JP3381257B2 - Projection exposure method - 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 method for illuminating a first object with light and reducing and projecting the illuminated pattern of the first object onto a substrate or the like as a second object. Especially, the reticle (mask) as the first object
The present invention relates to a projection exposure method suitable for projecting and exposing a circuit pattern formed on a substrate (wafer) as a second object.

[0002]

2. Description of the Related Art In recent years, as the pattern of integrated circuits has become finer, the performance required for a projection exposure apparatus used for printing a wafer has become more and more severe. Under such circumstances, the projection optical system is required to have high resolution, flatness of the image plane, and little distortion (hereinafter referred to as distortion). For these reasons, in addition to shortening the exposure wavelength λ, it has been attempted to increase the numerical aperture NA of the projection optical system, reduce the field curvature, and reduce distortion. Examples of such a method include JP-A-4-157512 and JP-A-5-17306.
There are things such as No. 5.

Further, as a method for adjusting only the magnification error, Japanese Patent Laid-Open Nos. 59-144127 and 62-356.
There is No. 20. In the former, it has been proposed that a film that does not affect the image performance, such as a pellicle, is very thin and is arranged in the optical path by curving it. In the latter, a rotationally symmetric plano-convex lens, or a rotationally symmetric plano-convex lens and a plano-concave lens are proposed. It is proposed to move the group of in the optical axis direction to adjust the overall magnification on the wafer surface isotropically.

[0004]

However, the high-performance projection optical system proposed in each of Japanese Patent Application Laid-Open Nos. 4-157412 and 5-173065 has a number of lenses of 15 to 24. In the case of a high-resolution projection optical system having a numerical aperture NA of 0.4 or more, the number of constituent elements is significantly increased to 20 or more. In this way, as the required performance becomes stricter, the number of projection optical systems is increasing and the structure is becoming extremely complicated. Therefore, we actually manufactured these projection optics,
In order to achieve high performance by installing it in a projection exposure apparatus and suppressing aberrations such as field curvature, astigmatism, and distortion as designed values, the accuracy of individual lens parts and the accuracy of assembly are extremely strict. Therefore, there is a problem in that the yield is bad, the manufacturing schedule is very long, or sufficient performance cannot be exhibited.

Further, in the method of correcting the magnification error proposed in Japanese Patent Laid-Open No. 59-144127, an extremely thin film or the like which does not affect the image forming performance of the optical system is curved and its prism action is used. Although the magnification error is corrected, it is not possible to finely adjust the correction amount and the correction direction of the directional asymmetric magnification error component remaining in the projection optical system. Moreover, since the thin film is used, it can be two-dimensionally held by being attached to a metal frame or the like when the exposure area is elongated as in the mirror projection method, but when the exposure area is rectangular or square. It is very difficult to hold such a thin film three-dimensionally and exhibit good reproducibility. Also, even if glass or the like is used instead of a thin film to maintain the shape, it is still difficult to make them thin and uniform so as not to affect the imaging performance. Durability of the film including damage accidents due to heat absorption of exposure light when the film is actually used, heat absorption of exposure light and changes in optical performance due to environmental changes are very problematic. .

Further, in Japanese Patent Laid-Open No. 62-35620, it is proposed to use a rotationally symmetric lens to adjust the magnification error. However, if the rotationally symmetric lens is moved in the optical axis direction, the wafer surface can be adjusted. It is possible to adjust only the entire magnification in 1) isotropically, and it is not possible to adjust the directional asymmetric magnification error component remaining in the projection optical system.

Further, in the method of correcting the magnification error proposed in JP-A-59-144127 and JP-A-62-356620, only the magnification error can be basically corrected, and the off-axis aberration It is not possible to correct astigmatism and the like, and it is also difficult to deal with a magnification error component and a distortion aberration component that are rotationally asymmetric and locally remain randomly in the projection optical system.

The present invention has been made in view of the above problems, and has rotationally asymmetric optical characteristics of the projection optical system, such as rotationally asymmetric off-axis aberration components (astigmatism components, field curvature, etc.), Its main purpose is to correct rotationally asymmetric magnification error components and the like.

[0009]

To achieve the above object, the projection exposure method of the first invention is a projection exposure method for projecting a pattern image of a reticle onto a substrate, wherein the projection optical system is rotationally asymmetric. Calculating a relative amount of rotation between two optical elements having rotationally asymmetric optical characteristics in the projection optical system to correct the optical characteristics; and performing the relative rotation based on the calculation result. And setting the two optical elements so as to obtain a quantity.

The projection exposure method of the second invention is characterized in that the two optical elements have rotationally asymmetric powers.

The projection exposure method of the third invention is characterized in that the two optical elements have different powers in directions orthogonal to each other.

The projection exposure method of the fourth invention is characterized in that each of the two optical elements has a predetermined refractive index distribution.

Further, the projection exposure method of the fifth invention further comprises a step of measuring an optical characteristic rotationally asymmetric with respect to the optical axis of the projection optical system, and the relative rotation amount is measured. It is characterized in that it is calculated based on rotationally asymmetric optical characteristics.

The projection exposure method of the sixth invention is characterized in that the pattern image of the reticle is projected onto the substrate after setting the two optical elements.

DETAILED DESCRIPTION OF THE INVENTION As shown in FIG. 1, a cylindrical lens 1 having a negative refractive power, which is a kind of toric lens having different powers in orthogonal directions, has a meridional direction (y).
The focal length in the y ′ plane direction is f1, the distance from the cylindrical lens 1 to the reticle surface 4 (xy plane) as the first object is d11, and the center position of the reticle surface 4 (reticle surface and optical axis A
When the object point is the position where x intersects with x, and the image point (virtual image) position formed between the object point (reticle surface 4) and the cylindrical lens 1 by the cylindrical lens 1 is d12, this cylindrical lens 1 The image forming magnification β1 in the Y direction (the plane direction including the optical axis Ax and the Y axis) rotated by θ from the y axis and the distance d12 from the cylindrical lens 1 to the image point position (hereinafter, simply referred to as the image forming position). Is as shown below. Although not shown in FIG. 1, a projection optical system for projecting the pattern of the reticle onto the wafer is arranged on the side opposite to the reticle surface 4 with respect to the cylindrical lens. Is also the same.

Β1 = f1 / (d11 · cos ² θ + f1) (1) d12 = d11 · f1 / (d11 · cos ² θ + f1) (2) Similarly, in the X direction (optical axis Ax and X axis) orthogonal to the Y direction. The imaging magnification β1 ′ and the imaging magnification d12 ′ (in the plane direction including and) are as shown below.

Β1 ′ = f1 / (d11 · sin ² θ + f1) (3) d12 ′ = d11 · f1 / (d11 · sin ² θ + f1) (4) Therefore, the astigmatism amount AS1 is AS1 = d12−d12 ′ ( 5) is given in.

Therefore, if the cylindrical lens 1 is moved,
Since d11 in the expressions (1) to (4) changes, (5)
It can be understood from the equation that the astigmatism amount changes and the magnifications of the equations (1) and (3) change.

