JP3090754B2 - Projection exposure equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a projection exposure apparatus for forming a fine resist pattern required for manufacturing a semiconductor integrated circuit, and more particularly to a projection exposure apparatus in which an optical system is optimized so that the depth of focus is maximized.

[0002]

2. Description of the Related Art In recent years, the progress of optical lithography has been remarkable, and there is a possibility that a 0.5 μm rule can be realized in a g-line (436 nm) or i-line (365 nm) projection exposure apparatus. This is because the performance of the projection exposure apparatus has been improved, and in particular, the NA of the lens has been increased. However, it is doubtful that the 0.3 μm rule of the next generation can be achieved with the extension up to now. Although the resolution is improved by increasing the NA of the lens and shortening the wavelength of the exposure light, the practical resolution is not significantly improved because the depth of focus is reduced. Therefore, development of a technology for improving the depth of focus is desired.

FIG. 25 shows a schematic configuration of a conventional projection exposure apparatus generally used. In this figure, 1 is a lamp composed of a mercury lamp, 2 is an elliptical reflecting mirror, 3 is a second focal point of the elliptical reflecting mirror 2, 4 is an input lens, 5 is an optical integrator (fly-eye lens), 6 is an output lens, 7 is a collimation lens, 8 is a reticle (mask), 9 is an aperture stop as a uniform stop, 10 is a filter for passing only light of a wavelength whose optical system is aberration-corrected, and 11 and 12 are optical path bending devices. Cold mirror for lowering the height of the lens, 13 is a lamp house, 14 is a projection optical system for projecting a pattern image on the reticle 8 onto a wafer by a lens, mirror or a combination thereof, 15 is a wafer, and 16 is a numerical aperture It is an aperture to do.

The basic structure of a conventional projection exposure apparatus is shown in FIG.
26, there are many other than those shown in FIG.
1, a light source 1, a first condensing optical system 18, a uniforming optical system 19, a second condensing optical system 20, a reticle 8, a projection optical system 14, and a wafer 15 are arranged in this order. The first condensing optical system 18 is a part corresponding to the elliptical reflecting mirror 2 and the input lens 4 in the example of FIG.
A plane mirror, a lens, and the like are appropriately arranged, and have a role of allowing the light beam emitted from the light source to enter the uniforming optical system 19 as efficiently as possible. The homogenizing optical system 19 is a part corresponding to the optical integrator 5 in FIG. 25, and an optical fiber, a polyhedral prism, or the like may be used as the other components.

The second condensing optical system 20 is a portion corresponding to the output lens 6 and the collimating lens 7 in FIG. 25, superimposes the light emitted from the uniformizing optical system 19, and further secures the image plane telecentricity. In addition, a filter corresponding to the filter 10 of FIG. 25 is inserted at a position where the light beam is close to the optical axis parallel, and a reflecting mirror corresponding to the cold mirrors 11 and 12 is also inserted, although the location is not unique.

When the light coming from the reticle 8 is viewed from the apparatus having the above-described configuration, the nature of the light becomes the nature of the light coming out of the homogenizing optical system 19 through the second condensing optical system 20. The exit side of the homogenizing optical system 19 appears as an apparent light source. For this reason, in the case of the above configuration, the exit side 24 of the homogenizing optical system 19 is generally called a secondary light source. When the reticle 8 is projected onto the wafer 15, the formation characteristics of the projection exposure pattern, that is, the resolution and the depth of focus,
It is determined by the numerical aperture of the projection optical system 14 and the properties of the light irradiating the reticle 8, that is, the properties of the secondary light source 24.

FIG. 26B is an explanatory diagram relating to a reticle illumination light beam and an image forming light beam in the projection exposure apparatus shown in FIG. In FIG. 26B, the projection optical system 14 usually has an aperture stop 16 inside, and regulates the angle θa through which the light passing through the reticle 8 can pass, and the angle of the light beam falling on the wafer 15. θ is determined.

Generally, the numerical aperture NA of a projection optical system is an angle defined by NA = sin θ, and when the projection magnification is 1 / m, the relationship is sin θa = sin θ / m. In this type of apparatus, it is common that the image plane telecentric, that is, the principal ray falling on the image plane is formed perpendicular to the image plane. In order to satisfy the condition of this "image plane telecentric", FIG. A real image of the exit surface of the homogenizing optical system 19, that is, the light source surface of the secondary light source 24 in FIG.

Under these conditions, a solid angle when the secondary light source surface is viewed from the reticle 8 through the second condensing optical system 20 is taken as a range of light incident on the reticle 8, and a half angle thereof is φ.
And the coherency σ of the illumination light is σ = sinφ / sin
When defined by θa, it was considered that the pattern formation characteristics were determined by NA and σ.

Next, the relationship between NA and σ and the pattern forming characteristics will be described in detail. The larger the NA, the higher the resolution, but the shallower the depth of focus, and it becomes difficult to secure a wide exposure area due to the aberration of the projection optical system 14. Without a certain exposure area and a certain depth of focus (for example, 10 mm square, ± 1 μm), it cannot be used for actual LSI manufacturing and the like, so that the conventional apparatus has a limit of about NA = 0.35. On the other hand, the σ value is mainly related to the pattern cross-sectional shape and the depth of focus, and is related to the resolution in correlation with the cross-sectional shape. When the σ value is small, the edge of the pattern is emphasized, so the cross-sectional shape becomes a good pattern shape with the side wall approaching perpendicular,
The resolution of a fine pattern deteriorates, and the focus range that can be resolved is narrowed. Conversely, when the σ value is large, the resolution in a fine pattern and the focus range that can be resolved are slightly improved. However, in the case of a thick resist, the cross section of the pattern is trapezoidal or triangular in the case where the pattern cross section is inclined slowly. For this reason, in a conventional projection exposure apparatus, a relatively balanced σ value is
σ is fixedly set to 0.5 to 0.7, and σ is experimentally
= 0.3 and the like are only being attempted. Since the size of the light source surface of the secondary light source 24 can be determined to set the σ value, the circular aperture stop 9 for setting the σ value is generally placed immediately after the light source surface of the secondary light source 24.

