 3301 and the imaging spatial filter 3302. However, the effect of the present invention can be further improved by devising the mask pattern 104 as described below. Can be improved.

FIG. 25 shows the contrast of the circuit pattern projected on the wafer 200 by the imaging optical system 3200 when the line width (light transmission) of a line space pattern having the same pitch formed on the mask 100 is changed. . As shown in FIG. 25, the contrast increases as the line width decreases. That is, the line width may be reduced. Here, when the line width is reduced, the light intensity decreases, so that it is necessary to increase the exposure time. Therefore, the line width of the circuit pattern formed on the mask 100 is determined by the balance between the contrast required to transfer the circuit pattern and the exposure time.

Therefore, in the exposure method according to the present invention, it is desirable that the line width for forming the mask pattern 104 be constant. Therefore, if a wide wafer pattern 204 is desired to be formed, there is some contrivance. Here, if the line width cannot be made constant and becomes wider, the light intensity on the image forming surface becomes larger than that of a pattern having a constant peripheral line width, and the shape of the transfer pattern after resist development is originally obtained. It will be far away from the circuit pattern. This is considered to be due to the influence of light circulating from a part of the circuit pattern having a high light intensity.

正 し く In order to correctly form this wide circuit pattern, it is necessary to form the light intensity with the same light intensity as the peripheral circuit pattern having a constant line width. A method for forming this wide circuit pattern will be described with reference to the drawings. When a circuit pattern is transferred using the imaging optical system 3200, a circuit pattern sufficiently smaller than the resolution of the imaging optical system cannot be resolved and is formed as a uniform image. This phenomenon is used when imaging the wide circuit pattern. That is, by forming a fine pattern having a resolution lower than the resolution of the present optical system on the mask 100 as shown by 105, 106 and 108 in FIGS. 20 and 21, a wide circuit pattern is formed on the wafer 200 as shown in FIGS. Can be transcribed.

FIG. 26 shows the contrast when the pitch of the line space pattern is changed. As shown in FIG. 26, the contrast decreases as the pitch decreases.
That is, there is a position 331 where the contrast becomes almost zero when the pitch is reduced. When it is desired to make the entire surface of a wide circuit pattern white, a pattern having a pitch of 331 is preferably used. Specifically, a pattern of about 1/2 of the pitch of the pattern to be resolved is best. At this pitch, the light intensity of the wide circuit pattern is almost equal to the light intensity of the minimum pattern.

Specifically, to obtain a wafer pattern 204 as shown in FIG. 18, a mask pattern as shown in FIG. 20 is manufactured and a negative resist is used or a mask pattern as shown in FIG. And a positive resist may be used. In this case, the patterns 105, 106, 107, and 108 through which light is to be transmitted are formed by patterns having a small pitch as shown in FIG. The mask pattern 104 for transmitting light over a wide range like these 105, 106, 107, and 108 is automatically generated by the pattern conversion unit 1104 and simulated by the pattern transfer simulator 1108 as necessary.

パ タ ー ン The pattern of the half pitch here may be a half pitch only in one of the X direction and the Y direction. Of course, it goes without saying that the X and Y directions may be a half-pitch grid pattern. Further, the pitch does not necessarily need to be 、, and may be another pitch that makes the light intensity on the wafer pattern 204 necessary and sufficient.

According to the method described above, even in the case of a mask in which an extremely fine circuit pattern and a large circuit pattern are mixed, the transfer can be performed by one reduced projection exposure. However, when an extremely fine circuit pattern and a large circuit pattern are transferred by two or more reduction projection exposures, the above method is not necessary, and the pattern 104 can be formed only by a pattern having a constant line width. In this case, there is an effect that a system such as the pattern conversion unit 1104 becomes unnecessary.

As described above, in the method according to the present invention, there is no need to dispose a phase shifter, so that the processing in the pattern conversion unit 1104 is simple, time is saved, and errors are reduced.