On the other hand, if the cylindrical lens 1 is rotated,
Since θ in the expressions (1) to (4) changes, the astigmatism amount changes from the expression (5), and the expressions (1) and (3) are changed.
It can be seen that the scale factor of the equation changes.

Further, as shown in FIG. 2, the focal length in the meridional direction (the yy'plane direction) of the cylindrical lens 2 having a positive refracting power, which is one type of toric lens, is f2, and the focal length is from the cylindrical lens 2 to the first object. Is the distance to the reticle surface 4 (xy plane) as d, and the image point position formed by the cylindrical lens 2 when the central position of the reticle surface 4 (the position where the reticle surface and the optical axis Ax intersect) is the object point. When d22, Y rotated by θ from the y-axis by this cylindrical lens 2
Imaging magnification β in the direction (plane direction including the optical axis Ax and the Y axis)
2 and the distance from the cylindrical lens 2 to the image point position d22 (hereinafter, simply referred to as an image forming position) are as follows.

Β2 = f2 / (d21 · cos ² θ + f2) (6) d22 = d21 · f2 / (d21 · cos ² θ + f2) (7) Similarly, the X direction orthogonal to the Y direction (optical axis Ax and X axis) The imaging magnification β2 ′ and the imaging magnification d22 ′ (in the plane direction including and) are as shown below.

Β2 ′ = f2 / (d21 · sin ² θ + f2) (8) d22 ′ = d21 · f2 / (d21 · sin ² θ + f2) (9) Therefore, the astigmatism amount AS2 is AS2 = d22−d22 ′ ( 10) is given.

Therefore, if the cylindrical lens 2 is moved,
Since d21 in the expressions (6) to (10) changes, (1
It can be understood from the expression (0) that the astigmatism amount changes and the magnifications of the expressions (6) and (8) change.

On the other hand, if the cylindrical lens 2 is rotated,
Since θ in the expressions (6) to (9) changes, it can be understood from the expression (10) that the amount of astigmatism changes and the magnifications of the expressions (6) and (8) change.

Now, AS1 and AS2 shown in the above equations (5) and (10) are the respective cylindrical lenses (1,
The astigmatism amount can be corrected by 2).

The best image planes at that time are given by (d12 + d12 ') / 2 (11) (d22 + d22') / 2 (12), and the best image plane changes depending on d11, d21, and θ. Therefore, it can be seen that the amount of field curvature also changes.

As described above, in order to adjust the amount and direction of the imaging magnification, astigmatism, and field curvature, a toric lens such as a cylindrical lens is moved in the optical axis direction, or a toric lens such as a cylindrical lens is used. It is understood that it is sufficient to rotate the lens itself. Note that the focal length of the toric lens itself may be changed in addition to the adjustment method described above.

When θ = 0 in order to estimate the maximum correction amount of astigmatism when using the cylindrical lens 2 shown in FIG. 2, the maximum astigmatism at that time is as follows.

AS2 _max = − (d21) ² / (d21 + f2) (13) Now, when the distance from the reticle as the first object to the wafer as the second object is L, a line width of 10 μm or less is printed. As a result of repeated trial printing on the projection exposure apparatus, it was found that the maximum astigmatism amount AS2 _max to be corrected should be 10 ⁻⁵ L or less.

Therefore, if d21 ≦ 10 ⁻² L, (1
From equation (3), | f2 | ≧ 10L (14), and it is desirable that the focal length of the positive cylindrical lens 2 satisfies the range of equation (14).

When combining two or more toric lenses such as the cylindrical lenses shown in FIGS. 1 and 2 or combining with other optical elements, the reticle 4 is used.
The object point of is the image formation position in the direction of interest formed by the first toric lens or other optical element as a new object point, and from this new object point to the next toric lens or another optical element. Recalculate the distance to the element and d1
It may be set to 1 or d21.

Now, consider a case where the negative cylindrical lens 1 shown in FIG. 1 and the positive cylindrical lens 2 shown in FIG. 2 are arranged in series along the optical axis direction.

Now, when the generatrix directions of the two cylindrical lenses (1, 2) coincide with each other and the product of the imaging magnifications of the two cylindrical lenses is 1, that is, | β1 · β2 | = 1, each direction The combined power of the two cylindrical lenses (1, 2) is almost zero, and the optical characteristics such as magnification and off-axis aberration (astigmatism, curvature of field, etc.) do not change at all.

On the other hand, when the generatrix directions of the two cylindrical lenses (1, 2) are orthogonal to each other, the maximum magnification and the maximum off-axis aberration can be generated.

Therefore, if the two cylindrical lenses (1, 2) are rotated relative to each other, the correction amount and correction of the directional asymmetric magnification error component and off-axis aberration component remaining in the projection optical system are corrected. It can be seen that the adjustment for the direction can be realized.

When two negative cylinder lenses 1 shown in FIG. 1 are arranged in series along the optical axis direction, or when two positive cylinder lenses 2 shown in FIG. 2 are arranged along the optical axis direction. When two cylindrical lenses are arranged in series, the maximum magnification and the maximum off-axis aberration can be generated when the generatrix directions of the cylindrical lenses coincide with each other, and the generatrix directions of the cylindrical lenses are orthogonal to each other. Then, it is possible to give the same lens action as that of the substantially one rotationally symmetric spherical lens.

As described above, by using at least two cylindrical lenses, which are one type of toric lens, and making at least one of the cylindrical lenses rotatable, magnification and off-axis aberrations (astigmatism, field curvature, etc.) can be obtained. It is possible to arbitrarily adjust the amount and direction of optical characteristics such as).

In the above, adjustment regarding astigmatism and field curvature has been mainly described. Next, the negative cylindrical lens 1 shown in FIG. 1 or the positive cylindrical lens 2 shown in FIG. The adjustment of the magnification error when rotated about the optical axis Ax will be described in detail with reference to FIGS.

FIG. 3 shows a state in which a parallel light beam having a radius R centered on the optical axis Ax is incident on the negative cylindrical lens 1 shown in FIG. Here, in FIG. 3, the optical axis A
A parallel light beam with a radius R centered at x is formed on the reticle surface 4 (xy
The locus when passing through the plane) is shown as a circle 13, and the optical axis A
An ellipse 11 indicates a locus when a parallel light flux having a radius R centered on x and having a diverging action by the cylindrical lens 1 passes through a virtual plane (x'y 'plane). Also,
FIG. 5 shows the state of the light beam diameter on the virtual plane (x'y 'plane) shown in FIG.

On the other hand, FIG. 4 shows a state in which a parallel light beam having a radius R centered on the optical axis Ax is made incident on the positive cylindrical lens 2 shown in FIG. Here, in FIG.
A collimated light beam with a radius R centered on the optical axis Ax forms a reticle surface 4
A locus when passing through the (xy plane) is shown as a circle 13,
A parallel light beam with a radius R centered on the optical axis Ax causes the cylindrical lens 2
An ellipse 12 indicates a locus when the light flux that has undergone the converging action by passes through a virtual plane (x'y 'plane). Further, FIG. 6 shows the state of the light flux diameter on the virtual plane (x'y 'plane) shown in FIG.