As one method of improving the depth of focus of such a general projection exposure apparatus, a special stop is inserted to make the intensity distribution of the secondary light source of the projection exposure apparatus higher at the peripheral portion than at the central portion. There is a ring illumination exposure method (JP-A-61-91662). That is, the filter shown in FIG. 27 is inserted instead of the circular aperture stop 9 for setting the σ value in FIG. The effect of improving the depth of focus obtained by the filter shown in FIG. 27A was evaluated by simulation.
FIG. 28 shows the depth of focus with respect to the mask pattern size, and shows the dependency of the annular illumination filter on the center occlusion ratio ε. The central shielding ratio is obtained by the following equation: ε = ε = R ₁ and R ₂ of the annular filter shown in FIGS.
represented by r ₂ / r _1. The NA of the exposure apparatus is 0.54, the coherence factor is 0.5, and the exposure wavelength is 436 nm (g
Line). It can be seen that the larger the ε, the greater the effect of improving the critical resolution and the depth of focus.

However, since the actual secondary light source is formed on the exit surface of the fly-eye lens and is spatially discrete, if ε is made too large (the annular zone is made thin), the annular aperture stop is increased. Light cannot pass through. Therefore, ε is usually set to about 0.5 to 0.7. Then, as can be seen from FIG. 28, the effect of improving the depth of focus is reduced to about 13% at the maximum. FIG. 28 also shows that the effect of improving the depth of focus on the pattern size is remarkable. The effect of improving the depth of focus is recognized when the pattern size is 0.6 to 0.9 μm, but becomes worse when the pattern size is 0.9 μm or more. I have. Further, with respect to the annular diaphragm having a transmittance distribution as shown in FIG. 27B, no specific method for determining the transmittance is described.

As another method for improving the depth of focus of a projection exposure apparatus, there is a superflex method in which a spatial frequency filter is inserted into the pupil of the projection exposure apparatus. , 29a-ZC-
8). According to this, when images are formed at different positions z = ± β in the optical axis direction and the amplitudes of the two images each having a phase shift of ± Δφ are combined, (1) an increase in the depth of focus due to the multiple imaging (FLEX) effect And (2) a pseudo phase shift effect due to cancellation of phases at the pattern edge can be obtained at the same time. The amplitude distribution in the horizontal (x) direction and the optical axis (z) direction is represented by U (x,
z), such a composite amplitude is given by U ′ (x, z) = [exp (iΔφ) U (x, z−β) + exp (−iΔφ) U (x, z + β)] / 2 . This amplitude is applied to the projection lens pupil (aperture)
This is equivalent to that obtained by providing a spatial filter having a radial distribution of complex amplitude transmittance represented by the following equation.

T (r) = cos (2πβr ² −θ / 2) (where θ = 2Δφ−8πβ / NA ² ) If β and θ are appropriately selected, the distance between the imaging planes between the two images And the interference effect (or the spatial frequency transfer characteristic) can be arbitrarily controlled.

In this method, since a filter having a small transmittance is inserted into a pupil inside the projection optical system, the filter absorbs exposure light and becomes heat, and the optical system causes thermal expansion and deteriorates transfer accuracy. There are disadvantages. Also,
As shown in FIG. 29, there is also a problem that the effect of increasing the depth of focus greatly depends on the pattern size. The abscissa of FIG. 29 indicates a value obtained by standardizing the mask pattern size by λ / NA, and the ordinate indicates a value obtained by standardizing the focus margin by λ / NA ² . When viewed from a focus margin that secures a pattern contrast of 60% or more on the image plane, the focus margin becomes larger in a normal exposure method when the pattern size is 0.8 (standardized dimension) or more.

[0016]

The above two techniques for improving the depth of focus have the following problems, respectively. That is, in the annular illumination exposure method, (1) when considering the actual configuration of the secondary light source, the effect of improving the depth of focus is as small as 13% at most. (2) As shown in FIG. 29, the effect of the depth of focus improvement on the pattern size is large. (3) The annular diaphragm having the transmittance distribution as shown in FIG. 27B has three problems that no specific method for determining the transmittance is described.

Further, in the superflex method, (1) the filter absorbs the exposure light and turns into heat, and the optical system causes thermal expansion to deteriorate the transfer accuracy. (2) As shown in FIG. 29, there is a two problem that the depth of focus improvement effect largely depends on the pattern size.

The present invention has been made in consideration of the above circumstances, and has as its object to solve the above problem when the numerical aperture of the projection optical system and the size of the secondary light source are given. It is an object of the present invention to provide a projection exposure apparatus which can optimize the depth of focus and improve the exposure accuracy.

[0019]

In order to achieve the above object, the present invention employs the following configuration.

[0020]

That is, according to the present invention, in a projection exposure apparatus for projecting and exposing a mask pattern onto a wafer via a projection optical system, the mask is illuminated so that the depth of focus becomes maximum according to a desired pattern size. the next source, a special diaphragm allowed to large the exit surface in the intensity distribution of the light source in the direction of the peripheral portion than the central portion is provided, and on the pupil plane of the projection optical system, the light
A filter is provided for changing the optical path length of the light passing through the pupil plane along the radial direction centered on the optical axis without changing the transmittance .