Further, in the present embodiment, the mask pattern is converted based on a circuit pattern having a constant line width. However, in a circuit pattern having a large number of repetitive portions such as a memory, an optimal wafer pattern 204 can be obtained while performing a simulation with the transfer simulator 1108. A simple mask pattern 104 may be obtained. That is, the mask pattern 104 is obtained by the transfer simulator 1108 for each memory cell.

Next, the transfer mechanism of the mask pattern on the wafer will be described with reference to FIGS. 1, 2, and 3. FIG. FIG. 1A is a cross-sectional view of a mask 100 in which a mask pattern 104 is formed on a glass substrate 101 by chrome 102. A waveform 301 in FIG. 1B is a strong signal intensity distribution of the image pattern of the mask pattern 104. The waveform 301 is considered to be divided into a waveform 302 due to the zero-order diffracted light and a waveform 303 due to the high-order diffracted light. In the case where the mask pattern 104 is a fine pattern 105 as shown in the figure, the waveform 301 detected by the higher-order diffracted light is smaller than the waveform 302 generated by the 0th-order diffracted light, so that the detected waveform 301 has a lower contrast AM / AV. Here, since the component of the waveform 302 can be removed by blocking the 0th-order diffracted light, the detected waveform has a high contrast like the waveform 302.

According to Babinet's principle, in a diffraction pattern other than the zero-order diffraction light, light from an adjacent pattern appears as if the phase of the adjacent pattern is inverted (shifted by π). In other words, if the 0th-order diffracted light is blocked on the diffraction pattern, the light that forms an image on the wafer surface is formed by the light from an adjacent pattern whose phase is inverted (shifted by π). Is equivalent to Based on this technical idea, as shown in FIGS. 2 and 3, at least a part of the 0th-order diffracted light is shielded by a light shielding plate 324 (imaging spatial filter 3302) on the diffraction image plane, as shown in FIGS. Since only the diffracted light whose phase has been inverted reaches the image plane, the intensity distribution on the image plane has a high contrast.

Next, the mechanism for improving the contrast will be described. That is, as an example, FIG. 28 shows a transfer result of the circuit pattern formed on the mask shown in FIG. 27, and FIG. 29 shows a change in contrast when the pitch PIT is changed. In the conventional reduction projection exposure method, the contrast sharply decreases as the pattern size becomes extremely fine as indicated by 342, whereas in the reduction projection exposure method according to the present invention, the contrast does not decrease as indicated by 341. I understand.

FIG. 30 shows an evaluation example of the depth of focus of the reduction projection exposure method according to the present invention. According to the present invention, as indicated by 343, the contrast shows a value of about 80% or more in a range of ± 1.5 μm. In the conventional reduced projection exposure method, as indicated by 344, the contrast sharply drops due to defocus. This indicates that the present invention can cope with a resist having a large resist film thickness, and consequently can form a wafer pattern using the resist with a high aspect ratio. As a result, a pattern having a good aspect ratio and a high aspect ratio can be formed during etching.

Similarly, FIG. 38 shows the depth of focus 362 of the present invention and the depth of focus 363 of the phase shifter method. Both are almost the same, and the depth of focus 362 of the present invention shows a sufficiently good result as compared with the depth of focus 361 of the conventional method.

