JP3303758B2 - Projection exposure apparatus and device manufacturing 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 apparatus and a method of manufacturing a device, for example, manufacturing a semiconductor device such as an IC or LSI, an imaging device such as a CCD, a display device such as a liquid crystal panel, or a device such as a magnetic head. In the lithography process, an electronic circuit pattern on a mask or a reticle (hereinafter, collectively referred to as a “reticle”) is projected or scanned on a wafer surface through a projection optical system to obtain a highly integrated device. It is suitable.

In particular, the present invention is suitable for a case where a reticle and a wafer are aligned with high accuracy (alignment) and an electronic circuit pattern on the reticle surface is projected onto a silicon wafer surface with high integration.

[0003]

2. Description of the Related Art Conventionally, when a semiconductor device or a liquid crystal panel is manufactured by using a photolithography technique, a pattern on a reticle surface is formed on a wafer or glass plate coated with a photoresist or the like via a projection optical system. For example, a projection exposure apparatus (stepper) for exposing and transferring an image onto a photosensitive substrate is used.

In particular, the recent semiconductor technology has been miniaturized, and for example, the resolution from a line width of 0.25 μm to a finer pattern has been discussed. Among them, the technology that plays a central role is a light exposure technology represented by a stepper. The performance of the projection lens (projection optical system), which is an index of the performance of the light exposure technology, can be roughly divided into three directions. That is, shorter wavelength, larger screen, and higher NA. From the viewpoint of a short wavelength, lithography using light from an ArF excimer laser has been actively developed as a next-generation technology.

[0005] One of the important items in the light exposure technique is to always maintain the same optical performance. In lithography using an ArF excimer laser, it is known that usable glass material absorbs light, and the optical performance of the projection optical system changes due to the absorption of the light. Light absorption is already known in lithography using light such as i-line.
In a stepper using an ArF excimer laser, since the depth of focus of the projection optical system is small, it is necessary to control the optical performance with higher precision than before. There are various aberrations of the projection optical system in such control of the optical performance. Among the various aberrations, the aberrations most difficult to correct include axial astigmatism and curvature of field, and third- and high-order distortions. Further, correction of the non-rotationally symmetric magnification of the object itself is another problem.

Next, axial astigmatism will be described with an example. In recent years, in order to enlarge the area to be exposed on the wafer surface, a scanning type projection exposure apparatus called a scanner which scans the reticle and the wafer while synchronizing the reticle and the wafer with respect to the slit shape has been actively developed. . When such slit-shaped exposure light is used, since the slit shape is not rotationally symmetric, heat absorption asymmetric with respect to the optical axis due to light absorption of the glass material causes astigmatism (on-axis astigmatism) in the projection optical system. Will come.

However, the projection optical system used in the conventional projection exposure apparatus does not include means for correcting the occurrence of asymmetrical aberrations due to absorption of exposure light by changing the optical performance of the projection optical system. Asymmetric astigmatism due to the influence of exposure light absorption could not be effectively corrected. Therefore, conventionally, only passive measures such as limiting the amount of exposure light incident on the projection optical system to reduce the occurrence of various aberrations have been taken.

Next, field curvature will be described. As a method of correcting the change in curvature of field caused by the projection optical system absorbing the exposure light, the power (refractive power) of an optical element such as a lens or a mirror is changed because the curvature of field is related to the Petzval sum. Methods are known. However, conventionally, an optical element having a variable refractive power has not been applied to a projection optical system of a projection exposure apparatus, and the effect of absorption of exposure light cannot be effectively corrected. Therefore, conventionally, only passive measures such as limiting the amount of exposure light incident on the projection optical system to reduce the occurrence of various aberrations have been taken.

Next, the non-rotationally symmetric magnification will be described.
Factors required for a light exposure technique for manufacturing a semiconductor element include not only high resolution but also mutual alignment accuracy between patterns superimposed over multiple layers.

[0010] A method generally referred to as global alignment is commonly used as a method of positioning. Global alignment errors are largely divided into inter-shot components, which are errors between shots to be printed, and intra-shot components, which are errors inside each shot. Recently, as the screen size increases, how to suppress the error of the intra-shot component has become a major problem. In an actual wafer, an asymmetric distortion occurs due to the process.

For example, if there is an error of 2 ppm in magnification as a component which is asymmetric and cannot be corrected for a screen size of 22 mm, an error of 22 mm × 2 ppm = 44 nm is obtained. It will have a value close to 5x. It is clear from the overlay budget that this value is out of tolerance, and suppressing the intra-shot component has become a major issue for the light exposure apparatus.

As for the distortion, there are known means for controlling the magnification and the third-order distortion. For example, there are techniques such as moving a plurality of elements in the projection optical system in the optical axis direction, and changing the pressure of gas sealed between the optical elements. Since the magnification is a basic amount of the optical system, it can be changed without changing other aberrations.However, the correction of the third-order distortion involves problems such as a change in aberration due to movement and a small adjustment range. Yes,
It is necessary to design from the beginning of design considering the correction. In particular, recently, when various imaging methods collectively called image improvement such as a modified illumination method or a phase shift mask are adopted, matching of distortion between the respective imaging methods becomes a problem. In this case, the design is arbitrarily 3 without any load.
Establishing a technology to control the next distortion is a major issue.

Next, a high-order distortion will be described. In a scanning type projection exposure apparatus, an asymmetric magnification difference, for example, an x-axis and y-axis magnification difference taken in a direction orthogonal to the optical axis can be corrected by scanning. In the scanning direction, the distortion is averaged by an averaging effect accompanying the scanning. For example, if the slit is long in the x direction and the scanning is performed in the y axis direction, the scanning result is adjusted by controlling the synchronization of the scanning in the y direction. It is suppressed to a small value by averaging in the slit.

However, since averaging in the x direction is not performed, it is necessary to optically control the distortion with high precision. The most important problem when correcting the third-order distortion by a known technique is the fifth-order or higher-order distortion. Correcting this is a major issue.

Japanese Patent Laid-Open No. 7-183190 discloses a projection exposure apparatus capable of adjusting optical characteristics that are rotationally asymmetric with respect to the optical axis of a projection optical system remaining in a projection optical system for projecting a pattern on a mask surface onto a wafer surface. No. pp. 139 to 163.

The publication discloses an illumination optical system for illuminating a first object, and a projection optical system for projecting an image of the first object illuminated by the illumination optical system onto a second object at a predetermined reduction magnification. Wherein an optical unit having a power that is rotationally asymmetric with respect to an optical axis of the projection optical system is arranged between the first object and the second object, and the optical unit includes the projection unit. To be able to rotate around the optical axis of the projection optical system or to move along the optical axis of the projection optical system, in order to correct optical characteristics that are rotationally asymmetric with respect to the optical axis of the projection optical system remaining in the optical system. It is configured to be provided as possible.

[0017]

There have been various methods for correcting a change in optical performance due to the fact that the projection optical system has absorbed exposure light, including focus correction associated with the simplest exposure. However, in the era of the scanning type projection exposure apparatus (scanner), it has been found that asymmetrical aberrations are newly generated with respect to the optical axis, and this is going to become a major problem. This is a phenomenon that occurs because the illumination light has a slit shape and is asymmetric with respect to the optical axis, that is, the light intensity distribution differs between the long direction and the short direction of the slit shape.In extreme cases, the vertical line and the horizontal line on the optical axis Axial astigmatism with different focus differences occurs. When the exposure load is not large, even if the exposure load is zero, the occurrence of axial astigmatism when a large exposure load is applied means the instability of the system, which is a major problem for the scanner.

One of the optical performances that changes when a projection optical system absorbs exposure light in a normal stepper or scanning projection exposure apparatus is field curvature. As described above, there is a method of correcting the field curvature by changing the power of the optical element. However, there has been a problem that it is generally very difficult to change the power of the optical element with high accuracy and satisfactorily correct the field curvature.

In order to increase the resolution, it is necessary to align the reticle and the wafer with high accuracy. However, when an asymmetric magnification difference occurs in the wafer as the target object, the positioning accuracy decreases.

The control of magnification is a problem in improving the alignment accuracy, but the projection optical system used in the stepper usually has only a magnification correction function that is rotationally symmetric with respect to the optical axis due to its configuration. is there. However, in the actual semiconductor process, when xy coordinates are taken on the wafer surface in accordance with the directionality of the pattern to be printed, there are cases where the magnification of expansion / contraction in the x direction and the y direction, that is, in the vertical and horizontal directions, is different. This imposes restrictions on the improvement of accurate alignment. Therefore, for example, if there is a difference of 2 ppm between the elongation in the x direction and the elongation in the y direction, there is a problem that the correction residual as described above remains.

Further, in a scanning type projection exposure apparatus, although the distortion value is improved in the form of averaging in the scanning direction, the advantage of averaging in the direction orthogonal to the scanning direction cannot be obtained. was there. In particular, for distortion, there are known means for controlling magnification and third-order distortion.

For example, there are techniques such as moving a plurality of optical elements in the projection optical system in the direction of the optical axis, and changing the pressure of gas sealed between the optical elements. However, it has been difficult to correct higher-order distortions of the fifth or higher order by these known methods. In some cases, it may be difficult to correct third-order distortion.

In particular, recently, when various imaging methods collectively referred to as image modification such as a modified illumination method or a phase shift mask are adopted, matching of distortion between the respective imaging methods becomes a problem. In this case, analysis has revealed that the most problematic elements are magnification differences in the x and y directions and higher-order distortions if the magnification and the third-order distortion are corrected by known means. Among them, the magnification difference between x and y can be easily corrected in the case of a scanning system, and thus there is a problem how to correct and control the higher-order components.

In the projection exposure apparatus proposed in Japanese Patent Application Laid-Open No. 7-183190, when an optical unit having rotationally asymmetric power is driven, a plurality of aberrations fluctuate, so that only the target aberration can be corrected. There was a problem that it was difficult.

According to the present invention, at least one of changes in optical performance due to absorption of exposure light by the projection system, such as axial astigmatism, curvature of field, symmetric or asymmetric magnification, and distortion, is appropriately determined. By using an optical means having at least two optical elements having an aspheric surface of a set shape, a projection exposure in which correction is performed with a minimum effect on other optical performances and a high-resolution pattern is easily obtained. It is an object of the present invention to provide an apparatus and a method for manufacturing a device using the same.

In addition to the above, the present invention provides at least two aspheric surfaces each having a shape appropriately set to correct the state of expansion and contraction after obtaining each process of a symmetrically baked wafer.
By using an optical means having two optical elements, it is possible to perform asymmetric correction in the projection system, that is, to independently control the magnification in the x direction and the y direction, and to reduce the influence on the optical performance caused by the control. It is an object of the present invention to provide a projection exposure apparatus which is minimized and a method for manufacturing a device using the same.

[0027]

According to a first aspect of the present invention, there is provided a projection exposure apparatus having a projection optical system for projecting a mask pattern onto a substrate, wherein at least one of the projection optical systems is a projection optical system. A pair of aspherical members that can be displaced in a direction perpendicular to the optical axis, wherein the pair of aspherical members are arranged such that the respective aspherical surfaces face each other, and each of the aspherical members of the pair of aspherical members is The shape of the spherical surface is determined such that the axial astigmatism of the projection optical system changes by changing the positional relationship in the orthogonal direction between the aspheric surfaces of the pair of aspherical members. The on-axis astigmatism of the projection optical system is adjusted by displacing the spherical member in a direction perpendicular to the optical axis.

According to a second aspect of the present invention, in the first aspect of the present invention, the shape of each of the aspherical surfaces of the pair of aspherical members is expressed as x
Is characterized by including the following third-order terms: According to a third aspect of the present invention, in the first aspect of the present invention, the pair of aspherical members move in opposite directions in the orthogonal direction. According to a fourth aspect of the present invention, in the second aspect, the projection exposure is performed while scanning the mask and the substrate, and the x direction is orthogonal to the scanning direction.

A photographic exposure apparatus according to a fifth aspect of the present invention has a projection optical system for projecting a mask pattern onto a substrate, wherein at least one of the projection optical systems has a direction orthogonal to an optical axis of the projection optical system. A pair of aspherical members that are displaceable to each other, the pair of aspherical members are arranged so that the respective aspherical surfaces face each other, and the shape of each aspherical surface of the pair of aspherical members is When the positional relationship in the orthogonal direction between the aspheric surfaces of the pair of aspherical members changes, the curvature of field of the projection optical system is determined to change, and the at least one aspherical member is orthogonal to the optical axis. The projection optical system is characterized by adjusting the field curvature of the projection optical system by displacing in the direction.

According to a sixth aspect of the present invention, in the fifth aspect of the present invention, the shape of each of the aspherical surfaces of the pair of aspherical members is expressed as x
And the optical characteristic includes the curvature of field in the x direction. A seventh aspect of the present invention is characterized in that, in the fifth aspect of the present invention, the pair of aspherical members move in the orthogonal direction in directions opposite to each other. The invention according to claim 8 is the invention according to claim 6, wherein
The projection exposure is performed while scanning the mask and the substrate, and the x direction is orthogonal to the scanning direction.

A projection exposure apparatus according to a ninth aspect of the present invention has a projection optical system for projecting a mask pattern onto a substrate, wherein at least one of the projection optical systems has a direction orthogonal to an optical axis of the projection optical system. A pair of aspherical members that are displaceable to each other, the pair of aspherical members are arranged so that the respective aspherical surfaces face each other, and the shape of each aspherical surface of the pair of aspherical members is The projection magnification of the projection optical system is determined to change by changing the positional relationship in the orthogonal direction between the aspheric surfaces of the pair of aspheric members, and the at least one aspheric member is orthogonal to the optical axis. The projection magnification of the projection optical system is adjusted by displacing the projection optical system in a direction in which the projection optical system moves.

According to a tenth aspect of the present invention, in the ninth aspect of the present invention, the shape of each of the aspherical surfaces of the pair of aspherical members is x when expressed in an expression, assuming that the direction of the movement is the x direction.
And the optical characteristic includes the projection magnification in the x direction. Claim 11
The invention of claim 9 is characterized in that, in the invention of claim 9, the pair of aspherical members move in opposite directions to each other in the orthogonal direction. According to a twelfth aspect of the present invention, in the tenth aspect, the projection exposure is performed while scanning the mask and the substrate, and the x direction is orthogonal to the scanning direction.