Here, as a preferred embodiment of the present invention,
The following are listed. (1) The filter passes through the periphery of the pupil plane of the projection optical system
Light is optical relative to light passing through other parts
The target optical path length is made longer. (2) The filter passes through the periphery of the pupil plane of the projection optical system
Light is optical relative to light passing through other parts
The objective optical path length is to be shortened. (3) The filter is located at the periphery and center of the pupil plane of the projection optical system.
Light passing through
In contrast, the optical path length is made longer. (4) The filter is located at the periphery and center of the pupil plane of the projection optical system.
Light passing through
On the other hand, the optical path length is shortened. (5) Each time the mask used for transfer is changed, the secondary light source is emitted
In-plane intensity distribution and complex amplitude transmittance of pupil function of projection optical system
Reset at least one of the cloths and use the secondary light for each mask.
Source intensity distribution and pupil function complex amplitude transmittance distribution of projection optics
Having means for changing at least one of the following.

[0023]

According to the present invention, there is provided a first light-collecting optical system for collecting light from a light source, a uniforming optical system for uniformizing light collected by the first light-collecting optical system, A second condensing optical system for condensing the light uniformized by the optical system and irradiating the mask with the light, and a projection optical system for projecting the light transmitted through the mask onto a wafer; In a projection exposure apparatus that transfers the light onto a wafer, as a secondary light source on the emission side of the uniformizing optical system, a special stop is provided that makes the intensity distribution in the emission surface of the light source larger in the peripheral part than in the central part, and No change in light transmittance on the pupil plane of the projection optical system
When a filter that generates a relative delay or advance of the phase with respect to light passing through the peripheral portion of the pupil plane than light passing through other portions is disposed, and the radius of the secondary light source is L The radius d of the central portion where the intensity of the secondary light source is small is d ≧ 0.
When 5L is satisfied and the intensity lc at the central portion where the intensity is small is lc ≦ 0.35 × lp with respect to the peripheral intensity lp, and the radius of the pupil plane is H, the radial coordinate r of the pupil is , R ≦
1 relative to the phase of light passing through the 6H / 7 region.
The region that causes a delay or advance of 30 ° or more is r> 6
The pupil function exists in the range of H / 7.

[0025]

According to the present invention, when the numerical aperture of the projection optical system and the size of the secondary light source are given, the secondary light source radial intensity distribution of the projection optical system, the complex amplitude transmittance distribution of the mask, and the projection By setting at least one of the pupil function radial direction complex amplitude transmittance distribution of the optical system using a successive approximation method,
It is possible to realize an optimum depth of focus according to a desired pattern size.

[0026]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The details of the present invention will be described below with reference to the illustrated embodiments.

FIG. 1 is a schematic diagram showing a projection exposure apparatus according to a first embodiment of the present invention. In FIG. 1, 1 is a lamp (light source), 2 is an elliptical reflecting mirror, 3 is a second focal point of the elliptical reflecting mirror 2, 4 is an input lens, 5 is an optical integrator, 6 is an output lens, 7 is a collimation lens, and 8 is A reticle, 9 'is a special aperture having a distribution in transmittance, 10 is a filter, 11 and 12 are cold mirrors, 13 is a lamp house, 14 is a lens, a mirror, or a combination thereof, and an image of a pattern on the reticle 8 is formed on a wafer. Projection optical system for projecting, 15 is a wafer, 16 '
Is a filter having a distribution in the optical path length.

A mercury lamp was used as the light source lamp 1 and g
Bright lines such as a line 436 nm, an h line 405 nm, and an i line 365 nm, or a continuous spectrum near these wavelengths are extracted and used. For this reason, the lamp 1 as the light source needs to have high luminance, and it is better to be close to a point light source in consideration of the light collection efficiency and irradiation uniformity. However, since such an ideal light source does not exist in practice, a lamp 1 having a finite size and a distribution of intensity must be used, and the light emitted from such a lamp 1 cannot be reduced. An issue is how to convert the light into light with good efficiency and uniform irradiation.

In the apparatus shown in FIG. 1, the lamp 1 is placed at the first focal point of the elliptical reflecting mirror 2, and the light beam is once collected near the second focal point 3 of the elliptical reflecting mirror 2. Then, the light beam is converted into a substantially parallel light beam by the input lens 4 which shares a substantially focal position with the second focal point 3, and is input into the optical integrator 5.
The optical integrator 5 is a bundle of a number of rod-shaped lenses, and is also called a fly-eye lens. Passing through the optical integrator 5 is a main factor for improving the irradiation uniformity, and the input lens 4 plays a role in reducing vignetting of the light beam passing through the optical integrator 5 and improving the light collection efficiency.

The light exiting the optical integrator 5 is
By the output lens 6 and the collimation lens 7, the light beams emitted from the respective small lenses of the optical integrator 5 are collected so as to be superimposed on the reticle 8. Although the light beam incident on the optical integrator 5 has an intensity distribution depending on the location, the light emitted from each small lens of the optical integrator 5 is superimposed almost equally, so that the irradiation intensity on the reticle 8 becomes almost uniform. As a matter of course, if the intensity distribution of the light incident on the optical integrator 5 is close to uniform, the illuminance distribution of the reticle 8 on which the emitted light is superimposed becomes more uniform. On the emission side of the optical integrator 5, a special stop 9 'described later is arranged.

When a mercury lamp is used as the lamp 1 and the light is condensed by the elliptical reflecting mirror 2, the structure of the mercury lamp is vertically elongated as shown in FIG. I can't take it out. Therefore, as shown in FIG. 1, the intensity distribution of light entering the central portion of the optical integrator 5 may be reduced only by using a convex lens as the input lens 4. Therefore, a biconvex or one-sided convex / concave conical lens may be inserted between the input lens 4 and the optical integrator 5 to make the intensity distribution of light entering the optical integrator 5 more uniform.

The filter 10 is for passing only light having a wavelength whose optical system has been aberration-corrected. The cold mirrors 11 and 12 bend the optical path to lower the height of the apparatus and to reduce long-wavelength photothermal rays. It plays the role of transmitting the light and absorbing it into the coolable portion of the lamp house 13. The light irradiated on the reticle 8 passes through the projection optical system 14, and the image of the fine pattern on the reticle 8 is projected and transferred to the resist on the wafer 15. In the projection optical system 14, there is a filter 16 'for changing the optical path length in the radial direction from the optical axis with respect to the transmitted light.