FIG. 31 shows an embodiment of a so-called excimer stepper using an excimer laser having a wavelength shorter than i-line as a light source. According to this embodiment, the effect of the light source spatial filter 3301 can be achieved by scanning the position of the light source spatial filter 3301 with an excimer laser beam so as to form an annular shape. According to this embodiment, the shape of the light source spatial filter 3301 can be easily controlled by controlling the scanning unit. That is, this embodiment is an example in which an excimer laser using a gas such as KrF (krypton fluoride) having a shorter wavelength is used for the light source unit 3100 of the embodiment shown in FIG. By using light having a shorter wavelength, a finer circuit pattern can be transferred. In this embodiment, it goes without saying that a laser of another wavelength may be used. The light source unit 3100 of this embodiment includes an excimer laser 3111, a shutter 3108, a beam expander 3112, and an X galvanometer mirror 3113.
, Y galvanometer mirror 3114, integrator 3104, integrator cooling means 3120, condenser lens 3106, and spatial filter unit 3300 comprises a scanning control system 3311, a liquid crystal display element 3312, and a liquid crystal control system 3313, as shown in FIG. The light source space filter adjustment system 3303 in the example corresponds to the X galvanometer mirror 3113, the Y galvanometer mirror 3114, and the scanning control system 3311. Here, the integrator cooling means 3120 is for preventing the temperature of the integrator from rising due to the light of the excimer laser being concentrated on the integrator. Specifically, the integrator cooling means 3120 may be a means for circulating cooling water or a nitrogen gas for cooling. May be sprayed. Other components are in accordance with the embodiment shown in FIG. That is, in this embodiment, the position of the light source spatial filter 3301 is the position of the integrator 3104, and the X-gavano mirror 3113 and the Y galvano mirror 3114 in the light source spatial filter adjustment unit 3303 are scanned to be in the same shape as the light source spatial filter 3301. Make an annular light source. Also, as described in the embodiment shown in FIG. 8, the imaging spatial filter 3302 has a ring-shaped light-shielding portion formed on the liquid crystal display element 3312 by the liquid crystal control system 3313 so as to shield part or all of the zero-order diffracted light. It is formed. Accordingly, the resolution of the liquid crystal display element 3303 needs to have a necessary and sufficient precision to form the above-mentioned annular shape. In addition, it is only necessary to form an imaging spatial filter. Even if it is not a liquid crystal display element, for example, even if it is a metal plate formed so as to block light in a ring shape from the beginning, it is possible to form a ring on glass. The light shielding film may be formed.

Further, in this embodiment, the excimer laser light is focused on one point on the imaging spatial filter 3302. Only by scanning on the annular zone on the light source spatial filter does the zone form on the imaging spatial filter. Therefore, the object of the present invention can be achieved by shielding light only on each point on the imaging spatial filter in accordance with scanning on the light source spatial filter. Thus, since the light is not blocked too much, there is an effect that the exposure time can be shortened and the throughput can be increased. By the way, the mask 100 is mounted on the mask stage 3401, and the mask stage 3401 is controlled by the mask stage control system 3404 to position the mask 100 at the reference position. The wafer stage control unit 3405 controls the wafer stage 3402 based on the detection signal, and aligns the mask 100 with the wafer 200. After the alignment, the shutter 3108 is opened, and the annular light source is formed by the light source space filter adjusting unit 3303, so that the mask 100 is illuminated and the mask pattern is transferred onto the wafer 200 to form a wafer pattern.

In this embodiment, since the light from the excimer laser 3111 has strong coherence, it is necessary to reduce this coherence to an appropriate value, and this is simultaneously achieved by forming an annular light source on the light source spatial filter 3301. There is also an effect.

In particular, in the present invention, uniform illumination within the exposure field (in the exposure area) is sufficient, and it is not always necessary to simultaneously illuminate annular illumination formed by arranging a large number of virtual point light sources. Obviously, the present invention may be realized by time division or by scanning as in the above-described embodiment. Also, it is obvious that the annular illumination formed by arranging a large number of point light sources may be realized by being divided into a plurality.

As described above, according to the present invention, a large circuit pattern portion and a small circuit pattern portion can be exposed twice. Further, a mask in which a phase shifter is arranged at least partially (partly or entirely) of the circuit pattern on the mask can be used. In this case, since the central value of the incident angle of the illumination light incident on the reticle is increased by using the annular light source, the thickness of the phase shifter is slightly reduced in order to set the phase shift by the phase shifter to π. There is a need. This value is calculated using the incident angle θ.

(2m + 1) π = (d / cos θ) · (n / λ) · 2π (Equation 8)
Here, m is an integer, d is the thickness of the phase shifter, n is the refractive index of the phase shifter, and λ is the exposure wavelength.

As described above, the present invention produces new effects such as being able to cope with various circuit patterns by using a combination of a conventional exposure apparatus or a phase shifter.