A projection exposure apparatus according to a thirteenth aspect of the present invention has a projection optical system for projecting a mask pattern onto a substrate, wherein at least one of the projection optical systems has a direction orthogonal to an optical axis of the projection optical system. A pair of aspherical members that are displaceable to each other, the pair of aspherical members are arranged so that the respective aspherical surfaces face each other, and the shape of each aspherical surface of the pair of aspherical members is By changing the positional relationship in the orthogonal direction between the aspherical surfaces of the pair of aspherical members, the distortion of the projection optical system is determined to change, and the at least one aspherical member is orthogonal to the optical axis. The distortion of the projection optical system is adjusted by displacing in the direction.

According to a fourteenth aspect, in the thirteenth aspect, each of the pair of aspherical members has a shape of an aspherical surface,
Assuming that the direction of the movement is the x direction, the expression includes a fifth-order term of x when expressed by an equation. Claim 15
In the invention according to claim 13, in the invention according to claim 13, the shape of each aspheric surface of the pair of aspherical members includes a seventh-order term of x when expressed by an expression, assuming that the direction of the movement is the x direction. It is characterized by: According to a sixteenth aspect, in the thirteenth aspect, the pair of aspherical members move in opposite directions to each other in the orthogonal direction. Claim 1
The invention of claim 7 is the invention of claim 14 or 15, wherein the projection exposure is performed while scanning the mask and the substrate, and the x direction is orthogonal to the scanning direction. The invention of claim 18 is the invention of claim 6, further comprising, in the projection optical path, a second pair of transparent aspheric members in which respective aspheric surfaces face each other. The aspherical member is movable in a direction orthogonal to the optical axis and in a y direction orthogonal to the x direction, and the mask pattern is projected onto a substrate by the movement of the second pair of aspherical members. The field curvature in the y-direction changes, and the shape of each aspheric surface of the second pair of aspherical members becomes 3 when expressed by an equation.
It is characterized by including the following items. According to a nineteenth aspect, in the eighteenth aspect, the field curvature and the axial astigmatism when the pattern of the mask is projected on a substrate using the two pairs of aspherical members are adjusted. It is characterized by. According to a twentieth aspect, in the invention according to the eighteenth aspect, the second pair of aspherical members move in the orthogonal direction in directions opposite to each other. Claim 21
The invention of claim 18 is characterized in that, in the invention of claim 18, each of the two pairs of aspherical members moves in the orthogonal direction in directions opposite to each other. A twenty-second aspect of the present invention is characterized in that, in the eighteenth aspect, one member of the two sets of aspherical members is one member having two aspheric surfaces on both surfaces common to each pair.

A projection exposure apparatus according to a twenty-third aspect of the present invention has a projection optical system for projecting a mask pattern onto a substrate, wherein at least one of the projection optical systems is in a direction orthogonal to the optical axis of the projection optical system. A pair of aspherical members that are displaceable to each other, the pair of aspherical members are arranged so that the respective aspherical surfaces face each other, and the shape of each aspherical surface of the pair of aspherical members is The refractive power as one system of the pair of aspherical members is changed by changing the positional relationship in the orthogonal direction between the aspherical surfaces of the pair of aspherical members, and the at least one aspherical surface is determined. It is characterized in that the refractive power of the projection optical system is adjusted by displacing the member in a direction orthogonal to the optical axis.

According to a twenty-fourth aspect, in the twenty-third aspect, the shape of the aspheric surface is determined so that the refractive power of the one system becomes zero when the positional relationship is a predetermined relationship. It is characterized by:

According to a twenty-fifth aspect of the present invention, in the first aspect of the present invention, the facing aspherical surface has a shape that matches when the pair of aspherical members are in a predetermined positional relationship. It is characterized by having.

According to a twenty-sixth aspect of the present invention, there is provided a method of manufacturing a device, comprising the steps of: transferring a device pattern onto a substrate by the projection exposure apparatus according to any one of claims 1 to 25; and developing the transferred substrate. It is characterized by including.

[0039]

[0040]

[0041]

[0042]

[0043]

[0044]

[0045]

[0046]

[0047]

[0048]

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056]

[0057]

[0058]

[0059]

[0060]

[0061]

[0062]

[0063]

[0064]

[0065]

[0066]

[0067]

[0068]

[0069]

[0070]

[0071]

[0072]

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082]

[0083]

[0084]

[0085]

[0086]

[0087]

[0088]

[0089]

[0090]

[0091]

[0092]

[0093]

[0094]

[0095]

[0096]

[0097]

[0098]

[0099]

[0100]

[0101]

[0102]

[0103]

[0104]

[0105]

[0106]

[0107]

[0108]

[0109]

[0110]

[0111]

[0112]

[0113]

[0114]

[0115]

[0116]

[0117]

[0118]

[0119]

[0120]

[0121]

[0122]

[0123]

[0124]

[0125]

[0126]

[0127]

[0128]

[0129]

[0130]

[0131]

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139]

[0140]

[0141]

[0142]

[0143]

[0144]

[0145]

[0146]

[0147]

[0148]

[0149]

[0150]

[0151]

[0152]

[0153]

[0154]

[0155]

[0156]

[0157]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a schematic view of a main part of a projection exposure apparatus according to a first embodiment of the present invention. This embodiment shows a case where the present invention is applied to a normal stepper or a scanning type (scan type) stepper (projection exposure apparatus).

In the figure, reference numeral 4 denotes an exposure illumination system,
A reticle (mask) 1 as an object is illuminated. The exposure illumination system 4 uses an ArF excimer laser (wavelength 193n).
m) or one of a KrF excimer laser (wavelength 248 nm) or a lamp emitting g-line (436 nm) or i-line (365 nm), and a known optical system.

1 is a reticle (mask) as a first object
It is. A projection optical system 2 such as a refraction or catadioptric system projects the circuit pattern of the reticle 1 illuminated by the exposure illumination system 4 onto a wafer 3 (substrate to be exposed) as a second object.

T1 is an optical means having a function of controlling axial astigmatism. As will be described later, two optical elements 11, 1 each having an aspheric surface and made of quartz or fluorite are used.
Two. The optical means T1 is arranged near the pupil of the projection optical system 2. Reference numeral 5 denotes a wafer holder, which holds the wafer 3. Reference numeral 6 denotes a wafer stage on which the wafer holder 5 is mounted, which performs well-known xyz drive, θ drive, tilt drive, and the like.

Reference numeral 7 denotes an interference mirror for monitoring the position of the wafer stage 6 with an interferometer (not shown). The wafer 3 is positioned at a predetermined position by a wafer stage drive control system (not shown) using the interferometer mirror 7 and signals obtained from the interferometer, and projection exposure is performed in that state.

In this embodiment, the optical means T1 and the projection optical system 2 constitute one element of a projection system for projecting a pattern on the surface of the mask 1 onto the substrate 3. The same applies to the following embodiments.

When the present embodiment is a scanning type stepper (projection exposure apparatus), the reticle stage (not shown) on which the reticle 1 is mounted and the wafer stage 6 are adjusted according to the imaging magnification of the projection optical system 2. Scanning exposure is performed by moving the optical system 2 at a speed ratio in a direction orthogonal to the optical axis of the optical system 2.

The stepper of this embodiment is different from an ordinary stepper or a scanning stepper in that an optical means T1 and an optical means T1 driving means (not shown) are provided in a projection optical path. Are the same.

In general, a projection optical system using an ArF excimer laser (wavelength: 193 nm) as an exposure light source
Since the wavelength of the luminous flux from the rF excimer laser is in a short wavelength range, the selection range of the glass material usable therefor is narrow, and at present, there are only quartz (SiO ₂ ) and fluorite (CaF ₂ ).

However, quartz also has a problem in transmittance in this short wavelength region, and absorbs exposure light to cause a thermal change (optical characteristic change) to change the imaging performance. This situation is similar to the thermal change that occurs when i-rays (wavelength 365 nm) are used as the exposure light flux. In the case of the i-line, various glass materials are used to correct the chromatic aberration, and among them, a glass material having a low transmittance is also included.

For this reason, a problem of thermal aberration occurs even with the i-line. However, in a projection optical system using an ArF excimer laser, the situation becomes more severe as the pattern becomes finer and the depth of focus becomes smaller.

The glass material of the projection system causes aberration fluctuation due to the absorption of the exposure light. One of the aberrations that is difficult to correct at this time is an axial astigmatism which is a non-rotationally symmetric component aberration. Correction. On-axis astigmatism is caused by a non-rotationally symmetric manner in which light passes through the projection system. This kind of non-rotational symmetry with respect to the projection system is further emphasized in the case of the scanning optical system, because the cross section of the illumination light is rectangular slit-like light and the light beam impinging on the projection optical system 2 is also slit-like. .

In particular, the ratio of the length in the longitudinal direction to the length in the lateral direction of the slit is usually about 5, and a non-rotationally symmetric distribution (heat distribution) exists in the projection system in a more emphasized form than in the case of the stepper. I do. Aberrations generated for this purpose include on-axis and off-axis astigmatism.

The on-axis astigmatism is one in which the focus positions are different with respect to two cross sections which include the optical axis of the projection optical system and are orthogonal to each other.

Conventionally, the optical system is originally configured on the assumption that its optical characteristics are axially symmetric (rotationally symmetric) with respect to the optical axis. It was not expected that the effect would extend to the axis.

According to the present inventors, it has been confirmed that the onset of axial astigmatism due to exposure is large in the initial stage, and the value decreases as exposure continues, due to diffusion of absorbed heat.

In general, in a projection system for manufacturing a semiconductor device, the amount of axial astigmatism caused by absorption of exposure light by a glass material constituting the semiconductor device is small, but the pattern to be exposed is half micron or quarter micron. When it becomes finer and the depth of focus becomes smaller, it cannot be ignored.

The present inventor has introduced a transmission type optical element having a new function since the amount of axial astigmatism actually occurring is a small amount on the order of 0.2 to 0.3 μm. Have the potential to correct .

As a specific configuration of the projection system according to the present embodiment, one or a plurality of optical units each having at least two sets of aspherical optical elements in the optical path of the projection optical system are used. insert as aspheric face, and then shifting the transverse optical element becomes said set in a direction perpendicular to the optical axis to change the relative position related to the direction, thereby adjusting the axial astigmatism, Has been corrected. An optical means comprising an optical element consisting of two aspherical surfaces in a set generates its non-rotationally symmetric power according to the amount of shift, thereby changing the axial astigmatism of the projection system. It is possible to correct.

In this embodiment, in particular, attention is paid to the small amount of on-axis astigmatism caused by the absorption of exposure light, and effective correction is performed using optical means having only a small aspherical amount. . Further, since the optical means of the present embodiment has a small absolute value of the aspherical amount, it can only correct axial astigmatism and can have no adverse effect on other optical characteristics. Is also a preferred means.

Next, the specific structure of the optical means for correcting axial astigmatism according to this embodiment will be described.
The optical means of this embodiment uses a pair of optical elements, and is configured such that a small power can be generated when the two optical elements are considered as an integral element, and this power can be finely changed. It is characterized by:

The projection exposure apparatus according to the first embodiment shown in FIG. 1 shows a case where such optical means is used to correct axial astigmatism of a projection optical system.

Next, the optical means T used in the first embodiment shown in FIG.
1 will be described. FIG. 2 is a sectional view of a main part of an optical unit T1 having a function of controlling axial astigmatism according to the present embodiment.

In the figure, two optical elements 11 and 12 arranged facing each other have outer surfaces 11a and 1a.
2a is a plane, and the facing surfaces 11b and 12b are aspherical surfaces of the same shape, and when both surfaces are superimposed,
Face to face. The surfaces 11b, 12b
May be aspherical surfaces having different shapes.

A in the figure is an optical axis. The x and y axes are taken perpendicular to the optical axis A, and the aspherical shape of the optical element 11 is defined as fa (x, y) as the aspherical shape facing each other.
Assuming that the aspherical shape of the optical element 12 is fb (x, y) and the shifting direction is the x direction, both aspherical shapes are given by the following equations that differ by a constant term. That is, fa (x, y) = ax ³ + bx ² + cx + d ₁ fb (x, y) = ax ³ + bx ² + cx + d ₂ ‥‥‥ (1a)

Here, the expression for only x is that the two optical elements 11 and 12 are planes in the y direction, aspherical in the x direction, and shifted relative to the x direction. This is because only the optical power (focal length) in the x direction is controlled. Since the optical power is generated by laterally shifting the element in the x direction, x is 3
Use up to the next section.

In the initial state, the aspherical shape fa (x, y) of the optical element 11 and the aspherical shape fb of the optical element 12
Since the (x, y) irregularities completely match, the optical element 11
The optical means composed of the optical element 12 and the optical element 12 has no optical power and merely functions as a plane parallel plate. The smaller the distance (i.e., the distance) between the optical element 11 and the optical element 12 in the direction of the optical axis A, the better, for example, a value of about 100 μm is typical. Here, it is assumed that the optical element 11 is moved in the x direction by the distance Δ. The effect at this time is that if a, b, and c are constants, fa (x + Δ, y) −fb (x, y) = 3aΔx ² + 2bΔx + cΔ + (d ₁ −d ₂ ) + 3aΔ ² + bΔ ² + aΔ ³ ‥‥‥ (2a ).

Here, the influence of the higher-order terms of Δ is small and ignored, and the effect of the embodiment is easily understood.
Let b = c = 0 ‥‥‥ (3a). As a result, the expression (2a) can be simply expressed by the following expression (4a).

Fa (x + Δ, y) −fb (x, y) = 3aΔx ² + (d ₁ −d ₂ ) ‥‥ (4a)

[0186] It (4a) formula has the term x ² is a basis of the present embodiment. Therefore, the optical elements 11 and 12 become optical elements having optical power only in the x direction by the lateral shift amount Δ, and the power can be freely changed by the lateral shift amount Δ.