Next, a method of designing the secondary light source intensity distribution and the pupil function in the apparatus of the embodiment will be described in detail with reference to FIGS. FIG. 2 shows an algorithm for optimizing a secondary light source and a pupil function of the projection exposure apparatus. This optimization algorithm is a method called simulated annealing (Simuleted Annealing). The simulated annealing method is a Monte Carlo optimization algorithm using an analogy of statistical thermodynamics, and is used to find a global minimum when an appropriate cost function is given.

Now, it is assumed that NA = 0.54 (numerical aperture), λ = 436 nm (exposure wavelength), and σ = 0.5 (coherence factor) as an exposure apparatus. Given the numerical aperture and the size of the secondary light source (coherence factor), the intensity distribution of the secondary light source and the complex amplitude transmittance distribution of the pupil function in the projection optical system are optimized so that the performance of this exposure apparatus is optimized. decide. The calculation procedure when the depth of focus of the pattern of L / S = 0.6 μm is to be improved will be described with reference to FIG.

In step S1, the initial distribution of the radial intensity distribution S (r) of the secondary light source and the radial complex amplitude transmittance distributions t (r) and φ (r) of the pupil function are determined. Where r
Is the spatial coordinate in the radial direction, t is the amplitude transmittance, and φ is the phase. Although the convergence speed varies depending on how the initial distribution is given, 0 ≦ S (r) ≦ 1, 0 ≦ t (r) ≦ 1, 0 ≦ φ
Any distribution may be used as long as the boundary condition (r) ≦ 180 ° is satisfied.

In step S2, as shown in FIG. 3, a small displacement (grain) having a magnitude of ε is applied to one portion of the initial distribution. The place to give the grain, ±, is chosen randomly. In step S3, an image intensity distribution 1 before and after grain application is calculated. Here, an ideal (target) image intensity distribution li is also calculated in advance.

In step S4, cost functions E and E 'are calculated. Here, it is given by ^{E = (C -Ci) 2 E} '= (C'-Ci) 2. Here, C is a contrast value obtained from the image intensity distribution calculated after applying the grain, C ′ is a contrast value obtained from the image intensity distribution calculated before the application of the grain, and Ci is an ideal (target) image intensity. This is a contrast value obtained from the distribution.

In step S5, ΔE = EE ′ is obtained. If ΔE <0, it implies that the image intensity distribution after applying the grain is closer to the ideal (target) image intensity distribution, and the applied grain is locally optimal. It indicates that it has contributed to asymptotically approaching the value. Conversely, if ΔE> 0, it indicates that the given grain contributed to keep away from the optimum value when viewed locally.

In step S6, the grain ε is calculated based on ΔE.
It is determined whether or not to accept. If ΔE <0, it means that the given grain has contributed to asymptotically approaching the optimal value locally, and thus all are accepted. Conversely, if ΔE> 0, it means that the given grain has contributed to keep away from the optimum value when viewed locally, and is basically not accepted. However, if all are not accepted, it will not be possible to escape from the local minimum value, so it will be accepted with a certain probability. If the probability is too large, the value of the cost function fluctuates, and an optimal solution cannot be obtained forever. Conversely, if it is too small, it will not be a sufficient force to escape from the local minimum point, and it will be difficult to obtain an optimal solution.

Then, when ΔE> 0, by using analogy from thermodynamics, a grain that raises the cost function is accepted with a probability according to the Boltzmann distribution of P (ΔE) = exp (−ΔE / T). In other words, grains that slightly increase the cost function are accepted with a relatively high probability, and grains that greatly increase the cost function are accepted with a small probability. Here, T is a variable called “temperature” for controlling the probability.

In step S7, when the grains are received, the radial intensity distribution of the secondary light source and the radial complex amplitude transmittance distribution of the pupil function after the grains are applied are re-determined to obtain S (r), Let t (r) and φ (r). When the grain is not accepted, the radial intensity distribution of the secondary light source and the radial complex amplitude transmittance distribution of the pupil function before giving the grain are re-examined, and S (r), t (r), φ (r) And An optimal solution can be obtained by repeating such an operation for all S (r), t (r), and φ (r). By decreasing the "temperature" T as the convergence point is approached, vibration near the convergence point is suppressed. T
Decreases, the probability of accepting a change in ε for a large positive ΔE sharply decreases.

S (r), which is optimized using the above method,
FIG. 4 shows φ (r). t (r) is 1 (constant)
And Based on this result, the transmittance distribution of the special diaphragm 9 'and the optical path length distribution of the filter 16' were determined. That is, the special aperture 9 'has a low transmittance at the center of the optical axis and is almost constant, and has a high transmittance at the periphery. The filter 16 'has a short optical path length at the center of the optical axis and is almost constant, and has a long optical path length at the periphery. As shown in FIG. 5, the image intensity distribution at a normal focus (S (r) = 1) and a normal pupil (t (r) = 1, φ (r) = 0) at a defocus of 0.5 μm is ideal. (Target) Image intensity distribution l
i. Then, the image intensity distribution at a defocus of 1 μm was optimized. That is, S for obtaining a contrast closer to the contrast at the time of defocus of 0.5 μm in the normal illumination and the normal pupil at the defocus of 1 μm.
(R) and φ (r) are obtained.

FIG. 6 shows the result of calculation of the focus margin for the pattern size in the exposure at S (r) and φ (r). From this figure, it can be seen that the depth of focus has been improved, and that the dependence on the pattern size has been reduced.

As described above, according to the present embodiment, since the amplitude transmittance of the pupil function of the projection optical system is 100%, one of the problems of the superflex method described in the section of the prior art, that is, the filter is There is no problem that the exposure light is absorbed to be turned into heat and the optical system is thermally expanded to deteriorate the transfer accuracy. In addition, as shown in FIG. 6, the dependence of the depth of focus improvement effect on the pattern size is very small, which solves another problem of the superflex method and has a very large depth of focus effect. In the above description, t
Although (r) = 1, the effect is further enhanced by optimizing t (r).