The ellipse 11 in FIG. 3 and the ellipse 12 in FIG. 4 rotate with the rotation of the cylindrical lens (1, 2) about the optical axis.

As shown in FIG. 7, the luminous flux in the y'direction (plane direction including the optical axes Ax and y'axis) which is the meridional direction on the virtual plane (x'y 'plane) by the negative cylindrical lens 1. When the change amount of the diameter is ΔR1, and the distance from the negative cylindrical lens 1 to the virtual plane (x′y ′ plane) is e1, the following relationship holds.

ΔR1 = −R · e1 / f1 (15) Similarly, as shown in FIG. 8, y ′, which is the meridional direction on the virtual plane (x′y ′ plane) by the positive cylindrical lens 2.
When the change amount of the light beam diameter in the direction (plane direction including the optical axis Ax and y ′ axis) is ΔR2, the distance from the positive cylindrical lens 2 to the virtual plane (x′y ′ plane) is e2 The relationship is established.

ΔR2 = −R · e2 / f2 (16) Therefore, as shown in FIGS. 5 and 6, the virtual plane (x ′
The diameter in the y'direction (half the major axis in FIG. 5, half the minor axis in FIG. 6) indicated by the solid line on the y'plane) is y '= R (1-e1 / f1) (17) y' = R (1-e2 / f2) (18), the circle formula, that is, y = ± [R ² + (x ′) ² ] ^0.5 (19) is substituted, and the x ′ y ′ coordinates are converted into xy coordinates. When converted, the ellipses 11 and 12 shown by the solid lines in FIGS. 5 and 6 can be expressed as follows, respectively.

X ² / R ² + y ² / [(1-e1 / f1) · R] ² = 1 (20) x ² / R ² + y ² / [(1-e2 / f2) · R] ² = 1 (21)

In this way, for example, as shown in FIG.
6 has an asymmetrical magnification error, the cylindrical lens 1 of FIG. 3 having the optical characteristics of FIG. 5 is rotated so that the luminous flux diameter as shown in FIG. Since it can be changed, an asymmetrical magnification error can be adjusted. On the contrary, when the projection optical system has an asymmetrical magnification error as shown in FIG. 5, for example, by rotating the cylindrical lens 2 of FIG. 4 having the optical characteristics as shown in FIG. Since the luminous flux diameter as shown can be arbitrarily changed from the ellipse to the circle, an asymmetric magnification error can be adjusted.

Here, the negative cylindrical lens 1 as shown in FIG.
When the distance from the reticle as the first object to the wafer as the second object is L, a projection exposure apparatus that prints a line width of 10 μm or less was subjected to trial printing and repeated studies. As a result, it was found that the maximum correction amount for the magnification error should be 10 ⁻⁴ (= 100 ppm) or less.

Further, when the above equation (1) showing the relationship between the focal length f1 of the cylindrical lens 1 and the magnification β1 of the cylindrical lens 1 is modified, the following equation is obtained.

F1 = (− d11 · β1) / (β1-1) (23) Therefore, the maximum magnification error correction amount 10 ⁻⁴ (= 100 pp
Converting m) to β1, β1 = 0.9999 (or 1.000)
1) and if d11 ≦ 10 ⁻² L, then from equation (23), | f2 | ≧ 10 ² L (24), and the focal length of the positive cylindrical lens 2 satisfies the above equation (24) range. Is desirable.

In the above description, the example in which one toric lens (cylindrical lens) is rotated about the optical axis direction to correct the magnification error is described. However, one toric lens (cylindrical lens) is rotated in the optical axis direction. It is clear from the equations (1), (3), (6) and (8) that the magnification error can be corrected by moving. In this case, (2
It is more preferable to satisfy the expression (4).

By the way, although it has been described above that the magnification error can be corrected by using one toric lens (cylindrical lens), at least one of the toric lenses is used, and at least one of them is used. By making the cylindrical lens rotatable, it is possible to arbitrarily adjust the amount and direction of optical characteristics such as magnification error.

Therefore, the negative cylindrical lens 1 shown in FIG.
2 and the positive cylindrical lens 2 shown in FIG. 2 may be arranged in series along the optical axis direction of the projection optical system, and they may be relatively rotated. In this case, since the negative cylindrical lens 1 has the optical characteristics shown in FIG. 5 and the positive cylindrical lens 1 has the optical characteristics shown in FIG. 6, it is formed by these cylindrical lenses (1, 2). The luminous flux diameter is as shown in FIGS.
It can be understood that the light flux diameters can be arbitrarily changed from the ellipse to the circle by synthesizing the light flux diameters shown in (3) and rotating these relatively, and asymmetrical magnification error can be corrected.

Further, when the projection optical system has an asymmetrical magnification error as shown in FIG. 5 or 6, for example, at least two or more cylindrical lenses are arranged in series along the optical axis direction. However, if at least one of the cylindrical lenses is rotatably provided, the light beam diameter as shown in FIG. 5 or 6 can be arbitrarily changed from the ellipse to the circle, so that the asymmetric magnification error is adjusted. can do.

When two or more toric lenses (cylindrical lenses) are combined, or when combined with other optical elements, the focused light flux is the first toric lens (cylindrical lens) or another optical element. The light beam that has passed through the element may be traced as a new light beam that has entered the next toric lens (cylindrical lens) or the like.

When two negative toric lenses (cylindrical lenses) are combined and a negative cylindrical lens 1 as shown in FIG. 1 and a positive cylindrical lens as shown in FIG. 2 are installed close to each other, the generatrix direction of each lens is When they match, the total lens power in each direction becomes almost 0, and the light flux shape does not change. However, when the generatrix directions of the respective lenses are orthogonal, the shape change becomes the maximum.

Further, when two negative cylinder lenses 1 shown in FIG. 1 are arranged in series along the optical axis direction, or two positive cylinder lenses 2 shown in FIG. 2 are arranged along the optical axis direction. When two cylindrical lenses are arranged in series, the maximum magnification and the maximum off-axis aberration can be generated when the generatrix directions of the cylindrical lenses coincide with each other, and the generatrix directions of the cylindrical lenses are orthogonal to each other. Then, it is possible to give the same lens action as that of the substantially one rotationally symmetric spherical lens.

As described above, by using at least two cylindrical lenses, which are one type of toric lens, and making at least one of the cylindrical lenses rotatable, magnification and off-axis aberrations (astigmatism, field curvature, etc.) can be obtained. It is possible to arbitrarily adjust the amount and direction of optical characteristics such as).

The above equations (14) and (24)
Expressing the expression in a general form, the focal length of the cylindrical lens that can effectively act on the astigmatism correction is fA, and the focal length of the cylindrical lens that can effectively act on the correction of the magnification error is fD.
Then, | fA | ≧ 10L (25) | fD | ≧ 10 ² L (26), and it is desirable to satisfy the above equation (25) for effective correction of astigmatism, and effective magnification error. It is desirable to satisfy the above equation (26) in order to correct However, the focal length (fA, fD) of the cylindrical lens in this case
Can be applied not only to a single cylindrical lens, but also to a combination of a toric lens such as a plurality of cylindrical lenses or a toric reflection member. That is, the focal length (fA, fD) of this cylindrical lens is the composite focal length of the plurality of cylindrical lenses when a plurality of toric optical members are combined.