Since the operation of laterally shifting and obtaining the difference is a differentiation itself, a third-order term is entered as the shape of the aspheric surface, and a second-order component that gives optical power by the effect of the differentiation is output. This is the function of the optical elements 11 and 12.

In this embodiment, for simplicity, b
= C = 0, but the term 2bΔx in equation (2a) corresponds to the shift. Since Δ is a known amount for the purpose of controlling the power, the shift can be corrected. When b ≠ 0, the shift becomes a problem specifically in the case of alignment. This problem can be avoided by giving an instruction to the stage to reversely correct the shift resulting from changing the relative position of the pair of optical elements 11 and 12 during global alignment.

When an appropriate value is given to the term of the constant c, fa
The effect of reducing the absolute value of the deviation from the plane with respect to the aspherical surface represented by (x, y), fb (x, y) is obtained. Therefore, depending on the value of the constant a, it is also effective to intentionally give values instead of setting the constants b and c to zero. Actually, the constant b may be set to zero, and the constant c may be given a value opposite to the constant a.

However, since the correction based on the value of c can be actually corrected by the inclination of the incident light beam at the time of measurement by the interferometer, there is no problem even if C = 0.

Generally, on-axis astigmatism is 0.2 to 0.3 μm
The absolute value of the amount of aspherical surface required to correct small values before and after is very small. The actual effective amount is the optical element 1
Although it depends on the position where 1 and 12 are placed, there are several Newton stripes. As a typical example, suppose that the amount generated as a power component is several lines, 1 μm, the diameter of the lens is 200 mm, and the shift amount Δ at this time is 5 mm. From the equation (4a), 3a × 5 × 100 × 100 = 0.001, and a value of a = 6.7 × 10 ⁻⁹ is obtained.

Since 100 has a diameter of 200 mm, the radius value is shown. If b = c = 0, (1a)
In the formula, the amount of the aspheric surface is 6.7 × 10 ⁻⁹ × 100 × 100 × 100 = 6.7 × 10 ⁻³ , and the amount of the aspheric surface of ± 6.7 μm is set to the optical element 1.
1, 12 have.

In order to reduce the actual amount of deviation from the plane, it is advisable to add a term of constant c to this. The value of the constant c giving a value of 6.7 μm at 100 mm is 6.7 × 10
^-5 So can When c = -6.7 × 10 ^-5 constants a and a constant c as opposite sign deviation of aspherical amount of plane is reduced to ± 2.6 [mu] m.

FIG. 2 shows the shape of the aspherical surface when the constant c is zero, and FIG. 3 shows the shape at the cross section of y = 0 when the above value is inserted into the constant c. The aspherical surface has a gentle shape within a diameter of 200 mm, and the optical element 11 and the optical element 1
2 are complementary to each other, the amount of other aberrations caused by a change in the relative position between the optical element 11 and the optical element 12 can be suppressed to a value that is almost negligible. Only astigmatism can be finely corrected.

In some cases, other optical characteristics such as focus and magnification may be minute but change may be required. Therefore, if the other optical characteristics are corrected, the performance of the projection system remains unchanged. Can be considered.

Conventionally, there has been no idea that axial astigmatism is controlled by using an aspheric surface and continuously changing the amount thereof. In this sense, the present embodiment is used for manufacturing a semiconductor element which requires high performance. This greatly contributes to further enhancement of the function of the projection exposure apparatus.

Another point of the present embodiment is that, since the amount of on-axis astigmatism of interest is small, the aspheric surface to be used is reduced to an amount that can be easily measured by an interferometer. .

In this embodiment, the desired on-axis astigmatism is generated from the difference between the two aspherical surfaces which are relatively shifted laterally. Therefore, the aspherical amounts of the optical elements 11 and 12 are smaller than the amount of the difference. Is also nearly one digit larger. In the above example, an aspherical surface of 6.7 μm is required in order to obtain a value of 1 μm as an effect of the lateral shift. It is a key point in applying the present embodiment that the tilt can be optimized to reduce the diameter to ± 2.6 μm, thereby entering the high-accuracy measurement range of the interferometer. In manufacturing an aspherical surface, it is important to confirm that the surface has been accurately processed to a desired shape. However, if the surface can be suppressed to the amount of the present embodiment, the current technology can be applied sufficiently.

The aspherical amount can be further reduced by reducing the correction amount (reducing the dynamic range) or increasing the shift amount.

Further, in the present embodiment, correction of non-rotationally symmetric optical power can be realized by an optical element having a refraction function, so that it is effective for both refraction type and catadioptric type used in a projection exposure apparatus. The ability to provide a means is also a great advantage.

The direction of occurrence of on-axis astigmatism is closely related to the directionality of the slit in a scan type stepper that performs exposure with slit-like light (hereinafter referred to as “slit”). Therefore, the above-mentioned x direction having the optical power to be controlled is set so as to coincide with the longitudinal or lateral direction of the slit. Practically speaking, the setting is made so as to match the direction of the outer shape of the reticle 1 having a square shape in FIG.

Such a setting is also effective when an aberration (exposure aberration) occurs due to exposure in the stepper. In the present embodiment, a scan type stepper is taken as an example to briefly explain on-axis astigmatism.However, if a pattern on a reticle has directionality even with a normal stepper, diffracted light is distributed non-rotationally symmetrically, On-axis astigmatism may occur. The pattern on the reticle has the characteristic of having edges parallel to the outer shape of the reticle from the characteristics of CAD when designing the circuit pattern. If the x direction is set parallel to the outer shape of the reticle, the direction in which exposure aberration occurs and the x direction And the generated aberration can be suppressed.

A drive mechanism (not shown) for adjusting the positional relationship between the aspherical surfaces of the optical elements 11 and 12 by setting the characteristics of the axial astigmatism stored in advance in the CPU in the main body of the projection exposure apparatus. Instructed and driven by The amount of drive can be controlled by calculating the amount of correction using the amount of exposure, the reticle pattern ratio, the exposure energy, etc., from the characteristics of the axial astigmatism stored in the apparatus in advance by experiment or simulation. A method of controlling the driving amount while measuring the characteristics of the projection optical system, instead of calculating from the characteristics of the system input in advance, can also be applied. The amount of on-axis astigmatism has a characteristic that correlates with the focus change due to the exposure. The amount of focus change due to the exposure is monitored, the correction amount is converted, and the drive amount of the optical element 11 and / or the optical element 12 is calculated. It is also applicable.

Of course, the amount of astigmatism may be measured directly and the value may be fed back to the relative position driving amount of the optical elements 11 and 12.

Since the amount of on-axis astigmatism as exposure aberration changes with time, the driving amount of the optical element 11 and / or the optical element 12 also changes with time.

In the above embodiment, one optical element is shifted in the x direction. Alternatively, one optical element may be shifted in the x direction by δ and the other may be shifted by −δ in the x direction. This is shown in FIG.

That is, since fa (x + δ, y) −fb (x−δ, y) = 2a (3δx ² + δ ³ ) + 4bδx + 2cδ + d ₁ −d ₂ , b = c = 0 and δ of Neglecting the influence of higher-order terms, fa (x + δ, y) −fb (x−δ, y) ≒ 6ax ² δ
+ D ₁ −d ₂ .

Therefore, if the shift amount Δ = δ, the power change amount is doubled. Alternatively, in order to obtain the same power change, the value of the coefficient (constant) a can be halved. This leads to halving the amount of aspherical surface, and has the effect of facilitating shape evaluation. Further, the shift amount can be halved to obtain the same power change. This is advantageous for the space provided in the drive system and the positioning accuracy.

Further, in the configuration of the above-described embodiment, a system in which one surface of the optical element is made aspherical has been described. However, an aspherical surface may be provided on both surfaces. This is shown in FIG. At this time, the optical element 2
Assuming that the thicknesses 1 and 22 are thin, the operation and effects used in the above description are simply added. Therefore, in the case of aspherical surfaces having the same shape, the value of the coefficient a can be halved in order to obtain the same change in optical power as in the above-described example.
Naturally, a method of driving one element or a method of driving both elements in opposite directions can be adopted. Further, the pair of optical elements may be aspherical surfaces having different shapes in addition to the same shape.

As a matter of course, the same effect can be obtained by using a plurality of sets of optical elements. An example is shown in FIG.

FIG. 7 is a schematic view of a main part of a second embodiment of the present invention. In the present embodiment, the optical unit T1 is disposed in the optical path between the projection optical system 2 and the wafer 3, and the insertion position of the optical unit T1 in the optical path is different from that of the first embodiment in FIG. The other configuration is the same.

In this embodiment, the projection optical system 2 and the wafer 3
In the meantime, an optical element that forms a desired aspherical surface (effect) from the combination of the two aspherical surfaces that are laterally shifted as described above is inserted to correct axial astigmatism. The operation of the optical element itself is the same as that of the first embodiment. However, the optical means used in the present embodiment finds a position to be appropriately inserted depending on the configuration of the projection optical system other than near the pupil plane of the projection optical system 2, and is provided at this position. ing.

It is preferable that the optical means T1 is disposed near the pupil plane of the projection optical system 2 and between the projection optical system 2 and the wafer 3 or the reticle 1. This is because the three points have high independence of aberration control and are suitable positions. However, the optical means T1 can be inserted at a position other than near the pupil plane in the projection optical system.

FIG. 8 is a schematic view of a main part of an optical section T13 used in the third embodiment of the present invention in an x-direction section and a y-direction section. Parts other than the optical unit T13 of the projection exposure apparatus of the present embodiment have the configuration shown in FIG. 1 or FIG.

The optical means T13 of this embodiment has three optical elements 21, 22, 23. In the embodiments described above, since the direction in which astigmatism occurs is related to the longitudinal direction of the slit and the scanning direction, only correction of optical power in one direction has been considered. However, the occurrence of curvature of field as well as axial astigmatism may not be negligible,
This embodiment is provided to deal with this.

The optical means T1 shown in FIG. 8 has the same optical power variable function not only in the x direction but also in the y direction orthogonal to the x direction, so that the optical power in the two directions x and y can be independently controlled. The control allows both the curvature of field and the axial astigmatism to be corrected.

In FIG. 8, the reticle and the projection optical system are located above the figure, and the wafer is located below the figure. The upper surface 21 a of the optical element 21 is a flat surface, and the lower surface 2
1b and the surface 22a on the optical element 22 are aspherical surfaces facing each other, and are shifted from each other in the y direction so that they have optical power in the y direction and can be changed. I have. The principle of controlling the optical power in the y direction is the same as that described in the above equations (1a) to (4a), except that the parameter x is replaced with y.

The lower surface 22b of the optical element 22 and the upper surface 23a of the optical element 23 are aspherical surfaces facing each other, and these are shifted from each other in the x direction, so that the optical power in the x direction is obtained. To have. The principle of controlling the optical power in the x-direction is the same as that described in the above-mentioned equations (1a) to (4a). Surface 23 below optical element 23
b is a plane. Parameters a, b, and c of the two types of aspherical surfaces constituting the optical elements 21 and 22, and the optical elements 22 and 23
May be the same or different, and the shift amount may be controlled according to the equation (2a) for calculating the power.

With the configuration shown in FIG. 8, the x direction and y
The optical power of each of the directions can be controlled independently. The amount of optical power generated in the x and y directions is
If they are the same, the field curvature is corrected, and if the amounts are different from each other, the field curvature and the axial astigmatism are corrected, and the x direction and the y direction are corrected.
By generating only one optical power in one direction, axial astigmatism is corrected.

In FIG. 8, the upper surface 22a of the optical element 22 is defined as y
Although the lower surface 22b is used for the control in the x direction in the direction, the optical element 22 may be divided into upper and lower two parts to completely separate the control in the x direction and the control in the y direction as a solid. Optical element 21
The driving instruction of the relative positional relationship from to the optical element 23 is the same as in the example described above, and may be calculated from the characteristics of the system that is recognized in advance or calculated from the actually measured data. . Since the number of controlled objects has increased from one-dimensional to two-dimensional, the way of movement is only a little complicated, and a detailed driving method will not be described here.

In the first, second, and third embodiments, on-axis astigmatism,
Alternatively, when correcting the curvature of field, the set value of the best focus changes by a very small amount, but the amount of change can be calculated from the driving amount of the aspherical optical element and becomes a known amount. There is no problem if the change amount is obtained by the CPU and reflected on the control value of the wafer surface position in the optical axis direction of the projection optical system. Similarly, the influence on other optical performance, for example, the magnification, can be similarly corrected because it is an amount that can be calculated from the driving amount of the aspherical surface.

As described above, the influence of the optical means of the present invention on other optical performances can be suppressed to almost negligible. On the other hand, for example, when on-axis astigmatism is generated, there is a possibility that a non-rotationally symmetric magnification may occur.However, in a scan type stepper, a non-rotationally symmetric magnification, that is, a magnification difference between the scanning direction and the slit longitudinal direction is determined. There is no problem because it can be corrected.

In a normal stepper, there is no problem if a non-rotationally symmetric magnification correction function described later is added and operated.

The projection exposure apparatuses of the first, second, and third embodiments can exhibit a stable predetermined performance regardless of the input energy, thereby improving the reliability of the projection exposure apparatus and restricting the input energy. As a result, the throughput is improved and the cost of the semiconductor chip is greatly reduced. Further, since the amount of generation is variable, it can cope with various fluctuations, so that it is versatile, and since the amount of aspherical surface is small, the influence on other optical performance can be suppressed to a negligible value. Also, depending on the configuration of the system, it is a great advantage that the curvature of field can be controlled in addition to the axial astigmatism.

FIG. 9 shows a fourth embodiment of the projection exposure apparatus of the present invention.
FIG. This embodiment shows a case where the present invention is applied to a projection exposure apparatus which is a normal stepper or a scan type stepper.

In the figure, reference numeral 4 denotes an exposure illumination system,
The reticle 1 as an object is illuminated. Exposure illumination system 4
Is an ArF excimer laser (wavelength 193 nm) or K
rF excimer laser (wavelength 248 nm) or g-line (4
36 nm) or a lamp that emits i-rays (wavelength 365 nm) and a known optical system.