Further, in practicing the present invention, the incident angle of light applied to the reticle becomes large, so that the thickness of the circuit pattern such as chrome on the mask becomes a problem. That is, depending on the irradiation angle, a shadow due to the thickness of the circuit pattern such as chrome on the mask may be formed. Therefore, in the practice of the present invention, in order to obtain an accurate transfer pattern dimension, it is desirable that the thickness of the circuit pattern such as chrome on the mask be thin. Therefore, when the thickness of the circuit pattern of chrome or the like is dm and the allowable value of the transfer pattern is pc, it is necessary to satisfy Expression (9).

dm ≦ pc / l (Equation 9)
Here, 1 is a reduction ratio of the reduction projection lens.
To achieve this, it is necessary to pattern the mask with another material having a lower transmittance than chromium. However, if the size of the circuit pattern of the mask is determined in consideration of the dimensional change due to the incident angle of illumination, the above problem can be reduced.

(4) Further, if the transmission portion of the mask is provided with a coating that maximizes the transmittance of the light beam incident at the used incident angle, the exposure amount increases, and the time required for exposure can be reduced.

Also,
Since the present invention utilizes wave nature as described above, it is apparent that the present invention can be applied to an exposure apparatus using an electron beam or X-ray.

FIG. 3 is a diagram showing a light diffraction phenomenon by a circuit pattern formed on a mask for explaining the principle of the present invention. It is a figure for explaining the principle of the present invention. FIG. 3 is a diagram for explaining the principle of the present invention, similarly to FIG. 2. 1 is a schematic perspective view showing an embodiment of an exposure system in a reduction projection exposure apparatus according to the present invention. FIG. 3 is a cross-sectional view illustrating an arrangement relationship between a light source spatial filter and an imaging spatial filter in the exposure system according to the present invention. FIG. 6 is a sectional view showing an arrangement relationship of an imaging spatial filter in the exposure system according to the present invention, similarly to FIG. 5. FIG. 7 is a diagram showing an image forming relationship of a circuit pattern on a mask in a conventional reduction projection exposure apparatus. FIG. 1 is a configuration diagram showing one embodiment of an entire exposure system according to the present invention. FIG. 1 is a perspective view showing a first embodiment of an integrator having a light source spatial filter in an exposure system according to the present invention. FIG. 7 is a perspective view showing a second embodiment of the integrator having the light source spatial filter in the exposure system according to the present invention. 1 is a perspective view showing a first embodiment of an imaging lens having an imaging spatial filter in an exposure system according to the present invention. FIG. 7 is a perspective view showing a second embodiment of the imaging lens having the imaging spatial filter in the exposure system according to the present invention. FIG. 2 is a plan view illustrating a relationship between a light source spatial filter according to the first embodiment of the present invention and an imaging spatial filter according to the first embodiment. FIG. 9 is a plan view illustrating a relationship between a light source spatial filter according to a second embodiment of the present invention and an imaging spatial filter according to the second embodiment. FIG. 11 is a plan view illustrating a light source spatial filter according to a third embodiment of the present invention and an imaging spatial filter according to the third embodiment. FIG. 14 is a plan view showing a light source spatial filter according to a fourth embodiment of the present invention and an imaging spatial filter according to the fourth embodiment. It is a top view showing the relation between the light source spatial filter of a 5th example concerning the present invention, and the imaging spatial filter of a 5th example. It is the top view and sectional view showing an example of the wafer pattern transferred on the wafer concerning the present invention. It is a top view and a sectional view showing other examples of a wafer pattern transferred on a wafer concerning the present invention. FIG. 19 is a plan view and a sectional view showing an example of a mask pattern for obtaining the wafer pattern shown in FIG. 18 according to the present invention. FIG. 20 is a plan view and a sectional view showing an example of a mask pattern for obtaining the wafer pattern shown in FIG. 19 according to the present invention. FIG. 3 is a view showing various forms of a mask pattern according to the present invention. FIG. 6 is a diagram illustrating a relationship between δ and ε related to a spatial filter of the present invention and contrast on a wafer. FIG. 7 is a diagram illustrating a relationship between σ related to the spatial filter of the present invention and contrast on a wafer. FIG. 4 is a diagram illustrating a relationship between a line width of a circuit pattern on a mask and a contrast on a wafer according to the present invention. FIG. 4 is a diagram illustrating a relationship between a pitch of a circuit pattern on a mask and a contrast on a wafer according to the present invention. It is a top view showing an example of a circuit pattern on a mask concerning the present invention. FIG. 27 is a diagram showing a transfer result of a circuit pattern in FIG. 27 according to the present invention. FIG. 4 is a diagram showing a relationship between a pitch of an extremely fine circuit pattern formed on a mask according to the present invention and a contrast on a wafer. FIG. 3 is a diagram illustrating a relationship between a depth of focus and a contrast according to the present invention. FIG. 11 is a configuration diagram showing another embodiment of the entire exposure system according to the present invention. FIG. 3 shows a response function of the system according to the invention. FIG. 4 is an explanatory diagram of a correlation function between a light source and a spatial filter according to the present invention. It is explanatory drawing of the correlation function of a light source and a spatial filter of a conventional example. FIG. 4 is an explanatory diagram of a calculation method of a correlation function between a light source and a spatial filter. FIG. 4 is an explanatory diagram of a calculation method of a correlation function between a light source and a spatial filter. It is a figure showing the calculation result of OTF of the present invention. FIG. 3 is a diagram illustrating the depth of focus of the present invention. It is a figure explaining the effect of an annular light source and an annular filter. It is a figure showing an annular filter. It is a figure which shows the Example of an annular light source and an annular filter. It is a figure which shows an Example of the intensity distribution of an annular light source and an annular filter. It is a figure showing other examples of intensity distribution of an annular light source and an annular filter. It is a figure showing other examples of an annular light source and an annular filter. It is a figure showing an MTF curve. It is a figure showing an MTF curve. It is a figure showing other examples of an annular light source and an annular filter. FIG. 48 is a diagram showing a block diagram of an embodiment of the light source in FIG. 47. It is a figure showing an example of an integrator. It is a figure showing other examples of an annular light source and an annular filter.