If the equation (25) or the equation (26) is not satisfied, the toric component is too strong, which affects other aberrations, causing a problem. For example, correction of astigmatism causes deterioration of field curvature and magnification error, and correction of magnification error causes deterioration of telecentricity and astigmatism. Therefore, within the above range, asymmetrical aberration can be effectively corrected.

By the way, the above formulas (25) and (26) show the range of the optimum focal length of the toric optical member. Next, from another viewpoint, the range of the optimum focal length of the toric optical member will be described. consider.

First, in FIG. 9, the projection optical system has an aperture stop S.
With the reticle 4 side being the front group G F and the wafer 5 side being the rear group G.
A configuration with R is shown, where the front group GF is f _GF
And the rear group GR has a focal length of f _GR , and the projection optical system is telecentric on the reticle side and the wafer 5 side.

FIG. 10 shows the front group G of the projection optical system shown in FIG.
The figure shows a state in which a cylindrical lens having a positive power as a toric optical member is arranged between F and the reticle 4, and the power of this cylindrical lens 2 is in the paper surface direction (meridional direction) of FIG. .

Here, as shown in FIG. 10, the cylindrical lens 2
Let f2 be the focal length of the lens, and let D ₁ be the distance between the cylindrical lens 2 and the front lens group GF (the distance between the principal points of both optical systems). ₁ is
The following relationship holds.

F ₁ = (f2 · f _GF ) / (f2 + f _GF −D ₁ ) (27) Further, the imaging magnification B _{1 of} the projection optical system (GF, GR) is set as
Assuming that the image forming magnification B ₁ 'in the combined system of the cylindrical lens 2 and the projection optical system (GF, GR), the following relationship is established.

B ₁ = −f _GR / f _GF (28) B ₁ ′ = −f _GR / F ₁ = B ₁ [1+ (f _GF −D ₁ ) / f 2] (29) Therefore, the sagittal of the projection optical system The difference in magnification ΔB ₁ between the direction and the meridional direction is as follows.

ΔB ₁ = B ₁ ′ −B ₁ = B ₁ (f _GF −D ₁ ) / f 2 (30) On the other hand, the principal point on the reticle side by the combined system of the cylindrical lens 2 and the front group GF is H ₁ , The focus position on the reticle side by the combined system of the cylindrical lens 2 and the front lens group GF is P ₁ , and its focus position P
Assuming that the distance between ₁ and the reticle 4 is Δs ₁ , and the distance between the image formation position Q ₁ of the reticle 4 by the combined system of the cylindrical lens 2 and the projection optical system (GF, GR) to the wafer 5 is Δs ₁ ′, The relationship is established.

Δs ₁ = (f _GF −D ₁ ) ² / (f ² + f _GF −D ₁ ) (31) Δs ₁ ′ = (B ₁ ′) ² · Δs ₁ (32) where Δs ₁ ′ is projection optics It means the difference between the image formation positions in the sagittal direction and the meridional direction of the system, that is, the amount of astigmatism (astigmatic difference).

When the numerical aperture on the reticle side of the projection optical system is NA _R and the wavelength of the exposure light is λ, the depth of focus DOF _R on the reticle side of the projection optical system is as follows.

DOF _R = λ / (NA _R ) ² (33) Therefore, in order to suppress the amount of astigmatism within the depth of focus on the reticle side of the projection optical system, the above equations (31) and (3) are used.
The following equation is derived from the equation 3).

F2 ≧ − (f _GF −D ₁ ) + [(NA _R ) ² (f _GF −D ₁ ) ² ] / λ (34) Therefore, the cylindrical lens 2 is configured to satisfy the expression (34). The astigmatism amount can be suppressed within the depth of focus.

If this expression (34) is generally expressed, the power difference in the orthogonal direction of the toric optical member is Δf.
Then, it becomes as follows. Δf ≧ | − (f _GF −D ₁ ) + [(NA _R ) ² (f _GF −D ₁ ) ² ] / λ | (35) As described above, when the toric optical member is used, It is understood that it is preferable to satisfy the above expression (35) in order to suppress the amount of point aberration within the depth of focus on the reticle side of the projection optical system. It goes without saying that the relationships of the above equations (34) and (35) are established even when the projection optical system has the same magnification, reduction or enlargement magnification.

As an example, the numerical aperture NA _R on the reticle side of the projection optical system is 0.1, the wavelength λ of the exposure light is 436 nm, and f
_{GF = 250mm, f GR = 250mm} , D 1 = 200mm
Then, from the above formula (34), the focal length f2 of the cylindrical lens in the meridional direction is generally the above (3
From the equation (5), the power difference Δf in the direction orthogonal to the toric optical member is 5.7 × 10 ⁴ mm or more, and the variable magnification correction amount (magnification difference ΔB ₁ ) at this time is 87.
It becomes 0 ppm (= 8.7 × 10 ⁻⁴ ) or less.

In the above description, the equation (35) is derived on the premise that the toric optical member is arranged between the reticle and the projection optical system. Since the same relation is established even when they are arranged between the two, it is desirable to satisfy the following relation in this case.

Δf ≧ | − (f _GR −D ₁ ′) + [(NA _W ) ² (f _GR −D ₁ ′) ² ] / λ | (36) where NA _W is the wafer side of the projection optical system. Is the numerical aperture of
D ₁ 'is a distance between the toric optical member and the rear group GR (distance between principal points of both optical systems).

Next, referring to FIG. 11, when the positive cylindrical lens 2 is arranged between the front lens group GF and the rear lens group GR in the projection optical system, in other words, in the vicinity of the aperture stop S. The optimum focal length range of the cylindrical lens 2 will be examined. Figure 11
Shows a state in which a cylindrical lens 2 having a positive power as a toric optical member is arranged between the front group GF and the rear group GR of the projection optical system shown in FIG. Is in the paper surface direction (meridional direction) of FIG.

Here, as shown in FIG. 11, the cylindrical lens 2
Let f2 be the focal length of F, and let D ₂ be the distance between the front lens group GF and the cylindrical lens 2 (distance between the principal points of both optical systems). ₂ is
The following relationship holds.

F ₂ = (f2 · f _GF ) / (f2 + f _GF −D ₂ ) (37) Further, the imaging magnification B _{2 of} the projection optical system (GF, GR)
Assuming that the image forming magnification B ₂ 'in the combined system of the cylindrical lens 2 and the projection optical system (GF, GR), the following relationship is established.

B ₂ = −f _GR / f _GF (38) B ₂ ′ = −f _GR / F ₂ = B ₂ [1+ (f _GF −D ₂ ) / f ₂ ] (39) Therefore, the sagittal of the projection optical system The difference in magnification ΔB ₂ between the direction and the meridional direction is as follows.