Reference numeral 1 denotes a reticle as a first object. 2
Denotes a refraction or catadioptric projection optical system, which projects the circuit pattern of the reticle 1 illuminated by the exposure illumination system 4 onto a wafer (substrate) 3 as a second object.

T2 is an optical means having a function of controlling a magnification which is rotationally symmetric or non-rotationally symmetric with respect to the optical axis, and has two optical elements 31 and 32 having aspherical surfaces as described later. The optical means T2 is arranged in the optical path between the reticle 1 and the projection optical system 2. Reference numeral 5 denotes a wafer holder, which holds the wafer 3. Reference numeral 6 denotes a wafer stage on which the wafer holder 5 is mounted, and xy
It performs z drive, θ drive, tilt drive, and the like.

Reference numeral 7 denotes an interference mirror for monitoring the position of the wafer stage 6 with an interferometer (not shown). The wafer 3 is positioned at a predetermined position by a wafer stage drive control system (not shown) using the interferometer mirror 7 and signals obtained from the interferometer, and projection exposure is performed in this state.

When the present embodiment is a scan type stepper, the reticle stage (not shown) on which the reticle 1 is mounted and the wafer stage 6 are optically controlled at a speed ratio corresponding to the imaging magnification of the projection optical system 2. Scanning exposure is performed by driving in a direction orthogonal to the optical axis of the system 2.

This embodiment is different from a normal stepper or a scanning stepper in that the optical means T2 is provided in the optical path, and the other structure is basically the same.

Generally, as a control method for controlling the imaging magnification of the projection system, a method of moving an optical element in the projection system in the optical axis direction, a method of controlling a pressure in a sealed space of a part of the projection system, and the like. Are known, and any of these can be applied to the present embodiment.

However, these methods control a magnification that is rotationally symmetric with respect to the optical axis, and correct the magnification to an anamorphic magnification, that is, with two axes that are each orthogonal to the optical axis and orthogonal to each other. It is not possible to control the magnification in each direction of a certain x-axis and y-axis to different values.

It has been confirmed that when actually manufacturing a semiconductor device, an anamorphic magnification is caused in the wafer itself. In the case of complex wafers, in the process of manufacturing more than 20 wafers, high heat processes such as film formation and diffusion are repeated. On the other hand, patterning by exposure is performed before such a high heat process. The pattern formed by the semiconductor element is not always isotropic for each process,
For example, there are many cases where the configuration is biased in a specific direction, such as only lines in the x direction in the bit line process and only lines in the y direction in the word line process.

Such a directionality is based on the fact that CAD in designing a semiconductor device is based on x, y coordinates, and the semiconductor device itself requires a large number of repetitive patterns. This is because the direction is offset. Therefore, the completed semiconductor device has the x direction and the y direction.
Even if the directional lines are used as much, the directionality may appear remarkably at the level of individual processes during manufacturing, and the elongation in the high heat process performed after patterning naturally depends on the directionality of the pattern. It has a corresponding anisotropy. This is the technical background of the present embodiment.

The anisotropic expansion and contraction of the wafer can be determined using the alignment detection function of the position detection sensor on the projection exposure apparatus side. Detection of expansion and contraction at this time is performed by reading a mark dedicated to alignment provided in advance on the wafer side by a position detection sensor on the projection exposure apparatus side in the alignment process.

For the detection of wafer anisotropy, all known alignment methods such as global alignment and die-by-die alignment can be applied. In the case of global alignment, it is preferable to use a calculation method in which the expansion and contraction of the entire wafer coincides with the expansion of each shot.

According to the analysis of the inventor, it was found that the anisotropy of the wafer due to the heat actually caused was extremely small. Average expansion and contraction of the wafer, ie x and y
The average value of the expansion and contraction in the direction reaches around 10 ppm. The proportion of the anisotropy among them depends on the process, but is up to 10%.
% To about 20%, which is around 2 ppm. Therefore, in order to correct the non-rotationally symmetric magnification of the wafer itself, it is necessary to control a value up to about 5 ppm in consideration of an optical system and other error factors described later.

The anisotropic expansion and contraction amount of about 2 ppm, which induces a non-rotationally symmetric component (magnification) of the wafer, has not been a problem so far. However, a larger screen size and a finer symmetrical line width have caused a new problem. It is emerging as a problem.

Next, a specific configuration of the optical means for correcting the anisotropy (expansion and contraction) of the wafer according to the present embodiment will be described. The present embodiment is characterized in that a projection optical system is provided with a rotationally symmetric or non-rotationally symmetric anisotropic imaging magnification by using optical means using a pair of optical elements.

The projection exposure apparatus according to the fourth embodiment shown in FIG. 9 corrects such a non-rotationally symmetric imaging magnification with respect to the optical axis using the optical means T2. In particular, in this embodiment, an optical means T2 composed of two sets of aspherical optical elements is inserted at a position closest to the reticle 1 side of the projection optical system 2 or at an equivalent position thereof, and the optical axis of the optical element is By changing and adjusting the positional relationship in the orthogonal direction, the magnification that is not rotationally symmetric with respect to the optical axis is corrected.

Further, in the present embodiment, the two directions in which the magnification can be controlled independently are made coincident with the X direction and the Y direction used when designing the semiconductor element pattern on CAD, so that the asymmetrical pattern that occurs during the actual manufacture of the semiconductor element is obtained. It is easy to control the optical system that matches the movement of the magnification.

Next, the optical means T used in Embodiment 4 of FIG.
2 will be described. FIG. 10 is a schematic cross-sectional view of a main part of an optical unit T2 having a function of supplying a desired magnification that is not rotationally symmetric with respect to the optical axis used in the present embodiment. In the figure, two optical elements 31 and 32 arranged facing each other have flat outer surfaces 31a and 32a,
The facing surfaces 31b and 32b have the same aspherical shape that coincides when they are superimposed on each other in the state shown in the figure.

In the figure, the x and y axes are taken in a form perpendicular to the optical axis A, and the aspherical shape of the optical element 31 is fa (x, y) and the optical element 32 is
Let fb (x, y) be the aspherical shape of, and both aspherical shapes are given by the same cubic expression of x that differs by a constant term. That is, fa (x, y) = ax ³ + bx ² + cx + d ₁ fb (x, y) = ax ³ + bx ² + cx + d ₂ ‥‥ (1b)

There is no y-term in the expression (1b) because optical power (finite focal length) is given to the optical system only in the x direction. The optical elements 31 and 32 are used by shifting their relative positions in the x direction. However, since such lateral shift generates optical power in the optical unit T2, the aspherical shape has a third-order term of x. Used up to.

In the initial state, the aspherical shape fa (x,
Since y) has an aspherical surface shape fb (x, y) and the irregularities completely coincide with each other, the optical means T2 including the optical element 31 and the optical element 32 has no optical power and functions as a parallel plane plate. Not just. The smaller the distance between the optical element 31 and the optical element 32 in the direction of the optical axis A, the better.
A value of the order of m is typical.

Here, it is assumed that the optical element 31 is moved by Δ in the x direction. The effect at this time is as follows: when a, b, and c are constants, fa (x + Δ, y) −fb (x, y) = 3aΔx ² + 2bΔx + cΔ + (d ₁ −d ₂ ) + 3aΔ ² x + bΔ ² + aΔ ³ ‥‥ (2b) Become.

Here, the higher-order term of Δ is ignored because the movement amount is small, and b = c = 0 ‥‥ (3b), the equation (2b) is simplified, and fa (x + Δ, y) − fb (x, y) = 3aΔx ² + (d ₁ −d ₂ ) ‥‥ (4b)

[0249] It Expression (4b) has the term x ² is a basis of the present embodiment. Therefore, the optical elements 31 and 32 become optical elements having optical power only in the x direction by the lateral shift amount Δ, and the optical power can be freely changed by changing the lateral shift amount Δ.

Since the operation of laterally shifting and obtaining the difference is a differentiation itself, a third-order component is provided as the shape of the aspherical surface, and a second-order component which gives power by the effect of the differentiation is output from the optical element 31, 32.

In this embodiment, for simplicity, b
= C = 0. The term 2bΔx in the equation (2b) corresponds to the shift. Since Δ is a known amount for the purpose of controlling the power, the shift can be corrected. Specifically, the shift becomes a problem when performing alignment. This problem can be avoided by giving a command to the stage to reversely correct the shift resulting from changing the relative position of the pair of optical elements 31 and 32.

If an appropriate value is given to the term of the constant c, it is also effective to give the absolute value of the deviation from the plane of the aspherical surface. In practice, the constant b is used to keep the value of the aspheric amount small.
Is preferably zero, and the constant c is given a value opposite to the constant a. Further, C = 0 may be set for the above-described reason.

In general, the absolute value of the amount of aspherical surface required to correct a small value having a magnification of about 2 ppm is very small. The actually effective amount depends on the position where the optical elements 31 and 32 are placed, but is a few Newton fringes. As a typical example, suppose that the amount to be generated as a power component is several lines, 1 μm, the diameter of the lens is 200 mm, and the shift amount Δ at this time is 5 mm. From the equation (4b), 3a × 5 × 100 × 100 = 0.001 and a value of a = 6.7 × 10 ⁻⁹ is obtained. Since 100 has a diameter of 200 mm, the value of the radius is shown. Assuming that b = c = 0,
In the equation (1b), the amount of the aspherical surface is 6.7 × 10 ⁻⁹ × 100 × 100 × 100 = 6.7 × 10 ⁻³ , and the amount of the aspherical surface of ± 6.7 μm is set to the optical element 3.
1, 32 have. In order to reduce the actual amount of deviation from the plane, a term of the constant c may be added to this. 100m
The value of the constant c giving a value of 6.7 μm at m is 6.6.
Since it is 7 × 10 ⁻⁵ , if c = −6.7 × 10 ⁻⁵ when the constants a and c are opposite signs, the deviation of the aspheric amount from the plane can be reduced to ± 2.6 μm.

FIG. 10 shows the shape of the aspherical surface when the constant c is zero, and FIG. 11 shows the shape when the above value is inserted into the constant c. Since the aspherical surface has such a gentle shape within the diameter of 200 mm, and the optical elements 31 and 32 are complementary (complementary) to each other, the optical element 31
The amount of aberration caused by the change in the relative position of the optical element 32 and the optical element 32 hardly affects other optical performances, and only the magnification can be finely corrected.

In the configuration of the fourth embodiment, one optical element has been described as being shifted laterally in the x direction. Alternatively, one optical element may be shifted in the x direction by δ and the other optical element may be shifted in the x direction by −δ. good. This is shown in FIG. That is, since fa (x + δ, y) −fb (x−δ, y) = 2a (3δx ² + δ ³ ) + 4bδx + 2cδ + d ₁ −d ₂ , b = c = 0 and the higher order of δ Neglecting the effect of the term, fa (x + δ, y) −fb (x−δ, y) ≒ 6ax ² δ
+ D ₁ −d ₂ .

Therefore, if the shift amount Δ = δ, the optical power change amount is doubled. Alternatively, in order to obtain the same optical power change when Δ = δ, a coefficient (constant) a
Can be halved. This leads to halving the amount of aspherical surface, and has the effect of facilitating shape evaluation. Further, if Δ = δ, the shift amount can be halved to obtain the same optical power change. This is advantageous for the space provided in the drive system and the positioning accuracy.

Actually, in this embodiment, the amount of aspherical surface is reduced to half by the effect of doubling the relative amount of displacement by shifting in the ± direction.

Further, in the configuration of the fourth embodiment, the system in which one surface of the optical element is made aspherical has been described. However, an aspherical surface may be provided on both surfaces. This is shown in FIG. At this time, assuming that the thicknesses of the optical elements 31 and 32 are small, the functions and effects used in the above description are simply added. Therefore, in the case of all the same shapes, the value of the coefficient a can be halved in order to obtain the same change in optical power as in the above example. Naturally, a method of driving one of them or a method of driving both in the opposite direction can be adopted.

Incidentally, it goes without saying that the same effect can be obtained by using a large number of sets of optical elements. Figure 1 shows an example
It is shown in FIG.

The difference between the magnification in the x direction and the magnification in the y direction may be used not only for the wafer process but also for distortion matching between a plurality of apparatuses, distortion matching in a plurality of exposure modes, or correction of an error in reticle creation. Can also. In this case, the amount of non-rotationally symmetric (anisotropic) correction of the magnification is also a few ppm, and the amount of correction is manually set in the exposure apparatus, for example, data is set as a parameter, and based on the set parameter. The relative positions of the optical elements 31 and 32 are adjusted by a drive mechanism (not shown), and setting of the device is performed. Of course, parameter settings can also be input directly to the exposure apparatus from values obtained by automatic measurement.

Since the circuit design of the semiconductor element is performed by CAD in the XY coordinate system, it is desirable that the x direction and the y direction orthogonal to the optical axis described above match the x direction and the y direction of the CAD.

In general, the x direction and the y direction coincide with the direction of the end face of the reticle.
The direction coincides with the vertical and horizontal side directions of the end face of the reticle. The x and y directions coincide with the xy directions in which the stage 6 moves,
It is possible to correspond to the x magnification and y magnification required when performing global alignment.

The intra-shot component at the time of global alignment is corrected by the calculation that the expansion and contraction of the wafer is the same as the expansion and contraction of the shot as described above. This assumption is based on the fact that the expansion ratio of the entire wafer is constant within the wafer, that is, good linearity is maintained, and it has been confirmed that a good correction result is obtained.

In fact, in the global alignment, when the expansion of sppm in the x direction and the expansion of tppm in the y direction are observed, the tppm is corrected by the rotationally symmetric magnification correcting means based on the known means inherent in the projection optical system. Driving the optical element 31 in the x direction in which non-rotationally symmetric magnification correction can be performed, (st) ppm is corrected, or sppm is corrected by the symmetric magnification correction means inherent to the projection optical system, The optical element 3 in the x direction that can perform asymmetric correction
By driving 1, it is sufficient to correct (ts) ppm.

Here, the correction of the non-rotationally symmetric magnification (component) is controlled while changing the position of the optical element using an optical element having an aspherical surface. This greatly contributes to further enhancement of the functions of the projection exposure apparatus for manufacturing.