In the above embodiment, L / S = 0.6 μm
Although the contrast of the m patterns at a defocus of 1 μm has been optimized, it is also possible to optimize a pattern at an arbitrary defocus with a pattern of another size. Further, in the embodiment, L / S is optimized, but any pattern such as an isolated pattern and a contact hole pattern can be optimized. Further, the cost function is not limited to the square of the difference between the contrast C and the optimum contrast Ci, but may be any appropriate cost function.

Next, a second embodiment of the present invention will be described with reference to FIG. FIG. 7 shows MTF (Modu
2 shows an algorithm in the case of optimizing by using a lation transfer function). MTF is a function indicating the spatial frequency transfer characteristic of the optical system, and indicates how much amplitude the spatial frequency fx forming the mask pattern can pass through the optical system.

Although the calculation procedure generally follows the first embodiment,
Since the calculations in steps S3, S4, and S5 are different from those in the first embodiment, these will be described.

In step S3, MTF and M (fx) before and after grain application are calculated. The ideal (target) Mi (fx) is also calculated in advance.

In step S4, cost functions E and E 'are calculated.

[0050]

(Equation 1) Given by Here, M (fx) is the MTF calculated after applying the grain, M '(fx) is the MTF calculated before applying the grain, and Mi (fx) is the MTF that is ideal (target). Fc is a cutoff frequency.

In step S5, ΔE = EE ′ is obtained. If ΔE <0, MTF after grain application
Indirectly indicates that the MTF is closer to the ideal (target) MTF, and indicates that the given grain has contributed to asymptotically approaching the optimum value locally. Conversely, if ΔE> 0, it indicates that the given grain contributed to keep away from the optimum value when viewed locally.

According to this embodiment, although the amount of calculation is larger than that of the first embodiment, the pattern size dependence of the transfer performance improving effect is further reduced since the optimization is not performed for a pattern of a specific size. .

Although the cost relationship in the second embodiment is MTF, the OTF (Optical Transfer Functio
n) may be a cost function.

If the layer to be exposed is mainly an L / S pattern, filters 9 'and 16' (shown in FIG. 1) for providing the light source intensity distribution and the pupil function phase distribution shown in FIG. 4 may be inserted. When transferring a pattern mainly composed of an isolated element such as a contact hole layer, the pattern may deviate from the optimum condition.
In this case, at least one of the secondary light source radial direction intensity distribution, the mask complex amplitude transmittance distribution, and the pupil function radial direction complex amplitude transmittance distribution of the projection optical system is optimized again for the main mask pattern. I do. The filter 9 'or 16' is switched for each layer to be transferred, and at least one of the secondary light source radial direction intensity distribution, the mask complex amplitude transmittance distribution, and the pupil function radial direction complex amplitude transmittance distribution of the projection optical system is used. One may be changed.

Next, a third embodiment of the present invention will be described.

This apparatus is a projection exposure apparatus for projecting and exposing a pattern of a mask onto a wafer via a projection optical system. In a projection exposure apparatus, a secondary light source for illuminating the mask is provided with an intensity distribution in an emission plane of the light source from a central portion to a peripheral portion. Provide a special stop to make it larger in the part, and in the pupil plane of the projection optical system, light passing through the periphery of the pupil plane is relatively delayed or delayed with respect to the phase of light passing through other parts. A filter is installed to cause the advance.

When the radius of the secondary light source is L,
The radius d of the central portion where the intensity of the secondary light source is small satisfies d ≧ 0.5 L, and the intensity Ic of the central portion where the intensity is small satisfies Ic ≦ 0.35 Ip with respect to the peripheral portion intensity Ip.
Further, assuming that the radius of the pupil is H, the radial coordinate r of the pupil is 130 ° or more relative to the phase of light passing outside the region of 5H / 7 <r <6H / 7 or It is characterized in that the region where the advance occurs is a pupil function that exists within the range of 5H / 7 <r <6H / 7.

FIGS. 8 to 10 show the effects of this embodiment. FIG. 8 shows S (r) optimized using the above method, and FIG. 9 shows φ (r) optimized in the same manner. t (r) was set to 1 (constant). The optimization was performed so that the defocus ± 1 μm in the mask pattern having the line width of 0.6 μm and the space width of 1.8 μm and the peak intensity of the image intensity distribution in the just focus were maximized. The NA of the exposure apparatus, the coherence factor σ, and the exposure wavelength λ are 0.54, 0.5, 4 respectively.
36 nm.

FIG. 10 shows the results of calculation of the image intensity profile in exposure at S (r) and φ (r) and the image intensity profile in normal exposure. The solid line is normal exposure, the short dotted line is exposure according to the present embodiment when using a normal Cr mask, and the long dotted line is exposure according to the present embodiment when a halfton mask is used. According to the present embodiment, it can be seen that the peak intensity of the image intensity profile increases. It can also be seen that the use of a halftone mask is more effective.

Next, a fourth embodiment of the present invention will be described.

In this embodiment, when the radius of the pupil is H in the third embodiment, the radial coordinate r of the pupil is 0 ≦
The region that causes a delay or advance of 130 ° or more relative to the phase of light passing outside the region of r <H / 7 and H / 7 <r <3H / 7 is 0 ≦ r <H / 7 and H / 7 <r
The pupil function exists within the range of <3H / 7.

The effects of this embodiment are shown in FIGS. FIG. 11 shows S (r) optimized using the above method, and FIG. 12 shows φ (r) optimized in the same manner. t (r) was set to 1 (constant). The optimization was performed so that the contrast calculated from the image intensity distribution in the defocus ± 1 μm in the mask pattern of L / S = 0.4 μm and the just focus was maximized. Further, the NA of the exposure apparatus, the coherence factor σ,
The exposure wavelength λ is 0.54, 0.5, 436 nm, respectively.