ΔB ₂ = B ₂ ′ −B ₂ = B ₂ (f _GF −D ₂ ) / f ₂ (40) On the other hand, the principal point on the reticle side by the combined system of the front lens group GF and the cylindrical lens 2 is H ₂ , The focus position on the reticle side by the combined system of the front lens group GF and the cylindrical lens 2 is P ₂ , and its focus position P
₂ to the reticle 4 is Δs ₂ , the projection optical system (GF
, GR) and a cylindrical lens 2 combined system 4
If the distance from the image forming position Q ₂ to the wafer 5 is Δs ₂ ′, the following relationship holds.

Δs ₂ = (f _GF ) ² / (f ² + f _GF −D ₂ ) (41) Δs ₂ ′ = (B ₂ ′) ² · Δs ₂ (42) where Δs ₂ ′ is the sagittal of the projection optical system. Direction and the meridional direction, the difference between the imaging positions, that is, the amount of astigmatism (astigmatic difference).

Therefore, in order to suppress the amount of astigmatism within the depth of focus on the reticle side of the projection optical system, the above equation (33) is used.
The following equation is derived from the equation and the equation (41).

F2 ≧ − (f _GF −D ₂ ) + [(NA _R ) ² (f _GF ) ² ] / λ (43) Therefore, the cylindrical lens 2 can be configured to satisfy the expression (43). Preferably, this makes it possible to suppress the amount of astigmatism within the depth of focus.

If this equation (43) is generally expressed, the power difference in the orthogonal direction of the toric optical member is Δf.
Then, it becomes as follows.

Δf ≧ | − (f _GF −D ₂ ) + [(NA _R ) ² (f _GF ) ² ] / λ | (44) Thus, when the toric optical member is used, It is understood that it is preferable to satisfy the above expression (44) in order to suppress the amount of point aberration within the depth of focus on the reticle side of the projection optical system. Needless to say, the relations of the above equations (43) and (44) are also established when the projection optical system has the same magnification, reduction or enlargement magnification.

As an example, the numerical aperture NA _R on the reticle side of the projection optical system is 0.1, the wavelength of the exposure light is λ of 436 nm,
f _GF = 250 mm, f _GR = 250 mm, D ₂ = 200 m
Assuming that m is the focal length f2 of the cylindrical lens in the meridional direction from the equation (43), generally speaking (4)
From the equation (4), the power difference Δf in the orthogonal direction of the toric optical member is 1.43 × 10 ⁶ mm or more, and the variable magnification correction amount (magnification difference ΔB ₁ ) at this time is 3
It becomes ⁵ ppm (= 3.5 × 10 ⁻⁵ ) or less.

From the results of the analysis shown in FIGS. 9 to 11,
When a toric optical member is arranged between the reticle and the projection optical system or between the projection optical system and the wafer, the contribution of the correction to the astigmatism is suppressed to be small while the contribution of the correction to the magnification error is increased. On the other hand, when a toric optical member is arranged at or near the pupil of the projection optical system, the contribution of the correction to the astigmatism can be increased while suppressing the contribution of the correction to the magnification error to be small. Understand that it will be possible.

The toric optical member referred to in the present invention may be a toric lens having a rotationally symmetric spherical surface polished in one direction and having different powers in the orthogonal directions, or in the orthogonal directions. It may be a reflecting mirror having different powers, or may be a gradient index lens having different powers in orthogonal directions.

By the way, in the above description, astigmatism which occurs rotationally asymmetrical by using a toric optical member having different powers in orthogonal directions as an aspherical surface rotationally asymmetrical with respect to the optical axis of the projection optical system, Although correction of field curvature, magnification error, etc. was described, in addition to these aberrations and magnification errors that are rotationally asymmetrical, the magnification error component locally remaining randomly due to rotational asymmetry in the projection optical system. When a distortion component is generated, the lens surface of the cylindrical lens, which is a type of toric optical member that can move along the optical axis direction or can rotate about the optical axis, is locally processed by polishing or the like. By placing the processed and processed cylindrical lens between the reticle and the wafer, in addition to the correction of astigmatism, field curvature, and magnification error that occur rotationally asymmetrically, it occurs randomly. It is possible to correct the rate error components or distortion components.

Further, when the projection optical system has only a magnification error component and a distortion aberration component which are rotationally asymmetric and locally remain randomly, an optical element (lens, reflecting mirror) constituting the projection optical system is provided. If processing such as polishing is locally performed on itself, it is possible to correct a magnification error component and a distortion component that are randomly generated.

Further, in the case where the projection optical system has only the magnification error component and the distortion aberration component which are rotationally asymmetric and locally remain at random, in order to correct the magnification error component and the distortion aberration component which are randomly generated. , The parallel plane plate having a predetermined thickness is locally subjected to processing such as polishing, and the processed parallel plane plate is placed between the reticle and the projection optical system.
It may be arranged inside the projection optical system or between the projection optical system and the wafer. However, in this case, since the plane-parallel plate has a predetermined thickness, spherical aberration occurs, but the projection optical system may be configured in advance so as to correct the spherical aberration.

Next, an embodiment of the present invention will be described in detail with reference to FIG. FIG. 12 shows the configuration of the projection exposure apparatus according to the embodiment of the present invention. As shown in FIG. 12, a reticle 35 held by a reticle stage (not shown) is arranged above both sides (or one side) of the telecentric projection lens 36, and the projection is provided between the reticle 35 and the projection lens 36. As an optical means having rotationally asymmetric power with respect to the optical axis of the lens 36, toric optical members having different powers are arranged in the orthogonal direction. This toric-type optical member has, in order from the reticle side, a negative cylindrical lens 1 having a concave surface facing the projection lens side and having a negative power in the paper surface direction, and a positive cylinder lens having a convex surface facing the reticle side and having a positive power in the paper surface direction. The cylindrical lens 2 is provided, and the cylindrical lens 1 and the cylindrical lens 2 are provided so as to be rotatable about the optical axis of the projection lens 36, respectively.

Further, regarding the projection lens 36, the reticle 3
The wafer 38 placed on the wafer stage 37 is arranged at a position conjugate with the wafer stage 3.
Reference numeral 7 is composed of a two-dimensionally movable XY stage and a Z stage movable in the optical axis direction of the projection lens 36.

On the other hand, above the reticle 35, there is an illumination optical system (2) for uniformly illuminating the reticle 35 by Koehler.
1, 22, 23, 24, 25, 32, 33, 34) are provided, and in the illumination optical system, a measurement system 42 for measuring the optical characteristics of the projection lens and the exposure light IL described later are provided. A first alignment system 47 that optically detects the relative position between the reticle 35 and the wafer 38 by using light of different wavelengths is provided.

On the outside of the projection lens 36, the off
An axis-type second alignment system 48 is provided, and the second alignment system 48 is used for the exposure light I described later.
The position of the wafer 38 is optically detected by light having a wavelength different from L.