In the present embodiment, in order to generate desired optical performance from the difference between the optical elements having two aspherical surfaces, the amount of the original aspherical surface, that is, the aspherical amount of the optical elements 31 and 32 themselves is The value is almost one digit larger than the desired final shape (shifted difference). In the above example, to obtain a value of 1 μm, an aspheric surface of 6.7 μm is required. The key point in applying the present embodiment is that the inclination is optimized to reduce the diameter to 2.6 μm, thereby entering the highly accurate measurement range of the interferometer. When manufacturing an aspherical surface, it is important to be able to measure whether or not the surface has been accurately processed into a desired shape. However, if the amount can be reduced to the extent of the present embodiment, the current technology is sufficient.

The command to drive the optical element having an aspherical surface used in the present embodiment may be based on the actual measured value of the wafer as described above or may be manual. Correction is performed immediately in the case of a manual operation, or immediately after the measurement is completed and the correction amount is calculated in the global alignment, and immediately before the exposure operation is started, by an instruction from the CPU controlling the entire exposure apparatus to the drive mechanism.

FIG. 15 is a schematic view of a main part of a fifth embodiment of the present invention. In the present embodiment, an optical unit T2 having a function of correcting a magnification that is not rotationally symmetric with respect to the optical axis is disposed in the optical path between the projection optical system 2 and the wafer 3, and is different from the fourth embodiment in FIG. In comparison, the insertion position of the optical means T2 into the optical path is different, and the other configuration is basically the same.

The optical means comprising a pair of optical elements each having an aspherical surface can be placed in various places as described above. However, depending on the insertion position, there are cases where the correction range of the magnification is restricted. Determine the insertion position taking into account.

In the fourth and fifth embodiments, the magnification (component) that is not rotationally symmetric with respect to the optical axis of the projection system can be adjusted, such as expansion and contraction due to the semiconductor process, distortion matching between apparatuses, and reticle creation error. In addition, the overlay accuracy when producing a semiconductor device is significantly improved. 256
It is predicted that the alignment accuracy rather than the resolving power will be limited at the time of the fine processing of the MDRAM or later. However, since the non-rotationally symmetric component which remains as the uncorrectable component can be corrected, the effect of the present embodiment is obtained. Is very large.

Also, since the magnification correction amount is variable,
Since it is possible to cope with various fluctuations, it is rich in versatility and the amount of aspherical surface is small. Therefore, there is a great advantage that only the magnification can be controlled while suppressing the influence on other performances to a negligible value.

FIG. 16 is a schematic view of a main part of a projection exposure apparatus according to a sixth embodiment of the present invention. This embodiment shows a case where the present invention is applied to a projection exposure apparatus which is a normal stepper or a scan type stepper.

In the figure, reference numeral 4 denotes an exposure illumination system,
The reticle 1 as an object is illuminated. Exposure illumination system 4
Is an ArF excimer laser (wavelength 193 nm) or K
rF excimer laser (wavelength 248 nm) or g-line (4
36 nm) or a lamp that emits i-rays (wavelength 365 nm) and a known optical system.

[0274] Reference numeral 1 denotes a reticle as a first object. 2
Denotes a projection optical system such as a refraction type or a catadioptric type, and projects a circuit pattern of the reticle 1 illuminated by the exposure illumination system 4 onto a wafer (substrate) 3 as a second object.

T3 is an optical means having a function of controlling the curvature of field, and has two optical elements 41 and 42 each having an aspherical surface, as described later. The optical means T3 is arranged near the pupil of the projection optical system 2. Reference numeral 5 denotes a wafer holder, which holds the wafer 3. Reference numeral 6 denotes a wafer stage on which the wafer holder 5 is mounted, which performs xyz drive, θ drive, tilt drive, and the like.

Reference numeral 7 denotes an interference mirror for monitoring the position of the wafer stage 6 with an interferometer (not shown). The wafer 3 is positioned at a predetermined position by a wafer stage drive control system (not shown) using the interferometer mirror 7 and signals obtained from the interferometer, and projection exposure is performed in that state.

When the present embodiment is a scanning stepper (projection exposure apparatus), the reticle stage (not shown) on which the reticle 1 is mounted and the wafer stage 6 are adjusted according to the imaging magnification of the projection optical system 2. Scanning exposure is performed by driving at a speed ratio in a direction orthogonal to the optical axis of the projection optical system 2.

The present embodiment is different from a normal stepper or a scanning stepper in that the optical means T3 is provided in the optical path, and the other configuration is basically the same.

Generally, in a projection optical system using an ArF excimer laser (wavelength 193 nm) as an exposure light source, A
Since the wavelength of the luminous flux from the rF excimer laser is in a short wavelength range, the selection range of the glass material usable therefor is narrow, and at present, there are only quartz (SiO ₂ ) and fluorite (CaF ₂ ).

However, quartz also has a problem in transmittance in this short wavelength region, and absorbs exposure light to cause a thermal change (optical change) to change the imaging performance. This situation is similar to the thermal change that occurs when i-rays (wavelength 365 nm) are used as the exposure light flux. In the case of the i-line, various glass materials are used to correct the chromatic aberration, and among them, a glass material having a low transmittance is also included.

For this reason, a problem of thermal aberration also occurs in the case of the i-line. However, in a projection optical system using an ArF excimer laser, the situation becomes more severe as the pattern becomes finer and the depth of focus becomes smaller.

Aberration fluctuation occurs because the glass material of the projection optical system absorbs the exposure light.
There are various aberrations such as coma, astigmatism, and field curvature.
One of these aberrations that is particularly difficult to correct is curvature of field.

Although the amount of such field curvature is actually small, it has become a particular problem because the pattern to be exposed is fine and the depth of focus is small.

Since the actually generated field curvature is of the order of 0.2 to 0.3 μm, the present inventor has a small correction amount, and introduces a transmission type optical element having a new function to perform correction. I found what I could do.

As a specific configuration of the present embodiment, optical means composed of a pair of optical elements each having an aspherical surface is inserted into the optical path of the projection system, and the positional relationship between the optical elements is orthogonal to the optical axis. The field curvature is corrected by changing the positional relationship in the direction.

The optical power of the two inserted optical elements each having an aspherical surface is made variable by being relatively shifted laterally. As a result, the Petzval sum of the entire optical system is changed to control the field curvature. I have. Particularly, in the present embodiment, attention is paid to a small change in field curvature caused by absorption of exposure light, and effective field curvature correction is performed using an optical element having only a small aspherical amount.

Since the optical means of this embodiment has a small absolute value of the aspherical amount, it can correct only the field curvature and does not affect other optical performances. Is also a preferred means.

Next, a specific configuration of an optical unit for correcting field curvature according to the present embodiment will be described. This embodiment has a small optical power when the two optical elements are considered as an integral unit by using an optical means composed of two aspherical optical elements, and this power is variable. It is characterized by having such a configuration. Optical means for varying the optical power for this purpose include, for example,
An optical power control element by lateral displacement as disclosed in Japanese Patent Application No. 10034 is used.

The projection exposure apparatus of the sixth embodiment shown in FIG. 16 shows a case in which such optical means is used to correct the field curvature of the projection system.

Next, the configuration of the optical means T3 shown in FIG. 16 will be described. FIG. 17 is a sectional view of a main part of an optical unit T3 having a function of controlling the field curvature of the present embodiment. In the figure, two optical elements 41 and 42 in which the aspherical surfaces face each other have outer surfaces 41a and 42a which are flat, and the facing surfaces 41b and 42b have the same aspherical shape. ing. In the figure, the x and y axes are taken in a form perpendicular to the optical axis A, and as the aspherical shapes facing each other, the aspherical shape of the optical element 41 is represented by fa (x, y),
Assuming that the aspherical shape of the optical element 42 is fb (x, y) and the direction in which the position is relatively shifted is the x direction, the aspherical shapes of the two are given by the following equations that differ by a constant term. That is, fa (x, y) = a (3xy ² + x ³ ) + b (2xy + x ² ) + cx + d ₁ fb (x, y) = a (3xy ² + x ³ ) + b (2xy + x ² ) + cx + d ₂ ‥‥‥ (1c) Become.

Here, the reason why the optical elements 41 and 42 are not symmetrical with respect to x and y in the equation (1c) is that
This is because the element shifts the relative position in the direction.
In order to generate optical power by shifting in the x direction, x is used up to the third order term.

In the initial state, the aspherical shape fa (x,
Since y) and the irregularities of the aspherical shape fb (x, y) match, the optical element 41 and the optical element 42 have no optical power as a whole, and merely function as a parallel plane plate. The smaller the distance between the optical element 41 and the optical element 42, the better, for example, a value of about 100 μm is typical. Here, it is assumed that the optical element 41 is moved in the x direction by Δ. The effect at this time is that if a, b, and c are constants, fa (x + Δ, y) −fb (x, y) = 3aΔ (x ² + y ² ) + 2bΔ (x + y) + cΔ + (d ₁ −d ₂ ) + 3aΔ ² + bΔ ² + aΔ ³ ‥‥‥ (2c).

Here, the influence of the higher-order term of Δ is ignored because the movement amount is small, and to make it easier to understand the effect of the embodiment, if b = c = 0 ‥‥‥ (3c), then (2c ) Is simply expressed as: fa (x + Δ, y) −fb (x, y) = 3aΔ (x ² + y ² ) + (d ₁ −d ₂ ) ‥‥ (4c)

The basis of the present embodiment is that equation (4c) has a term of (x ² + y ² ). Therefore, the optical elements 41 and 42 become optical elements having powers that are rotationally symmetric with respect to the optical axis by the lateral shift amount Δ, and the power can be freely controlled by the lateral shift amount Δ.

Since the operation of shifting and taking the difference is the differentiation itself, a third-order term is inserted as the shape of the aspherical surface, and the second-order component that gives power by the effect of the differentiation is output from the optical element 41, 42.

In this embodiment, for the sake of simplicity, b
= C = 0, but the term 2bΔ (x + y) in equation (2c) corresponds to the shift. Even when b ≠ 0, Δ is a known amount for the purpose of controlling the power, so that the shift can be corrected. Specifically, the shift becomes a problem when performing alignment. In global alignment, this problem can be avoided by giving an instruction to the stage to reversely correct the shift resulting from adjusting the relative relationship between the pair of optical elements 41 and 42.

If an appropriate value is given to the term of the constant c, fa
There is an effect that the absolute value of the deviation from the plane of the aspheric surface represented by (x, y), fb (x, y) can be reduced. Therefore, depending on the value of the constant a, it is also effective to intentionally give values instead of setting the constants b and c to zero. Actually, in order to keep the value of the aspherical amount small, the constant b is set to zero, and the constant c is given a value opposite to the constant a.

Also, the absolute amount of the aspherical surface is a problem at the time of measurement rather than processing, and c = 0 may be used since the influence of c can be canceled by tilting the measurement beam.

In general, the absolute value of the amount of aspherical surface required to correct a small value where the field curvature is about 0.2 to 0.3 μm is very small. The actual effective amount is the optical element 41,
Although it depends on the position where 42 is placed, there are several Newton stripes. As a typical example, suppose that the amount generated as a power component is several lines, and is 1 μm, and the diameter of the lens is 2 μm.
If the horizontal shift amount Δ at this time is 5 mm, then from equation (4a), 3a × 5 × 100 × 100 = 0.001, and a value of a = 6.7 × 10 ^-9 is obtained.

Since 100 has a diameter of 200 mm, the value of the radius is shown. If b = c = 0, (1c)
In the equation, the amount of the aspherical surface at the section of y = 0 is a value of 6.7 × 10 ⁻⁹ × 100 × 100 × 100 = 6.7 × 10 ⁻³ , and the original amount of the aspherical surface of ± 6.7 μm is obtained. This means that the optical elements 41 and 42 have.

To reduce the actual deviation from the plane, it is advisable to add a term of constant c to this. The value of the constant c giving a value of 6.7 μm at 100 mm is 6.7 × 10E
Since the constant a and the constant c are opposite signs, and c = −6.7 × 10 ⁻⁵ , the deviation of the aspherical amount from the plane is ± 5 in the section of y = 0.
It can be reduced to 2.6 μm.

FIG. 17 shows the shape of the aspherical surface when the constant c is zero, and FIG. 18 shows the shape of the cross section at y = 0 when the above value is inserted into the constant c. Since the aspherical surface has a gentle shape within the diameter of 200 mm, and the optical element 41 and the optical element 42 have complementary shapes that complement each other, the relative position of the optical element 41 and the optical element 42 changes. The amount of generation of other aberrations hardly affects other optical characteristics, and only the field curvature can be finely corrected. In some cases, it may be necessary to make some slight corrections on other optical characteristics on the device side, but there is no change in the performance of the entire projection system.

As described above, the correction of the field curvature fluctuation is controlled while making the correction amount variable by using the aspherical surface.
The present embodiment greatly contributes to further enhancement of the function of a projection exposure apparatus for manufacturing a semiconductor element which requires a high function.

Another point of this embodiment is that the amount of curvature of field to be corrected is small, and the aspheric surface used is reduced to an amount that can be easily measured by an interferometer.

In the present embodiment, in order to generate desired optical characteristics from the difference between the two aspherical surfaces, it is desired to obtain the original amount of aspherical surface, that is, the amount of aspherical surface of the optical element 41 and the optical element 42 themselves. The value is almost one digit larger than the final form (difference). 5. To obtain a value of 1 μm in the above example,
A 7 μm aspheric surface is required. It is a key point in applying the present embodiment that the tilt can be optimized to reduce the diameter to ± 2.6 μm, thereby entering the high-accuracy measurement range of the interferometer. In manufacturing an aspherical surface, it is important to confirm that the surface has been accurately processed to a desired shape. However, if the surface can be suppressed to the amount of the present embodiment, the current technology can be applied sufficiently.

In this embodiment, the optical power can be adjusted by an optical element having a refracting action, so that it is effective for both the refraction type and the catadioptric type used in a projection exposure apparatus for manufacturing semiconductor devices. It is also a great advantage that such means can be provided.