FIG. 13 shows the result of calculation of the focus margin for the pattern size in the exposure at S (r) and φ (r). The short dotted line indicates normal exposure, and the solid line indicates exposure according to this embodiment when a normal Cr mask is used.
A long dotted line indicates exposure according to this embodiment when a halfton mask is used. According to the present embodiment, it is understood that the depth of focus is improved as compared with the normal exposure. It can also be seen that the use of a halftone mask is more effective. At the above NA and exposure wavelength, the line width of L / S = 0.4 μm is close to the resolution limit, and it can be seen that when the line width is optimized to increase the depth of focus, the limit resolution also improves.

Next, a fifth embodiment of the present invention will be described.

In this embodiment, when the radius of the pupil is H in the third embodiment, the radial coordinate r of the pupil is 0 ≦
Regions that cause a delay or advance of 130 ° or more relative to the phase of light passing outside the regions of r <H / 7 and H / 7 <r are 0 ≦ r <H / 7 and H / 7 < The pupil function exists within the range of r.

FIGS. 14 to 16 show the effects of this embodiment. FIG. 14 shows S (r) optimized using the above method, and FIG. 15 shows φ (r) optimized in the same manner. t (r) was set to 1 (constant). In the optimization, the defocus ± 1 μm in the mask pattern of L / S = 0.4 μm and the dark part of the image intensity distribution in the just focus are brought close to 0. Bringing the dark part of the image intensity distribution close to 0 is almost equivalent to maximizing the contrast. The NA of the exposure apparatus, the coherence factor σ, and the exposure wavelength λ are 0.54, respectively.
1,436 nm.

FIG. 16 shows the result of calculation of the focus margin with respect to the pattern size in the exposure at S (r) and φ (r). The short dotted line indicates normal exposure, and the solid line indicates exposure according to this embodiment when a normal Cr mask is used.
An alternate long and short dash line indicates exposure according to this embodiment when a halfton mask is used. According to this embodiment, it can be seen that the depth of focus is improved as compared with the normal exposure. It can also be seen that the use of a halftone mask is more effective. NA above
Also, at the exposure wavelength, the line width of L / S = 0.4 μm is close to the resolution limit, and it can be seen that when the line width is optimized to increase the depth of focus, the limit resolution also improves.

Next, a sixth embodiment of the present invention will be described.

In this embodiment, in a projection exposure apparatus for projecting and exposing a mask pattern onto a wafer via a projection optical system, the phase of light passing through the periphery of the pupil plane is defined as a pupil function of the projection optical system. Is set to generate a delay or advance relative to the phase of light passing through other portions, and when the radius of the pupil is H, the radial coordinate r of the pupil
Is a pupil function in which a region that causes a delay or advance of 130 ° or more relative to the phase of light passing through r ≦ 6H / 7 exists in the range of r> 6H / 7. It is characterized by.

FIGS. 17 and 18 show the effect of this embodiment. FIG. 17 shows φ (r) optimized using the above method. t (r) was set to 1 (constant). In the optimization, the dark part of the image intensity distribution at the defocus of ± 1 μm in the mask pattern of L / S = 0.6 μm was made to approach 0. Bringing the dark part of the image intensity distribution close to 0 is almost equivalent to maximizing the contrast. The NA of the exposure apparatus, the coherence factor σ, and the exposure wavelength λ are 0.54, 0.5, and 436 n, respectively.
m.

FIG. 18 shows the result of calculation of the focus margin with respect to the pattern size in the exposure at φ (r). Short dotted line is normal exposure, solid line is normal Cr
Exposure according to the present embodiment when a mask is used, and the dashed line indicates exposure according to the present embodiment when a halfton mask is used. According to the present embodiment, it is understood that the depth of focus is improved as compared with the normal exposure. It can also be seen that the use of a halftone mask is more effective. That is, there is a feature that the pattern size dependence of the depth of focus improvement effect is very small.

Next, a seventh embodiment of the present invention will be described.

In this embodiment, when the radius of the pupil is H in the sixth embodiment, the radial coordinate r of the pupil is 5H.
/ 7 <r <6H / 7 A region that causes a delay or advance of 130 ° or more relative to the phase of light passing through the region other than 6H / 7 is within the range of 5H / 7 <r <6H / 7. It is characterized by an existing pupil function.

FIGS. 19 and 20 show the effect of this embodiment.
Shown in FIG. 19 shows a graph obtained by optimizing φ using the above-described method.
(R). t (r) was set to 1 (constant). The optimization was performed by defocusing ± 1 in a mask pattern having a line width of 0.6 μm and a space width of 1.8 μm.
The peak intensity of the image intensity distribution at μm and just focus was set to be maximum. In addition, the N
A, coherence factor σ, and exposure wavelength λ are 0.54, 0.5, and 436 nm, respectively.

FIG. 20 shows the calculation results of the image intensity profile in the exposure at S (r) and φ (r) and the image intensity profile in the normal exposure. The solid line is normal exposure, the short dotted line is exposure according to the present embodiment when using a normal Cr mask, and the long dotted line is exposure according to the present embodiment when a halfton mask is used. According to the present embodiment, it can be seen that the peak intensity of the image intensity profile increases. It can also be seen that the use of a halftone mask is more effective.

Next, an eighth embodiment of the present invention will be described.

In this embodiment, when the radius of the pupil is H in the sixth embodiment, the radial coordinate r of the pupil is 0 ≦
The region that causes a delay or advance of 130 ° or more relative to the phase of light passing outside the region of r <H / 7 and H / 7 <r <3H / 7 is 0 ≦ r <H / 7 and H / 7 <r
The pupil function exists within the range of <3H / 7.