The embodiment shown in FIG. 12 will be specifically described. Exposure light IL emitted from a light source 21 such as a mercury lamp.
Is condensed by the elliptical mirror 22, reflected by the reflection mirror 23, converted into a substantially parallel light flux by the collimator lens 24, and is incident on the optical integrator 25 including a fly-eye lens. Second of the elliptical mirror 22
A shutter 26 is arranged in the vicinity of the focal point, and the exposure light IL can be interrupted at any time by rotating the shutter 26 via a drive unit 27 such as a motor.

When the exposure light IL is blocked by the shutter 26, the exposure light I reflected by the shutter 26
Since L is emitted in a direction substantially perpendicular to the optical axis of the elliptic mirror 22, the exposure light IL emitted in this way is made incident on one end of the light guide 29 by the condenser lens 28. Therefore, the exposure light IL emitted from the light source 21 enters either the optical integrator 25 or the light guide 29.

When the exposure light IL enters the optical integrator 25, a large number of secondary light source images (hereinafter, simply referred to as secondary light sources) are formed on the reticle side focal plane of the optical integrator 25. A variable aperture stop 30 is arranged on the forming surface. The exposure light IL emitted from these secondary light sources passes through the half mirror 31 arranged at an angle of 45 degrees with respect to the optical axis, and then passes through the first condenser lens 32, the dichroic mirror 33, and the second condenser lens 34. The pattern area on the lower surface side of the reticle 35 is illuminated with a uniform illuminance.

At the time of exposure, the toric optical member (1,
2) and the pattern of the reticle 35 by the projection lens 36
Is formed on the wafer 38. In this case, Opti
The secondary light source forming surface of the cal integrator 25 is a projection lens
It is conjugate with the pupil plane of 36 and is placed on the secondary light source formation surface.
By adjusting the aperture of the variable aperture stop 30
The coherency of the illumination optical system that illuminates the chicle 35 is shown.
The σ value can be changed. Illuminates the reticle 35
The maximum incident angle of the exposure light IL_IL, The projection lens 36
The half angle of the opening on the side of the chicle 35_PLThen, the σ value is sin
θ_IL/ Sin θ _PLCan be expressed as Where the σ value is
It is set to about 0.3 to 0.7.

Although not shown, an aperture stop is provided at the pupil position of the projection lens 36, and the aperture of this aperture stop may be variable. An adjusting plate 39 made of, for example, a glass plate is fixedly provided near the wafer holder of the wafer stage 37.
A reference pattern is formed on the surface of the projection lens 36 side. Correspondingly, in the image field of the projection lens 36 and in the vicinity of the pattern area of the reticle 35, the reference pattern on the adjustment plate 39 and the projection lens 36 are provided.
The reticle mark RM is formed at a position conjugate with respect to. As an example, the reference pattern on the adjustment plate 39 side is a cross-shaped opening pattern formed in the light-shielding portion, and the reticle mark RM on the wafer 35 side is the toric optical member (1, 2) and the projection on the reference pattern. The pattern is obtained by inverting the lightness and darkness of the pattern obtained by multiplying the projection magnification by the lens 36.

A condenser lens 41 and a reflection mirror 40 are arranged on the lower surface of the adjusting plate 39 of the wafer stage 37, and the exit end of the light guide 29 is fixed to the rear focal plane of the condenser lens 41. This light guide 2
The exit end surface of 9 is conjugate with the pupil plane of the projection lens 36, and thus is also conjugate with the variable aperture stop 30. Further, the light emitting surface at the exit end of the light guide 29 is such that the size of the projected image on the variable aperture stop 30 is set to be substantially equal to the aperture of the variable aperture stop 30, whereby the reference pattern on the adjusting plate 39 is set. The illumination σ value is approximately equal to the illumination σ value for the exposure light IL. Further, in the illumination optical system of the exposure light IL, the light receiving unit of the photomultiplier 42 is arranged at a position conjugate with the variable aperture stop 30 with respect to the half mirror 31. That is, the light receiving portion of the photomultiplier 42 is arranged so as to be conjugate with the pupil surface of the projection lens 36 and the exit end surface of the light guide 29. The detection surface of the light receiving portion is made larger than the image of the light emitting surface at the exit end of the light guide 39 projected on the detection surface to prevent light amount loss. Therefore, when the reference pattern of the adjusting plate 39 is illuminated from the lower surface side, most of the light emitted from the reference pattern of the adjusting plate 39 is irrespective of where the adjusting plate 39 is in the image field of the projection lens 36. The light enters the projection lens 36 and the toric optical member (1, 2), and enters the light receiving surface of the photomultiplier 42 through the reticle mark RM of the reticle 35.

The central processing unit 43 (hereinafter referred to as CPU) is electrically connected to the photomultiplier 42, and the photoelectric conversion signal output from the photomultiplier 42 is supplied to the CPU 43. Further, an X-direction mirror and a Y-direction mirror (not shown) are fixed to the upper surface of the wafer stage 37, and a laser interference system 44 is provided.
And, by using these two mirrors, the coordinates of the position on the wafer stage 37 can be constantly monitored. The coordinate information from the wafer stage 37 is supplied from the laser interference system 44 to the CPU 43, and the CPU 4
3 is the wafer stage 3 via the stage drive unit 45.
The position 7 can be moved to a desired coordinate position.

Now, the operation of this embodiment will be described. Optical characteristics (astigmatism, field curvature, magnification error, distortion aberration) that are rotationally asymmetrical with respect to the optical axis of the projection optical system remaining in the projection lens 36 and the toric optical member (1, 2) due to assembly error or the like. ), A reference reticle 35 'as shown in FIG. 13 is arranged in advance on a reticle stage (not shown). In the pattern area of the reference reticle 35 ', as shown in FIG.
Cross-shaped light-shielding patterns such as chrome are two-dimensionally arranged at predetermined intervals.

The CPU 43 blocks the exposure light IL by the shutter 26 via the drive unit 27, and then moves the adjustment plate 39 on the wafer stage 37 into the image field of the projection lens 36 via the stage drive unit 45. As a result, the exposure light IL (hereinafter, simply referred to as illumination light) reflected from the shutter 26 is emitted to the inside of the wafer stage 37 via the condenser lens 28 and the light guide 29. This illumination light is reflected by the reflection mirror 40, then converted into a substantially parallel light flux by the condenser lens 41, and illuminates the reference pattern formed on the adjustment plate 39 from the lower surface side. The reference pattern of the adjusting plate 39 is projected onto the light-shielding pattern of the reference reticle 35 'by the projection lens 36 and the toric optical member (1, 2).
The matching state between the two patterns is photoelectrically detected by the photomultiplier 42 via the second condenser lens 34, the dichroic mirror 33, the first condenser lens 33, and the half mirror 31. And
The CPU 43 sequentially detects, via the photomultiplier 42, the coordinates of the positions of the plurality of light-shielding patterns two-dimensionally arranged in the reference reticle 35 ′, and the coordinates of the wafer stage 37 via the laser interference system 44. While constantly monitoring the position, the wafer stage 37 is sequentially moved via the stage drive unit 45. As a result, the photomultiplier 42 photoelectrically detects the matching state between the plurality of light-shielding patterns two-dimensionally arranged in the reference reticle 35 'and the reference pattern of the adjusting plate 39, and CP
The U 43 sequentially stores the coordinate positions in the respective matching states in the first memory section inside the CPU 43 via the laser interference system 44. Further, the CPU 43 has a second memory unit and a first correction amount calculation unit (not shown) inside the CPU 43. The second memory unit has optical characteristics (non-rotationally asymmetric with respect to the optical axis of the projection optical system). Correlative information regarding point aberration, field curvature, magnification error, distortion, and the relative amount of rotation of the toric optical members (1, 2) is stored in advance.
Therefore, the first correction amount calculation unit, based on the information from the first and second memory unit, the toric optical member (1,
The optimum relative rotation amount to be corrected in 2) is calculated. Then, the CPU 43 outputs a drive signal to the drive unit 46 such as a motor based on the correction information from the first correction amount calculation unit, and the drive unit 46 outputs a predetermined correction amount (rotation amount) to the toric optical member (1 , 2) Rotate relatively.