A drive mechanism (not shown) for adjusting the relative position between the aspheric surfaces of the optical elements 41 and 42 is instructed and driven by the setting of the exposure image plane (curve) characteristic stored in advance in the CPU of the projection exposure apparatus main body. Is done. In this case, the characteristics of the exposure image plane (curvature) are stored in the apparatus in advance through experiments or simulations, and the amount of change in the field curvature and the amount of field curvature due to exposure can be determined from the amount of exposure, the reticle pattern rate, and the exposure energy. It is calculated and the wafer stage is controlled. As another method, a method of controlling the driving amount while calculating the characteristics of the projection optical system instead of the a priori characteristics can be applied. The amount of curvature of field and the amount of curvature due to exposure have a deep correlation with the focus change due to exposure. Calculation can also be applied.

Of course, the field curvature may be measured directly and the value may be fed back to the driving amount of the optical element 41 and / or the optical element.

In the sixth embodiment, one optical element is shifted in the x direction. However, one optical element may be shifted in the x direction by δ and the other optical element may be shifted by −δ in the x direction. This is shown in FIG.

That is, fa (x + δ, y) −fb (x−δ, y) = 2a [3δ (x ² + y ² ) + δ ³ ] + 2bδ (2y ² +1) + 2cδ + d ₁ −d _2. If b = c = 0 and | δ | ≒ 0, then fa (x + δ, y) −fb (x−δ, y) ≒ 6aδ (x ² + y ² ) + d ₁ −d ₂ .

Therefore, if Δ = δ, the amount of change in optical power is doubled. Alternatively, when Δ = δ, the value of the coefficient (constant) a can be halved to obtain the same power change. This leads to halving the amount of aspherical surface, and has an effect of facilitating evaluation of the shape of the aspherical surface. Furthermore, Δ =
With δ, the amount of shift to obtain the same optical power change can be halved. This is advantageous for the space provided in the drive system and the positioning accuracy.

Further, in the configuration of the sixth embodiment, a system in which one surface of the optical element is an aspheric surface has been described, but this may be provided on both surfaces. This is shown in FIG. At this time, assuming that the thicknesses of the optical elements 41 and 42 are small, the functions and effects used in the above description are simply added. Therefore, in the case of all aspherical surfaces having the same shape, the value of the coefficient a can be halved in order to obtain the same optical power change as in the above-described example. Also,
Naturally, a method of driving one or a method of driving both in the opposite direction can be adopted.

Incidentally, it goes without saying that the same effect can be obtained by using a large number of sets of optical elements. An example is shown in FIG.

FIG. 22 is a schematic view of a main part of a seventh embodiment of the present invention. In the present embodiment, the optical means T3 is disposed in the optical path between the projection optical system 2 and the wafer 3, and the insertion position of the optical means T3 in the optical path is different from that in Embodiment 6 in FIG. The other configuration is the same.

In this embodiment, the above-mentioned optical means T3 comprising a pair of optical elements each having an aspherical surface is inserted between the projection optical system 2 and the wafer 3 to correct the field curvature. The operation of the optical means T3 itself is as described in the sixth embodiment. The position where the optical means T3 is inserted is
Other than the vicinity of the pupil plane as in the present embodiment, it can be appropriately set according to the configuration of the projection optical system.

The preferred arrangement position of the optical means T3 is between the reticle and the projection optical system in addition to the vicinity of the pupil plane and between the projection optical system and the wafer.

This is because the three places are highly independent and convenient when controlling aberrations.

When correcting the field curvature in the sixth and seventh embodiments,
Although the set value of the best focus changes by a small amount, the amount can be calculated from the driving amount of the aspherical optical element and becomes a known amount. This amount is obtained by the CPU in the projection exposure apparatus, and the light amount of the projection optical system is obtained. There is no problem if it is reflected on the control value of the wafer position in the axial direction. Similarly, the influence on the other optical performance, for example, the magnification, can be calculated because the magnification change amount can be calculated from the driving amount of the aspherical optical element and becomes a known amount. Therefore, the influence on the other optical performances due to the introduction of the optical means T3 of this embodiment can be suppressed to almost negligible.

Since the amount of curvature of field is variable,
Since it is possible to cope with various fluctuations, it has a great advantage that it is rich in versatility, and since only a small amount of aspheric surface is used, it is possible to control only the curvature of field while suppressing the influence on other optical performance to a negligible value.

In the present invention, the optical means T1 having the function of controlling the axial astigmatism, the optical means T2 having the function of controlling the non-rotationally symmetric magnification, and the optical means having the function of controlling the curvature of field are provided. At least two optical means are selected from the means T3, and the optical means having a configuration independently or combined are arranged in the optical path (the optical means T13 of the third embodiment in FIG. Two or three of axial astigmatism, non-rotationally symmetric magnification, and field curvature may be corrected.

FIG. 23 shows an eighth embodiment of the present invention, in which a projection optical system of a normal stepper or a scan type stepper which is a projection exposure apparatus for semiconductor production performs two-dimensional distortion correction. This is an example in which a set of aspherical optical elements is inserted between a projection optical system and a reticle. Here, the third-order distortion is defined as a distortion component having a third-order characteristic called a barrel type or a pincushion type in a manual of magnification and optics.

As for the distortion, there are known means for controlling the magnification and the third-order distortion. For example, there are techniques such as moving a plurality of elements of the projection optical system in the optical axis direction, and changing the pressure of gas sealed between the optical elements of the projection optical system. Magnification is a basic amount of the optical system, and can be changed without changing other aberrations. However, with respect to correction of third-order distortion, a change in other aberrations due to correction and a small adjustment range are small. Therefore, it is necessary to perform the design in consideration of the correction from the beginning of the design.

In particular, recently, when various image forming methods collectively called image modification such as a modified illumination method or a phase shift mask are adopted, matching of distortion between the respective image forming methods becomes a problem. This matching error mainly occurs due to a processing error of the optical system or the like. In this case, correction to the third-order distortion is a basic requirement. Here we minimize the constraints on the design,
Provided is an optical system characterized by freely correcting and controlling the third-order distortion and minimizing the influence on other performances caused by the correction and control.

This embodiment relates to a scan type stepper, in which a set of two aspherical optical elements is inserted into the projection optical system, for example, at a position closest to the reticle side.
Correcting the third-order distortion of the projection optical system in the direction orthogonal to the scanning direction by adjusting at least one of the two aspherical elements in the direction orthogonal to the optical axis and adjusting the positional relationship between the two elements. It is characterized by being able to.

The present invention is characterized by correcting the third-order distortion, which is a problem occurring in the scanning type exposure apparatus, by making the direction of the lateral shift coincide with the direction orthogonal to the scanning direction.

The projection optical system to which the present invention is applied may be a refraction type or a catadioptric system, and the projection optical system has means for controlling the magnification.

When various new image forming methods such as a modified illumination and a phase shift mask are used in combination, distortion matching between the respective image forming methods becomes a problem.

In the case of a scan type exposure apparatus, the magnification difference in the x and y directions can be easily corrected by adjusting the ratio of the speed of the reticle and the speed of the wafer which are scanned synchronously. The correction of the third-order distortion component by the aspherical elements is performed in a direction orthogonal to the scanning direction.

In the present embodiment, by inserting a simple optical element, the third-order distortion whose design or correction range is restricted is freely controlled, and the matching accuracy is improved. For this reason, for a magnification component that is optically easily corrected, a known method, for example, moving a part of the projection optical system in the optical axis direction or controlling the pressure in the optical system is used. Is corrected for the distortion component.

Next, the configuration of an optical system for generating and controlling tertiary distortion will be described. A third-order distortion is generated in the projection optical system by using an optical unit including a pair of optical elements as in the above-described embodiment.

FIG. 23 shows a main part of a scanning type exposure apparatus, wherein 1 is a reticle, 2 is a projection optical system, 3 is a wafer, 4 is an illumination optical system, 5 is a wafer holder, 6
Is a stage, and 7 is a mirror for a laser interferometer mounted to control the stage 6. Since it is a scanning type, the reticle and the wafer are scanned synchronously, and the pattern on the reticle is transferred onto the wafer. The configuration of FIG. 23 is the same as that of a normal scan type stepper, but differs in this embodiment in the optical element T4 for adjusting the third-order distortion indicated by 51 and 52.
Is the existence of

FIG. 24 shows the details. The optical elements 51 and 52 which are arranged to face each other are aspherical elements in which the outer surfaces are flat and the facing surfaces are a pair having the same aspherical shape. In the figure, x and y axes are taken perpendicular to the optical axis A, and the y direction is made to coincide with the scanning direction of the apparatus. In the aspherical shape facing each other, 51 is fa (x, y), and 52 is fb (x, y).
Then, both are given by a quintic expression of the same x that differs only in the constant term. That is, fa (x, y) = ax ⁵ + b ₁ fb (x, y) = ax ⁵ + b ₂ (1d) (1d) and there is no y term in the scanning direction consists of two aspherical elements This is for giving optical characteristics to the optical means only in the x direction orthogonal to the scanning direction y direction with respect to the projection optical system.
In the y direction for scanning, higher-order distortion is canceled due to averaging due to scanning differences, so in the present embodiment, only the x-direction needs to control the third-order distortion. Although 51 and 52 are used by shifting their relative positions in the x direction, a fifth order term of x is indispensable as the shape of the aspherical surface because a third order distortion is generated by lateral shifting.

In the initial state, the aspherical shape fa (x, y) cancels out because the asperities coincide with the aspherical shape fb (x, y), so that the optical elements 51 and 52 have a total optical power of zero. Yes, it simply acts as a plane-parallel plate. The smaller the distance between the optical elements 51 and 52 is, the better, for example, a value of about 100 μm is typical. Here, it is assumed that the element 51 is moved by Δ in the x direction. The effect of this time, fa (x + Δ, y) -fb (x, y) = 5aΔx 4 + 10aΔ 2 x 3 + 10a Δ 3 x 2 + 5aΔ 4 x + (b 1 -b 2) ‥‥‥ (2d) where If the higher-order terms of Δ are ignored because the movement amount is small, the equation (2d) becomes simple, and fa (x + Δ, y) −fb (x, y) = 5aΔx ⁴ + (b ₁ −b ₂ ) ‥ ‥‥ (3d). The basis of the present invention is that equation (3d) has a term of the fourth power of x. Therefore, the optical elements 51 and 52 become elements having fourth-order characteristics only in the x direction by the shift amount Δ, and the characteristics can be freely controlled (changed) by changing the lateral shift amount Δ. It is characteristic.

Since the operation of shifting and taking the difference is the differentiation itself, it is the function of the optical elements 51 and 52 that the seventh order is entered as the shape of the aspherical surface and the fourth order component is output by the effect of the differentiation. The absolute value of the amount of aspherical surface required to correct a small distortion value for the aforementioned matching is very small. As a typical example, suppose that the aspherical amount for generating a desired distortion is 1 μm.
m, the lens diameter is 200 mm, and the shift amount Δ at this time is 5 m
Assuming m, from (3d), 5a × 5 × 10 ^-8 = 0.001, and a = 4.00 × 10 ^-13 is obtained. 100 indicates the radius value because the diameter is 200 mm. Therefore, the amount of aspherical surface is obtained from the formula (1d) as 4.00 × (10 ⁻¹³ ) × (10 ⁻¹⁰ ) = 4.00 × 10 ^−3. You have it. To reduce the actual amount of deviation from the plane, it is advisable to add a term of cx which is a first-order term of x to this. The value of c that gives a value of 4.00 μm at 100 mm is 4.00 × 10 ^-5, so if a and c are reversed and c = −2.86 × 10 ^-5 , the deviation of the aspheric amount from the plane is ± 2.14 μm. Can be reduced to Since the aspherical surface has such a gentle shape within the diameter of 200 mm, and the aspherical shapes of the optical elements 51 and 52 are complementary to each other, the relative positions of the optical elements 51 and 52 are relatively small. Does not substantially affect other optical characteristics of the projection optical system, and can correct only the third-order distortion.

The present embodiment is characterized in that the correction direction of the tertiary distortion is set to a direction orthogonal to the scanning direction of the apparatus. However, the correction amount is small and the absolute value of the shift amount can be appropriately selected. The aspherical surface used can be reduced to an amount that can be easily measured by an interferometer. In the present invention, in order to generate desired optical characteristics from the difference between the two aspheric surfaces, the amount of the original aspheric surface, that is, the amount of aspheric surface of the optical elements 51 and 52 itself is larger than the final shape to be obtained. Becomes In the above example, each optical element requires an aspheric surface of 4.00 μm in order to obtain an amount of aspheric surface of 1 μm.

By optimizing the inclination, the diameter can be reduced to 2.14 μm, which is within the highly accurate measurement range of the interferometer.
In manufacturing an aspherical surface, it is important to be able to measure whether or not the surface has been accurately processed into a desired shape, but if the amount can be suppressed to the level of the present invention, the current technology is sufficient. To further reduce this value here 5m
The shift amount set in m may be set to a larger value.

The command to drive at least one aspherical optical element may be changed according to a procedure stored in the apparatus in advance, or may be performed based on a measured value of an actual wafer. For example, if the third-order distortion changes due to a difference in the illumination mode or NA, the amount of change is stored in advance, and when the illumination mode or NA is changed, the above-described aspherical optical element is laterally shifted. Good. At this time, the magnification component is also corrected at the same time by known correction means.

Also, instead of setting the distortion components from the beginning, a distortion component in the screen is analyzed using a reticle capable of measuring distortion, and a magnification component and a third-order distortion component are respectively determined based on the values. It may be corrected. In this case, the measurement may be automatically performed on the apparatus, or may be an off-line type in which an image of the reticle is once printed on the wafer and then measured.

FIG. 25 shows a ninth embodiment of a semiconductor exposure apparatus embodying the present invention. The difference from the eighth embodiment is that two optical elements 51 and 52 constituting an optical unit T4 for controlling tertiary distortion are provided between the wafer and the projection optical system. Other operations are almost the same as those of the above-described embodiment.

In the description of the eighth and ninth embodiments, the case where only one of the two aspherical optical elements placed opposite to each other is moved to obtain a desired characteristic has been described. Is not limited to this, and may be symmetrical with respect to the optical axis. For example, the upper optical element 51 may be moved to the right by Δ and the lower optical element 52 may be moved to the left by Δ.