The effects of this embodiment are shown in FIGS. FIG. 21 shows φ optimized using the above method.
(R). t (r) was set to 1 (constant). The optimization was performed so that the contrast calculated from the image intensity distribution in the defocus ± 1 μm in the mask pattern of L / S = 0.4 μm and the just focus was maximized. The NA of the exposure apparatus, the coherence factor σ, and the exposure wavelength λ are 0.54, 0.
5,436 nm.

FIG. 22 shows the result of calculation of the focus margin with respect to the pattern size in the exposure at φ (r). Short dotted line is normal exposure, solid line is normal Cr
The exposure according to the present embodiment using a mask and the long dotted line represent the exposure according to the present embodiment using a halfton mask. According to this embodiment, it can be seen that the depth of focus is improved as compared with the normal exposure. It can also be seen that the use of a halftone mask is more effective. According to this embodiment, since the amplitude transmittance of the pupil function is 100%, one of the problems of the superflex method described in the section of the prior art, that is, the filter absorbs the exposure light to become heat, and the optical system However, there is no problem that thermal transfer causes thermal transfer to deteriorate transfer accuracy. Also, at the above NA and exposure wavelength, the line width of L / S = 0.4 μm is close to the resolution limit. Therefore, it can be seen that when the line width is optimized to increase the depth of focus, the limit resolution also improves. .

Next, a ninth embodiment of the present invention will be described.

In this embodiment, when the radius of the pupil is H in the sixth embodiment, the radial coordinate r of the pupil is 0 ≦
Regions that cause a delay or advance of 130 ° or more relative to the phase of light passing outside the regions of r <H / 7 and H / 7 <r are 0 ≦ r <H / 7 and H / 7 < The pupil function exists within the range of r.

FIGS. 23 and 24 show the effects of this embodiment.
Shown in FIG. 23 shows a graph obtained by optimizing φ using the above method.
(R). t (r) was set to 1 (constant). In the optimization, the defocus ± 1 μm in the mask pattern of L / S = 0.4 μm and the dark part of the image intensity distribution in the just focus are brought close to 0. Bringing the dark part of the image intensity distribution close to 0 is almost equivalent to maximizing the contrast. The NA, coherence factor σ, and exposure wavelength λ of the exposure apparatus are 0.54, 1.0, and 436 nm, respectively.

FIG. 24 shows the result of calculation of the focus margin with respect to the pattern size in the exposure at φ (r). Short dotted line is normal exposure, solid line is normal Cr
Exposure according to the present embodiment when a mask is used, and the dashed line indicates exposure according to the present embodiment when a halfton mask is used. According to this embodiment, it can be seen that the depth of focus is improved as compared with the normal exposure. It can also be seen that the use of a halftone mask is more effective.

According to this embodiment, since the amplitude transmittance of the pupil function is 100%, one of the problems of the superflex method described in the section of the prior art, that is, the filter absorbs the exposure light and becomes heat. In addition, there is no problem that the optical system causes thermal expansion and deteriorates the transfer accuracy. Also, at the above NA and exposure wavelength, the line width of L / S = 0.4 μm is close to the resolution limit. Therefore, it can be seen that when the line width is optimized to increase the depth of focus, the limit resolution also improves. .

The present invention is not limited to the above embodiments. In the embodiment, the light source intensity distribution and the pupil function are optimized for the purpose of improving the focal seismic intensity. However, the optimization can be performed for the purpose of improving other evaluation amounts such as the limit resolution or the dimensional accuracy. It is.

In the above embodiment, the simulated annealing method was used as the optimization algorithm, but this does not limit the present invention in any way.
S method (Dumped Least Squre Method), GA (Genetic Al
gorithm), Simplex method, or any other appropriate optimization algorithm.

In the above embodiment, the N of the projection optical system
A (aperture ratio), σ of illumination system (coherence factor)
Is determined, and the secondary light source radial intensity distribution S (r),
Pupil function radial direction complex amplitude transmittance distribution t (r), φ (r)
Has been optimized, it is also possible to optimize the NA of the overlay projection optical system, the σ of the illumination system, and the complex amplitude transmittance distribution of the reticle pattern. It is also possible to assume a phase shift mask such as a Levenson type phase shift mask, a halftone mask, a self-alignment type phase shift mask as the mask.

The method of the present invention based on the above-described embodiments is not limited to the improvement of the transfer performance of the projection optical system for semiconductor manufacturing.
It is possible to improve the performance of all optical devices configured based on a partial coherent optical system, such as an optical microscope and an optical drawing device.

[0089]

As described above in detail, according to the present invention, the secondary light source radial intensity distribution of the projection optical system and the mask radial direction are adjusted so that a desired depth of focus can be obtained at a desired pattern size. Since at least one of the complex amplitude transmittance distribution and the pupil function radial amplitude complex amplitude transmittance distribution of the projection optical system is optimized, the depth of focus and the pattern resolution performance can be optimized, and the exposure accuracy can be improved. It becomes possible to measure.

[Brief description of the drawings]

FIG. 1 is a schematic configuration diagram showing a projection exposure apparatus according to a first embodiment of the present invention;

FIG. 2 is a schematic diagram showing an algorithm for optimizing a secondary light source and a pupil function in the first embodiment;

FIG. 3 is a schematic view showing a state where a grain having a size of ε is given to one portion of an initial distribution;

FIG. 4 is a characteristic diagram showing changes in secondary light source intensity and pupil phase with respect to the distance from the optical axis at optimized S (r) and φ (r);

FIG. 5 is a characteristic diagram showing a relationship between a defocus amount and a contrast in the embodiment.

FIG. 6 is a characteristic diagram showing a relationship between a pattern size and defocus in the embodiment.

FIG. 7 shows MTF as a cost function in the second embodiment.
Schematic diagram showing an algorithm when optimizing using

FIG. 8 is a diagram showing optimized S (r) in the third embodiment;

FIG. 9 is a diagram showing an optimized φ (r) in the third embodiment;

FIG. 10 is a diagram showing a result of calculating an image intensity profile in the third embodiment.