Explanation of reference numerals

1000: pattern generation system, 2000: mask production system, 3000: exposure system 3101: Hg lamp, 3104: integrator, 3301: light source spatial filter (ring-shaped) 100: mask, 3201: imaging lens, 3302: imaging spatial filter , 200 ... wafer

Claims (7)

A method for inspecting a pattern formed on a substrate, comprising:
Forming an annular secondary light source from the light emitted from the light source,
Irradiating the substrate on which the pattern is formed with light emitted from the formed annular secondary light source,
Imaging the substrate irradiated with the light to obtain an image of the substrate,
A pattern inspection method for detecting a defect in the pattern by comparing the image obtained by the imaging with a reference image. The pattern inspection method according to claim 1, wherein the secondary light source having the annular shape is formed by scanning the light emitted from the light source in an annular shape using a pair of galvanomirrors. 2. The pattern inspection according to claim 1, wherein the light emitted from the light source is light having coherence, and the coherence of the light is reduced by a coherence reducing unit and the light is irradiated onto the substrate. 3. Method. 4. The pattern inspection method according to claim 1, wherein the light emitted from the light source is a laser in an ultraviolet region. A method for inspecting a pattern formed on a substrate, comprising:
Fire a laser from the light source,
Transforming the optical path of the emitted laser to form a secondary light source;
Irradiating the substrate with a laser emitted from the formed secondary light source,
Imaging a diffracted light image by diffracted light from the pattern of the substrate irradiated with the laser,
A pattern inspection method for detecting a defect of the pattern by comparing the diffracted optical image obtained by the imaging with a reference image. 6. The pattern inspection method according to claim 5, wherein the secondary light source has an annular shape. The pattern inspection method according to claim 6, wherein the annular secondary light source is formed by scanning light emitted from the light source in an annular shape using a pair of galvanomirrors.

Images