Up to this point, the third-order distortion correction has been performed only in the direction orthogonal to the scanning direction. However, when the width of the scanning slit is large or when applied to a stepper, the shape of the aspherical surface is changed to ga (x , y) = a ^{a (x 5 + 5x 4 y} ) + b 1 gb (x, y) = a (x 5 + 5x 4 y) + b 2 ‥‥‥ (4d). In the initial state, the aspheric surface ga (x, y) is the aspheric surface gb
Since the projections and depressions coincide with (x, y) and are canceled, the optical power of the optical element 51 and the optical element 52 is zero as a whole, and only functions as a parallel plane plate.
The smaller the distance between the optical elements 51 and 52, the better, for example,
A value on the order of 100 μm is typical. Here, it is assumed that the element 51 is moved in the x direction by Δ. If the effect of this time ignoring the small effect of higher order term of Δ, ga (x + Δ, y) -gb (x, y) = 5aΔ (x 4 + y 4) + (b 1 -b 2) ‥‥ ‥ (5d) That (5d) expression has a term of x ⁴ + y ⁴ is a basis of the present invention. For this reason, the optical means composed of the optical elements 51 and 52 is a rotationally symmetric optical element having the fourth order characteristic by the shift amount Δ, and the characteristic can be freely controlled by the lateral shift amount Δ. It is.

Since the operation of shifting and taking the difference is the differentiation itself, the function of the optical elements 51 and 52 is to put the seventh order as the shape of the aspheric surface and to produce the fourth order component by the effect of the differentiation. . As a result, the third-order distortion can be freely controlled. The optical elements 51 and 5
The insertion position of 2 may be between the projection optical system and the reticle, or between the projection optical system and the wafer, as in the above embodiment. In some cases, the present element can be incorporated in the projection optical system.

As described in the eighth and ninth embodiments, instead of moving only one of the optical elements, the optical elements 51 and 52 are symmetrical with respect to the optical axis (in the opposite direction) as described in the other embodiments. It is good also as composition which moves.

FIG. 26 is a schematic view of a main part of a projection exposure apparatus according to a tenth embodiment of the present invention. This embodiment shows a case where the present invention is applied to a projection exposure apparatus which is a normal stepper or a scan type stepper.

In the figure, reference numeral 4 denotes an exposure illumination system,
The reticle 1 as an object is illuminated. Exposure illumination system 4
Is an ArF excimer laser (wavelength 193n) as a light source
m) or KrF excimer laser (wavelength 248 nm),
Alternatively, a lamp that emits i-rays (wavelength 365 nm) is used, and is constituted by a known optical system.

Reference numeral 1 denotes a reticle as a first object. 2
Denotes a projection optical system such as a refraction or catadioptric system, which projects the circuit pattern of the reticle 1 illuminated by the exposure illumination system 4 onto a wafer (substrate) 3 as a second object.

T5 is an optical means having a function of controlling higher-order distortion, and has two optical elements 61 and 62 having aspherical surfaces as described later. The relationship is changed by shifting at least one element in a direction orthogonal to the optical axis. For example, the element is shifted in a direction orthogonal to scanning to correct higher-order distortion in this direction. The optical means T5 is arranged in the optical path between the reticle 1 and the projection optical system 2. 5
Denotes a wafer holder, which holds the wafer 3. Reference numeral 6 denotes a wafer stage, and a wafer holder 5
, And performs xyz drive, θ drive, tilt drive, and the like.

Reference numeral 7 denotes an interference mirror for monitoring the position of the wafer stage 6 with an interferometer (not shown). Using a signal obtained from the interferometer mirror 7 and the interferometer, the wafer 3 is always positioned at a predetermined position by a wafer stage drive control system (not shown), and exposure is performed from this position.

When the present embodiment is a scanning type projection exposure apparatus, the reticle stage (not shown) on which the reticle 1 is mounted and the wafer stage 6 are moved at a speed ratio corresponding to the imaging magnification of the projection optical system 2. Scanning exposure is performed by driving in a direction perpendicular to the optical axis A.

In the present embodiment, the provision of the optical means T5 in the optical path is different from a normal stepper or a scanning stepper, and the other constitutions are basically the same.

In the embodiment shown in FIG. 26, an optical unit T5 comprising a pair of optical elements each having an aspherical surface for performing high-order distortion correction is combined with the projection optical system 2 and the reticle 1
And between them.

Here, the higher-order distortion means a higher-order or higher-order component excluding a component having a tertiary characteristic with respect to the value of the image height called a barrel type or a pincushion type in a manual of optics or the like. It is defined as a higher-order distortion component.

The projection optical system to which this embodiment is applied may be a refraction type or a catadioptric system, and the projection optical system has a function of controlling magnification and tertiary distortion in advance. As means for controlling magnification and third-order distortion, for example, a method of moving an optical element (a lens or the like) in the projection optical system in the optical axis direction, a method of controlling pressure in a part of the space of the projection optical system, and the like. Known means are known, and all of them are applicable to the present embodiment. However, these means control magnification and low-order distortion components up to the third order, and cannot control high-order distortion components.

However, as described above, if various new image forming methods such as a modified illumination and a phase shift mask are used in combination, distortion matching between the respective image forming methods becomes a problem. This matching error mainly occurs due to a processing error of the optical system or the like.

According to the error analysis of the match ring performed by the present inventor, when the magnification and the third-order distortion are corrected, the main component of the error is the higher-order distortion component and the magnification difference in the x and y directions orthogonal to each other. Were found to be primary. In the case of a scanning type exposure apparatus, such a difference in magnification in two directions can be easily corrected by adjusting the speed ratio at the time of synchronous scanning. However, it was difficult to adjust higher-order components with ordinary optical systems. This is based on the fact that simply moving the optical element in the direction of the optical axis or controlling the pressure in the space changes only the third-order distortion component, while the fifth- and higher-order components hardly change. .

However, if a new imaging method such as a modified illumination or a phase shift mask is used to increase the depth of focus or to resolve a finer pattern, a higher-order image is added to the number of imaging methods used. The distortion value also changes. Therefore, free control of higher-order values has become an essential requirement for improving matching accuracy.

In the tenth embodiment, the higher-order distortion, which was difficult to control with the conventional projection optical system, is freely controlled using an aspherical optical element to improve the matching accuracy. Let me.

For this reason, in this embodiment, the magnification component and 3
For the next distortion component, a known method, for example, moving a part of the optical element of the projection optical system in the optical axis direction,
A method such as moving the reticle in the optical axis direction or controlling the pressure in the optical system is used, and higher-order components are corrected using the optical unit T5.

Next, the configuration of the optical means T5 for generating and controlling higher-order distortion will be described.

The present embodiment is characterized in that high-order distortion is generated in the projection optical system 2 by using the optical means T5 composed of a pair of optical elements. As the optical element for this purpose, for example, an optical element disclosed in Japanese Patent Publication No. 43-10034 is applied.

FIG. 29 is a schematic view of a main part of the optical means T5 of this embodiment. The optical elements 61 and 62 of the optical unit T5 of the present embodiment, which are disposed to face each other, have flat outer surfaces, and the facing surfaces have the same aspherical shape.

In the figure, x and y axes are set so as to be orthogonal to the optical axis A, and the y direction coincides with the scanning direction of the apparatus. The optical element 61 has an aspherical shape facing each other.
If a (x, y) and the optical element 62 are fb (x, y), both are given by the same 7th order expression of x which differs by a constant term. That fa (x, y) = ax 7 + b 1 fb (x, y) = ax 7 + b 2 ‥‥‥ (1e) (1e) optical means have no term of y in the scanning direction in formula T5
This is because the optical characteristics are given to the projection optical system only in the x direction orthogonal to the scanning direction y direction. In the y direction in which scanning is performed, higher-order distortion is canceled due to averaging by scanning, and therefore, in the present embodiment, only higher-order distortion needs to be controlled in the x direction. The optical elements 61 and 62 are used by shifting at least one of them in the x direction to change the relative position in the same direction. However, since the shift causes high-order distortion, the shape of the aspheric surface is the seventh order or more of x. Is required.

In the initial state, the aspheric surface fa (x, y)
Since the concave and convex portions of the aspherical surface fb (x, y) coincide with each other, the optical elements 61 and 62 have zero optical power as a whole, and merely function as parallel plane plates. The smaller the distance between the optical elements 61 and 62 is, the better, for example, a value of about 100 μm is typical. Here, it is assumed that the optical element 61 is moved in the x direction by Δ.
The effect of this time, fa (x + Δ, y) -fb (x, y) = 7aΔx 6 + 21aΔ 2 x 5 + 35a Δ 3 x 4 + 35a Δ 4 x 3 + 21aΔ 5 x 2 + 7aΔ 6 x + Δ 7 + (b 1 −b ₂ ) ‥‥ (2e) Here, if the higher-order terms of Δ are ignored because the movement amount is small, the equation (2e) becomes simple, and fa (x + Δ, y) −fb (x, y) = 7aΔx ⁶ + (b ₁ −b ₂ ) ‥‥ (3e).

[0364] It (3e) equation has a term x ⁶ is a basis of the present embodiment. Therefore, the optical elements 61 and 62 become optical elements having sixth-order characteristics only in the x direction by the lateral shift amount Δ, and the characteristics can be freely controlled by the shift amount Δ.

Since the operation of shifting and taking the difference is the differentiation itself, it is the function of the optical elements 61 and 62 that the seventh order is given as the shape of the aspheric surface and the sixth order component is output by the effect of the differentiation. .

The absolute value of the amount of aspherical surface required to correct a small distortion value for the above-mentioned matching is very small. As a typical example, suppose that the amount of aspherical surface for generating a desired distortion is 1 μm, the diameter of the lens is 200 mm, and the lateral shift amount Δ at this time is 5 m.
Assuming m, from equation (3e), 7a × 5 × 10 ¹² = 0.001, and the value a = 2.86 × 10 ^-17 is obtained.

Numeral 100 indicates the radius value because the diameter is 200 mm. Therefore, the amount of the aspherical surface is 2.86 × (10 ⁻¹⁷ ) × (10 ¹⁴ ) = 2.86 × 10 ⁻³ from the expression (1e), and the original optical elements 61 and 62 have an aspherical amount of ± 2.86 μm. It has been.

To reduce the actual amount of deviation from the plane, it is advisable to add a term of cx which is a first-order term of x to this. 100mm
The value of c that gives a value of 2.86 μm at that time is 2.86 × 10 ^-5, so if a and c are reversed and c = −2.86 × 10 ^-5 , the deviation of the aspheric amount from the plane is up to ± 1.77 μm Can be reduced.

FIG. 27 shows the shape of the aspherical surface when c is zero.
FIG. 28 shows the shape when the above values are put in c. 200mm
Since the aspherical surface has such a gentle shape in the diameter of the optical element and the aspherical surfaces of the optical elements 61 and 62 are complementary to each other, the relative position of the optical elements 61 and 62 changes. The amount of aberration caused by the influence hardly affects other optical characteristics, and only high-order distortion can be corrected.

The present embodiment is characterized in that the correction direction of the higher-order distortion is set to a direction orthogonal to the scanning direction of the apparatus. However, the correction amount is small and the absolute value of the shift amount can be appropriately selected. Another point of the present embodiment is that the aspherical surface used can be reduced to an amount that can be easily measured by an interferometer.

In this embodiment, in order to generate a desired optical characteristic from the difference between the two aspheric surfaces, the amount of the original aspheric surface, that is, the amount of aspheric surface of the optical elements 61 and 62 itself is obtained in the final form. This value is larger than (difference). In the above example, a 2.86 μm aspherical surface is required to obtain a value of 1 μm. By optimizing the inclination, the size can be reduced to 1.77 μm. It is a key point in applying the present embodiment that these values enter the highly accurate measurement range of the interferometer.

In manufacturing an aspherical surface, it is important to be able to measure whether or not the surface has been accurately processed into a desired shape. However, if the amount can be suppressed to the extent of the present embodiment, the current technology is sufficient. . In order to further reduce this value, the shift amount set to 5 mm may be set to a larger value.

A command for driving the aspherical optical element may be changed in accordance with a procedure (data) stored in the apparatus in advance, or may be performed based on a measured value of an actual wafer. For example, if the higher-order distortion changes due to the difference in the illumination mode or NA, the amount of change is stored in advance, and when the illumination mode or NA is changed, the above-described aspherical optical element is shifted laterally. Good. At this time, at the same time, a distortion component of a lower order than a higher-order distortion, for example, a third-order distortion or a magnification component is also corrected by a known correction unit.

Also, instead of setting from the beginning as described above, the distortion component in the screen is analyzed using a reticle capable of measuring distortion, and the distortion from the magnification to the higher order is analyzed based on the value. Each may be corrected. In this case, the measurement may be automatically performed on the apparatus, or may be an off-line type in which an image of the reticle is once printed on the wafer and then measured.

FIG. 29 is a schematic view of a main part of Embodiment 11 of the present invention. This embodiment is different from the tenth embodiment in FIG. 23 only in that an optical unit T5 is provided between the projection optical system 2 and the wafer 3, and the other configuration is the same.

In the description of the tenth and eleventh embodiments, the case where only one of the two aspherical optical elements placed opposite to each other is moved to obtain desired characteristics has been described. The movement is not limited to this. For example, the upper optical element 1 is symmetrical with respect to the optical axis (in the opposite direction).
1 may be moved to the right by Δ and the lower optical element 12 may be moved to the left by Δ.

In the embodiment, the fifth-order distortion is considered as the higher-order distortion.
When considering even higher-order characteristics, the shape of the aspheric surface may be a shape obtained by adding a higher-order term, for example, a ninth-order or higher term.

As described above, Embodiments 10 and 11
With this technology, a high-order distortion component, which could not be corrected conventionally, can be variably corrected by a projection exposure apparatus. As a result, it is possible to variably adjust higher-order distortion components that could not be corrected conventionally, such as changes in distortion due to changes in the illumination mode and NA, distortion matching between devices, and errors in reticle creation. The overlay accuracy when doing so is greatly improved. Since there is a prediction that the positioning accuracy rather than the resolving power will be restricted at the time of fine processing of 256 MDRAM or later, it is possible to correct the higher-order components remaining as components that cannot be corrected conventionally. The effect is very large.