After the above operation is completed, the normal reticle 35 used in the actual process is set on the reticle stage (not shown), and the CPU 43 switches the shutter 26 via the drive unit 27. As a result, the exposure light IL illuminates the reticle 35 via the illumination optical system, and the reticle 3 is illuminated.
The pattern image 5 is faithfully transferred onto the wafer 38 via the toric optical members (1, 2) and the projection lens 36. As described above, when the exposure transfer is continuously performed by the projection exposure apparatus, thermal energy due to the exposure light IL is accumulated in the projection lens 36 and the optical characteristics of the projection lens 36 may fluctuate. As described above, the optical characteristics of the projection lens 36 may be measured periodically during the operation, and the toric optical members (1, 2) may be rotated based on the measured results. At this time, it is more desirable to use the well-known technique of controlling the pressure between the lenses forming the projection lens 36 to adjust the magnification of the projection lens 36 itself.

The rotationally asymmetric optical characteristics (astigmatism, field curvature, magnification error, etc.) remaining in the projection lens 36 depend on the relative rotation amount of the toric optical members (1, 2).
It is desirable to confirm whether or not the distortion is corrected in a completely optimized state. In this case, more complete correction can be achieved by repeating the above-described operation.

When measuring the magnification error and distortion remaining in the projection lens 36, the wafer stage 37 is two-dimensionally moved to find the coordinate position of each light shielding pattern in the reference reticle 35 '. Although it is good, when the astigmatism and the field curvature remaining in the projection lens 36 are measured more accurately, the output from the photomultiplier 42 while moving the wafer stage 37 in the optical axis direction of the projection lens 36. The coordinate position of each light-shielding pattern within the reference reticle 35 'that maximizes the signal contrast may be obtained.

In the projection exposure apparatus of this embodiment, the non-linear expansion / contraction of the wafer 38 is performed by the semiconductor manufacturing process or the like.
Even when a semiconductor is manufactured by a plurality of projection exposure apparatuses, a difference in magnification error and distortion between the projection exposure apparatuses can be sufficiently dealt with. Specifically, first,
The CPU 43 sequentially and optically detects the coordinate positions of the plurality of wafer marks formed on the wafer via the second alignment 48 provided outside the projection lens 36, and thus the laser interference system 44. The wafer stage 37 is sequentially moved via the stage drive unit 45 while constantly monitoring the coordinate position of the wafer stage 37 via. As a result, the CPU 43 causes the second alignment 48
And the coordinate position of each wafer mark formed on the wafer obtained from the laser interference system 44 to the third position inside the CPU 43.
Sequentially stored in the memory section. Further, the CPU 43 has a fourth memory section and a second correction amount calculation section (not shown) inside the CPU 43. The fourth memory section has optical characteristics (non-rotationally asymmetrical with respect to the optical axis of the projection optical system) Point aberration, field curvature,
Magnification error, distortion) and toric optical member (1,
Information related to the relative rotation amount of 2) is stored in advance. Therefore, the second correction amount calculation unit calculates the optimum relative rotation amount of the toric optical member (1, 2) to be corrected based on the information from the third and fourth memory units. Then, based on the correction information from the correction amount calculation unit, the CPU 43 sends a drive signal to the drive unit 46 such as a motor.
The drive unit 46 causes the toric optical members (1, 2) to relatively rotate by a predetermined correction amount (rotation amount).

In the embodiment shown in FIG. 12, the rotationally asymmetrical optical characteristics (astigmatism, field curvature, etc.) remaining in the projection lens 36 due to the relative rotation amount of the toric optical members (1, 2). Although the example of correcting the magnification error and the distortion aberration has been described, it goes without saying that the toric optical members (1, 2) may be relatively moved in the optical axis direction of the projection lens 36. Further, in the embodiment of FIG. 12, an example of automatically correcting the rotationally asymmetric optical characteristics (astigmatism, field curvature, magnification error, distortion aberration) remaining in the projection lens 36 is shown, but a toric optical member ( It is also possible to manually rotate or move 1) or 2).

Further, instead of the light source 21, the elliptic mirror 22 and the collimator lens 24 in this embodiment, a laser light source such as an excimer laser which supplies a parallel light flux may be used.
Furthermore, this laser and a beam expander that converts this laser light into light having a predetermined cross section of the light flux may be combined. Now, in the embodiment shown in FIG. 12, a toric optical member (1,
The example in which 2) is arranged has been described, but the arrangement is not limited to this, and the structure shown in FIG. 14 may be adopted.

FIG. 14A shows an example in which toric optical members (1, 2) are arranged between the projection lens 36 and the wafer 38. As shown, the toric optical members (1, 2) are arranged on the reticle 35 in order from the wafer 38 side.
It has a negative cylindrical lens 1 having a concave surface facing the side and a positive cylindrical lens 2 having a convex surface facing the wafer 38 side. According to this configuration, similarly to the embodiment shown in FIG. 12, it is possible to largely contribute to the correction of the magnification error without affecting the astigmatism. Therefore, it is effective when the magnification error remains large in the projection lens 36 (effective as in the embodiment shown in FIG. 12).

FIG. 14B shows the front group 36A and the rear group 36.
In the projection lens 36 including B, the toric optical members (1, 2) are arranged between the front group 36A and the rear group 36B, that is, at or near the pupil position of the projection lens 36. As shown in the figure, the toric optical member (1, 2) comprises, in order from the reticle 35 side, a negative cylindrical lens 1 having a concave surface facing the wafer 38 side and a positive cylindrical lens 2 having a convex surface facing the reticle 35 side. have. According to this configuration, the error in magnification is not affected so much,
It can greatly contribute to the correction of astigmatism. Therefore, it is effective when a large amount of astigmatism remains in the projection lens 36.

FIG. 14C shows an example in which the toric optical members (2A, 2B) are arranged on the reticle 35 side and the wafer 38 side with the projection lens 36 interposed therebetween. As shown in the figure, a first positive cylindrical lens 2A having a convex surface facing the wafer 38 is provided between the reticle 35 and the projection lens 36, and the projection lens 36 and the wafer 38 are provided.
A second positive cylindrical lens 2B having a convex surface facing the reticle 35 side is provided between and. According to this configuration, similarly to the example shown in FIGS. 12 and 14A, it is possible to largely contribute to the correction of the magnification error without affecting the astigmatism.