FIG. 11 is a diagram showing optimized S (r) in a fourth embodiment;

FIG. 12 is a diagram showing an optimized φ (r) in the fourth embodiment;

FIG. 13 is a diagram showing a result of calculating a focus margin in the fourth embodiment.

FIG. 14 is a diagram showing optimized S (r) in the fifth embodiment;

FIG. 15 is a diagram showing an optimized φ (r) in the fifth embodiment;

FIG. 16 is a diagram showing a result of calculating a focus margin in the fifth embodiment;

FIG. 17 is a diagram showing an optimized φ (r) in the sixth embodiment;

FIG. 18 is a diagram showing a result of calculating a focus margin in the sixth embodiment.

FIG. 19 is a diagram showing an optimized φ (r) in the seventh embodiment;

FIG. 20 is a diagram showing a result of calculating an image intensity profile in the seventh embodiment;

FIG. 21 is a diagram showing an optimized φ (r) in the eighth embodiment;

FIG. 22 is a diagram showing a result of calculating a focus margin in the eighth embodiment.

FIG. 23 is a diagram showing an optimized φ (r) in the ninth embodiment;

FIG. 24 is a diagram showing a result of calculating a focus margin in the ninth embodiment;

FIG. 25 is a schematic configuration diagram showing a conventional projection exposure apparatus.

FIG. 26 is a diagram for explaining a problem of the conventional device;

FIG. 27 is a diagram showing an example of a filter used in place of an aperture stop,

FIG. 28 is a characteristic diagram showing a relationship between a pattern size and a depth of focus in a conventional device;

FIG. 29 is a characteristic diagram showing a relationship between a pattern size and defocus in a conventional device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 ... Lamp, 2 ... Elliptical reflecting mirror, 3 ... Second focus of elliptical reflecting mirror 2, 4 ... Input lens, 5 ... Optical integrator, 6 ... Output lens, 7 ... Collimation lens, 8 ... Reticle (mask), 9 ' ... Special aperture, 10 ... Filter, 11 and 12 ... Cold mirror, 13 ... Lamp house, 14 ... Projection optical system, 15 ... Wafer, 16 '... Filter.

Continuation of front page (56) References JP-A-61-91662 (JP, A) JP-A-61-276217 (JP, A) JP-A-49-53449 (JP, A) JP-A-4-348017 (JP) JP-A-5-102006 (JP, A) JP-A-4-369208 (JP, A) JP-A-4-267515 (JP, A) JP-A-4-251914 (JP, A) 1991 (Heisei 9) 3) Spring of the 38th Annual Meeting of the Japan Society of Applied Physics Related Proceedings, Volume 2 (issued March 28, 1991) 29a-ZC-8 (58) Fields surveyed (Int. Cl. ⁷ , DB name) H01L 21/027

Claims (8)

(57) [Claims] In a projection exposure apparatus for projecting and exposing a mask pattern onto a wafer via a projection optical system, a depth of focus is maximized according to a desired pattern size.
As a secondary light source for illuminating the mask, a special stop for increasing the intensity distribution in the emission surface of the light source in the peripheral part from the central part is provided, and the transmittance of light is provided on the pupil plane of the projection optical system. Change
No, the projection exposure apparatus characterized by comprising placing a filter for varying the optical light path length along the radial direction around the optical axis with respect to light passing through the pupil plane. 2. The filter according to claim 1, wherein light passing through a peripheral portion of a pupil plane of the projection optical system has a longer optical path length than light passing through other portions. 2. The projection exposure apparatus according to claim 1, wherein: 3. The filter according to claim 1, wherein light passing through a peripheral portion of a pupil plane of the projection optical system has a shorter optical path length than light passing through other portions. 2. The projection exposure apparatus according to claim 1, wherein: 4. The filter according to claim 1, wherein light passing through a peripheral portion and a central portion of a pupil plane of the projection optical system has an optical path length longer than light passing through other portions. 2. The projection exposure apparatus according to claim 1, wherein 5. The filter according to claim 1, wherein light passing through a peripheral portion and a central portion of a pupil plane of the projection optical system has a shorter optical path length than light passing through other portions. 2. The projection exposure apparatus according to claim 1, wherein 6. Each time the mask used for transfer is changed ,
2. The projection exposure apparatus according to claim 1, further comprising means for changing at least one of a secondary light source intensity distribution and an optical path length in the filter. 7. A projection exposure apparatus for projecting and exposing a mask pattern partially illuminated by coherent light onto a wafer through a projection optical system, wherein a secondary light source illuminating the mask is an emission surface of the light source. A special stop is provided to make the inner intensity distribution larger in the peripheral part than in the central part, and light transmittance is provided on the pupil plane of the projection optical system.
A filter that causes a relative delay or advance of the phase of light passing through the periphery of the pupil plane with respect to light passing through other portions without changing the radius of the secondary light source. Is L, the radius d of the central portion where the intensity of the secondary light source is small satisfies d ≧ 0.5 L, and the intensity lc of the central portion where the intensity is small is lc ≦ lc with respect to the peripheral intensity lp.
0.35 × lp and the radius of the pupil plane is H, the radial coordinate r of the pupil is 130 relative to the phase of light passing through the region of r ≦ 6H / 7. A projection exposure apparatus characterized in that an area in which a delay or advance of more than ° is generated is a pupil function that exists within the range of r> 6H / 7. 8. The mask comprises a light-transmitting substrate and a pattern of a translucent material disposed on the light-transmitting substrate.
The phase difference of the translucent material with respect to the translucent substrate is 180 × (2n + 1) ± 10 (degrees): n is an integer, and the transmissivity of the translucent material is the amplitude transmission T of the translucent substrate. The projection exposure apparatus according to claim 1, wherein the ratio T0 satisfies 0.01 × T0 ≦ T ≦ 0.16 × T0.