FIG. 14D shows an example in which FIG. 14C is further applied, and positive cylindrical lenses (which are respectively arranged on the reticle 35 side and the wafer 38 side with the projection lens 36 interposed therebetween ( 2A, 2B) with a negative cylindrical lens (1
An example in which A, 1B) are combined is shown. According to this configuration, it is possible to greatly contribute to the correction of the magnification error without affecting the astigmatism. In this case, the magnification error remaining in the projection lens 36 is mainly corrected by one of the first toric optical member (1A, 2A) and the second toric optical member (1B, 2B), and the other is corrected by the other. A magnification error with respect to the expansion and contraction of the wafer 38 may be corrected. Furthermore, based on this configuration, the first toric optical member (1A, 2A) and the second toric optical member
If the power of one of the toric optical members (1B, 2B) is increased and the other is weakened, the toric optical member of one strong power will not affect the astigmatism. Rough adjustment of magnification error can be done without affecting
On the other hand, with a toric optical member having weak power, fine adjustment of the magnification error can be performed without affecting the astigmatism.

FIG. 14 (e) is the same as FIG. 14 (a).
(B) is further applied in combination, and as shown in the figure, the reticle 35 and the projection lens (front group 36
A) is a first toric optical member (1) composed of a negative cylindrical lens 1A and a positive cylindrical lens 2A.
A, 2A) are provided, and between the front group 36A and the rear group 36B in the projection lens 36 (at or near the pupil position of the projection lens 36), a negative cylindrical lens 1B and a positive cylindrical lens 2B are provided. A second toric optical member (1B, 2B) constituted by is provided. With this configuration, the first toric optical member (1A, 2A) can adjust the magnification error without affecting astigmatism so much, and the second toric optical member (1B, 2B) can be adjusted. In (), the astigmatism can be adjusted without affecting the magnification error so much, that is, the magnification error and the astigmatism can be independently corrected.

FIG. 14 (f) is the same as FIG. 14 (d).
(E) is shown in combination, and in addition to independent correction of magnification error and astigmatism, coarse adjustment and fine adjustment of magnification error and astigmatism can be performed.

According to the present invention, the rotationally asymmetrical optical characteristics of the projection optical system, such as the rotationally asymmetrical off-axis aberration component (astigmatism component, field curvature, etc.), rotationally asymmetrical magnification error component, etc., can be corrected. You can

[Brief description of drawings]

FIG. 1 is a principle diagram when a toric lens is a negative cylindrical lens.

FIG. 2 is a principle diagram when a toric lens is a positive cylindrical lens.

FIG. 3 is a diagram showing an operation of the negative cylindrical lens of FIG.

FIG. 4 is a diagram showing the operation of the positive cylindrical lens of FIG.

5 is a plan view showing a state of a light beam cross section on the virtual plane shown in FIG. 3. FIG.

6 is a plan view showing a state of a light beam cross section on the virtual plane shown in FIG. 4. FIG.

FIG. 7 is a diagram showing a geometrical optical relationship of the negative cylindrical lens shown in FIG.

8 is a diagram showing a geometrical optical relationship of the positive cylindrical lens shown in FIG.

FIG. 9 is a diagram showing a geometrical optical relationship of a projection optical system.

10 is a diagram showing a geometrical optical relationship when a cylindrical lens as a toric lens is arranged between the projection optical system and the reticle shown in FIG.

11 is a diagram showing a geometrical-optical relationship when a cylindrical lens as a toric lens is arranged near the pupil of the projection optical system shown in FIG.

FIG. 12 is a diagram showing a configuration of an exemplary embodiment according to the present invention.

FIG. 13 is a plan view showing a state of a reference reticle.

FIG. 14A is a diagram showing a state in which a positive cylindrical lens and a negative cylindrical lens as toric lenses are arranged between a reticle and a projection lens, and FIG. 14B is a pupil position of the projection lens or The figure which shows a mode when a positive cylinder lens and a negative cylinder lens as a toric lens are arrange | positioned in the vicinity, (c) is a toric to each between a reticle and a projection lens, and between a projection lens and a wafer. FIG. 3D is a diagram showing a state in which a positive cylindrical lens as a lens is arranged, and FIG. 7D shows a positive cylindrical lens as a toric lens and a negative cylindrical lens as a toric lens between the reticle and the projection lens and between the projection lens and the wafer. Diagram showing how the cylindrical lens is arranged,
(E) is a diagram showing a state in which a positive cylindrical lens and a negative cylindrical lens as toric lenses are arranged between the reticle and the projection lens and at or near the pupil position of the projection lens, respectively, (f) Shows a state in which a positive cylindrical lens and a negative cylindrical lens as toric lenses are arranged between the reticle and the projection lens, and at or near the pupil position of the projection lens and between the projection lens and the wafer. FIG.

[Explanation of symbols]

1 ... Negative cylindrical lens, 2 ... Positive cylindrical lens, 21 ... Light source, 22 ... Elliptic mirror, 23 ... Reflecting mirror, 24 ... Collimator lens, 25 ... Optical integrator, 32 ...
1st condenser lens, 34 ... 2nd condenser lens, 36 ... Projection lens, 37 ... Wafer stage, 38
... Wafer, 42 ... Photomultiplier, 48 ... Second
Alignment system.

─────────────────────────────────────────────────── ─── Continuation of front page (56) Reference JP-A-7-183190 (JP, A) JP-A-59-144127 (JP, A) JP-A-62-21118 (JP, A) JP-A-6- 89842 (JP, A) JP-A-4-134813 (JP, A) JP-A-3-88317 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) H01L 21/027 G02B 3 / 06 G02B 13/18 G03F 7/20

Claims (6)

(57) [Claims] 1. A projection exposure method for projecting a pattern image of a reticle onto a substrate, wherein two projections have rotationally asymmetrical optical characteristics in the projection optical system in order to correct rotationally asymmetrical optical characteristics of the projection optical system. A projection including: a step of calculating a relative rotation amount between the optical elements; and a step of setting the two optical elements so as to have the relative rotation amount based on the calculation result. Exposure method. 2. The projection exposure method according to claim 1, wherein the two optical elements have rotationally asymmetric powers. 3. The projection exposure method according to claim 1, wherein the two optical elements have different powers in directions orthogonal to each other. 4. The projection exposure method according to claim 1, wherein each of the two optical elements has a predetermined refractive index distribution. 5. The method further includes measuring a rotationally asymmetric optical characteristic with respect to an optical axis of the projection optical system, wherein the relative rotation amount is calculated based on the measured rotationally asymmetric optical characteristic. Claim 1 characterized by the above-mentioned.
~ The projection exposure method according to claim 4. 6. The projection exposure method according to claim 1, wherein the pattern image of the reticle is projected onto the substrate after setting the two optical elements. .