Further, since the high-order distortion component is variable, it is possible to cope with various fluctuations, and it is also a great advantage that the influence on other performances can be suppressed to a negligible value due to the small amount of aspherical surface.

In the tenth and eleventh embodiments, the direction in which higher-order distortion is generated is a direction orthogonal to the scanning direction of the scanning projection exposure apparatus. The distortion can be corrected in a direction in which the averaging effect does not exist in the scanning direction, so that the balance of the entire system is improved, and an effect that greatly contributes to the positioning accuracy as a whole can be obtained.

Up to this point, the correction of high-order distortion components has been performed only in the direction orthogonal to the scanning direction.
And when the width of the scanning slit is increased, ga (x, y) and the aspherical shape when applied to a stepper ^{= a (x 7 + 7x 6} y) + b 1 gb (x, y) = a ( x ⁷ + 7x ⁶ y) + b ₂ ‥‥‥ (4e). In the initial state, the aspheric surface ga (x, y) is an aspheric surface
Since gb (x, y) and the concavity and convexity are canceled to cancel each other, the optical power of the optical elements 61 and 62 is zero as a whole, and only functions as a plane-parallel plate. The smaller the distance between 61 and 62 is, the better, for example, a value of about 100 μm is typical. Here, it is assumed that the element 61 is moved in the x direction by Δ. If the effect at this time is ignored because the effect of the higher order term of Δ is small, then ga (x + Δ, y) −gb (x, y) = 7aΔ (x ⁶ + y ⁶ ) + (b ₁ −b ₂ ) ‥‥ ‥ (5e) The basis of the present invention is that the expression (5e) has a term of x ⁶ + y ⁶ . For this reason, the optical means T5 composed of the optical elements 61 and 62 becomes an optical system having a sixth order characteristic in rotational symmetry by the shift amount Δ, and the characteristic is changed by the lateral shift amount Δ
It can be controlled freely by using

Since the operation of shifting and taking the difference is the differentiation itself, the function of the optical elements 61 and 62 is to put the seventh order as the shape of the aspheric surface and to give the sixth order component by the effect of the differentiation. . Thereby, higher-order distortion can be controlled freely. The optical elements 61 and 6
The insertion position of 2 may be between the projection optical system and the reticle, or between the projection optical system and the wafer, as in the above embodiment. In some cases, the present element can be incorporated in a projection optical system.

A configuration may be adopted in which not only one of the optical elements 61 but also the optical elements 61 and 62 are moved symmetrically (in the opposite direction) with respect to the optical axis.

As described above, the present invention is characterized in that two surfaces represented by the same shape formula are arranged to face each other, and the two surfaces are shifted from each other by a desired amount to obtain desired optical characteristics. I have. The two surfaces complement each other in the reference state where the lateral shift is zero, and have no optical characteristics, and have optical characteristics that are generated by a differential effect with respect to the shift direction in the laterally shifted state.

That is, in a rectangular coordinate system (x, y) taken on a plane perpendicular to the optical axis of the optical system, if the direction of lateral displacement is the x direction, the surface shape p (x, y) of the present invention It is characterized by having a section that p (x, y) = ax n. When n is 3, it is possible to change various characteristics represented by the second-order terms, for example, the magnification and the focus position. Since the expression is one-way, the magnification that can be changed and the focus position can be changed. Specifically, a magnification difference between the x direction and the y direction and a focus difference between the x direction and the y direction pattern, that is, astigmatism, are to be controlled. Which aberration is controlled depends on the position where the element of the present invention is inserted, as described in the previous embodiment.

As described above, in the x direction, in the case of a scanner, a direction parallel to or orthogonal to the direction of the scanning slit is preferable. This is also the direction parallel or perpendicular to the sides of the rectangular area of the patterned active area of the mask. In addition, a method of moving the corresponding two elements so as to have a symmetrical movement amount with respect to the optical axis, or fixing one of them and moving the other can be applied.

In a scanner, astigmatism may occur while changing over time due to the absorption of exposure light even on the axis. Therefore, astigmatism correction when n is 3 is a distinctive feature of the element of the present invention. It is. The correction amount can be freely controlled to be positive or negative depending on the shift amount. Further, since the amount of displacement is a large value such as 5 mm as shown in the numerical example, it is not necessary to take a mechanically special high-precision structure with a loose set tolerance.

If n is set to 2 as another application example, the inclination of the image plane can be corrected.

The distortion can be easily corrected by inserting the present element near the reticle. When n is 4, 2 due to eccentricity
The following distortion, the third-order distortion when n is 5, the fourth-order distortion due to eccentricity when n is 6, and the fifth-order distortion when n is 7 Distortion can be corrected.

In order to control these aberrations independently, the elements of the present invention are required by the number to be controlled independently. When n is an odd number, x ^n-1 which is a function to be formed is not an even function. Therefore, the shift amount must be determined by paying attention to the fact that the shift effect differs in the positive direction and the negative direction of x.

If the known means can control, for example, magnification and third-order distortion, the present invention may be applied to control of components beyond the known means.

On the other hand, in the present invention, an example in which the shape of q (x, y) = a (x ⁿ + nxy ^n-1 ) is suitable as another shape q (x, y) has been shown. When the shape is q (x,
In the case of y), a characteristic based on x ^n-1 + y ^n-1 can be provided shifted in the x direction. The characteristic including y as compared with the case of p (x, y) is effective when the slit width is large in a stepper or a scanner and simply one-directional correction of x is insufficient.

As described above, in the x direction, in the case of a scanner, a direction parallel to or orthogonal to the direction of the scanning slit is preferable. This is also the direction parallel or perpendicular to the sides of the rectangular area of the patterned active area of the mask. As for the method of shifting, as in all the embodiments described above, a method in which the corresponding two elements are moved so as to be symmetrical with respect to the optical axis, or one is fixed and the other is moved.

It is a major feature of the present invention that the field curvature can be corrected when n is 3. If n is 5, higher order field curvature can be corrected.

If n is set to 2 as another application example, the inclination of the image plane can be corrected.

The distortion can be easily corrected by inserting the present element near the reticle. When n is 4, 2 due to eccentricity
The following distortion, the third-order distortion when n is 5, the fourth-order distortion due to eccentricity when n is 6, and the fifth-order distortion when n is 7 Distortion can be corrected.

In order to control these aberrations independently,
The number of the elements of the present invention is required to be controlled independently. Further, when n is an odd number, the function formed is x ^n-1 + y ^n-1
Is not an even function, the shift amount has to be determined in consideration that the shift effect differs in the positive and negative directions of x.

Next, an embodiment of a method of manufacturing a semiconductor device using the above-described projection exposure apparatus will be described.

FIG. 30 is a flowchart for manufacturing a semiconductor device (a semiconductor chip such as an IC or LSI, or a liquid crystal panel or a CCD).

Step 1 (circuit design) in this embodiment
Now, the circuit design of the semiconductor device will be performed. Step 2
In (mask production), a mask on which a designed circuit pattern is formed is produced.

In step 3 (wafer manufacture), a wafer is manufactured using a material such as silicon. Step 4
The (wafer process) is called a pre-process, and an actual circuit is formed on the wafer by lithography using the prepared mask and wafer.

The next step 5 (assembly) is called a post-process, and is a process of forming a semiconductor chip using the wafer produced in step 4, and includes an assembly process (dicing and bonding) and a packaging process (chip encapsulation). And the like.

In step 6 (inspection), inspections such as an operation confirmation test and a durability test of the semiconductor device manufactured in step 5 are performed. Through these steps, a semiconductor device is completed and shipped (step 7).

FIG. 31 is a detailed flowchart of the wafer process in step 4 described above. First, in step 11 (oxidation), the surface of the wafer is oxidized. Step 12 (C
In VD), an insulating film is formed on the wafer surface.

In step 13 (electrode formation), electrodes are formed on the wafer by vapor deposition. In step 14 (ion implantation), ions are implanted into the wafer. Step 15
In (resist processing), a photosensitive agent is applied to the wafer. Step 16 (exposure) uses the above-described exposure apparatus to print and expose the circuit pattern of the mask onto the wafer.

In step 17 (developing), the exposed wafer is developed. In step 18 (etching), portions other than the developed resist are removed. In step 19 (resist stripping), the resist that has become unnecessary after the etching is removed. By repeating these steps, multiple circuit patterns are formed on the wafer.

By using the manufacturing method of this embodiment, a highly integrated semiconductor device can be easily manufactured.

[0408]

As described above, according to the present invention, a change in optical performance due to absorption of exposure light by the projection optical system, for example, a wafer caused by axial astigmatism, curvature of field, a process factor of the wafer, etc. Minimizing the influence on other optical characteristics by utilizing an optical means having at least two optical elements having an aspheric surface of a shape in which at least one of its own asymmetric magnification and distortion is appropriately set. Thus, it is possible to achieve a projection exposure apparatus capable of easily obtaining a high-resolution pattern by performing correction while pressing, and a device manufacturing method using the same.

In addition to correcting the expansion / contraction state after obtaining each process of the wafer on which the fine pattern is to be printed, distortion matching for the apparatus and correction of drawing errors of the reticle are also performed appropriately. Using optical means having at least two optical elements having an aspherical surface,
Projection exposure apparatus and device using the same, which are made possible by independently controlling the magnifications in the x direction and the y direction orthogonal to each other and minimizing the influence on other optical characteristics caused by the control. Can be achieved.

[Brief description of the drawings]

FIG. 1 is a schematic diagram of a main part of a first embodiment of the present invention.

FIG. 2 is an enlarged explanatory view of a part of FIG. 1;

FIG. 3 is an enlarged explanatory view of a part of FIG. 1;

FIG. 4 is an explanatory view of another embodiment of a part of FIG. 1;

FIG. 5 is an explanatory view of another embodiment of a part of FIG. 1;

FIG. 6 is an explanatory view of another embodiment of a part of FIG. 1;

FIG. 7 is a schematic diagram of a main part of a second embodiment of the present invention.

FIG. 8 is an explanatory diagram of an optical unit used in Embodiment 3 of the present invention.

FIG. 9 is a schematic diagram of a main part of a fourth embodiment of the present invention.

FIG. 10 is an enlarged explanatory view of a part of FIG. 9;

FIG. 11 is an enlarged explanatory view of a part of FIG. 9;

FIG. 12 is an explanatory view of another embodiment of a part of FIG. 9;

FIG. 13 is an explanatory diagram of another embodiment of a part of FIG. 9;

FIG. 14 is an explanatory view of another embodiment of a part of FIG. 9;

FIG. 15 is a schematic view of a main part of a fifth embodiment of the present invention.

FIG. 16 is a schematic view of a main part of a sixth embodiment of the present invention.

FIG. 17 is an enlarged explanatory view of a part of FIG. 16;

FIG. 18 is an enlarged explanatory view of a part of FIG. 16;

FIG. 19 is an explanatory view of another embodiment of a part of FIG. 9;

FIG. 20 is an explanatory view of another embodiment of a part of FIG. 9;

FIG. 21 is an explanatory view of another embodiment of a part of FIG. 9;

FIG. 22 is a schematic view of a main part of a seventh embodiment of the present invention.

FIG. 23 is a schematic view of a main part of an eighth embodiment of the present invention.

FIG. 24 is an enlarged explanatory view of a part of FIG. 23;

FIG. 25 is a schematic view of a main part of a ninth embodiment of the present invention.

FIG. 26 is a schematic view of a main part of a tenth embodiment of the present invention.

FIG. 27 is an enlarged explanatory view of a part of FIG. 26;

FIG. 28 is an enlarged explanatory view of a part of FIG. 26;

FIG. 29 is a schematic view of a main part of an eleventh embodiment of the present invention.

FIG. 30 is a flowchart of a method of manufacturing a semiconductor device according to the present invention.

FIG. 31 is a flowchart of a method for manufacturing a semiconductor device according to the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 1st object (reticle) 2 Projection optical system 3 2nd object (wafer) 4 Exposure illumination system 5 Wafer holder 6 Wafer stage 7 Interference mirror T1, T2, T3, T4, T5 Optical means 11, 12, 21, 22 , 23,31,32,41,4
2,51,52,61,62 Optical element

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int. Cl. ⁷ , DB name) H01L 21/027 G03F 7/20

Claims (26)

(57) [Claims] A projection for projecting a mask pattern onto a substrate.
A projection optical system , wherein at least one of the projection optical systems is
A pair of aspherical members that can be displaced in a direction perpendicular to the optical axis
Has the pair of aspherical members, each aspherical surface facing
And the shape of each aspherical surface of the pair of aspherical members is
A positional relationship between the aspheric surfaces of the pair of aspheric members in the orthogonal direction.
By changing the relationship, the axial astigmatism of the projection optical system is changed.
Is determined to change, and the at least one aspherical member is moved in a direction orthogonal to the optical axis.
Axial astigmatism of the projection optical system
A projection exposure apparatus, wherein 2. An aspherical surface of each of said pair of aspherical members.
The shape is expressed by an equation, assuming that the direction of the movement is the x direction.
2. The method according to claim 1, further comprising a third-order term of x.
Projection exposure equipment. 3. The pair of aspherical members are arranged in opposite directions to each other.
2. The projection device according to claim 1, wherein the moving member moves in the orthogonal direction.
Shadow exposure device. 4. A method for scanning the mask and the substrate while scanning the substrate.
Projection exposure, and the x direction is the direction of the scanning.
3. The projection exposure according to claim 2, wherein the direction is orthogonal to the direction.
apparatus. 5. A projection for projecting a mask pattern onto a substrate.
A projection optical system , wherein at least one of the projection optical systems is
A pair of aspherical members that can be displaced in a direction perpendicular to the optical axis
Has the pair of aspherical members, each aspherical surface facing
And the shape of each aspherical surface of the pair of aspherical members is
A positional relationship between the aspheric surfaces of the pair of aspheric members in the orthogonal direction.
By changing the relationship, the field curvature of the projection optical system changes.
The at least one aspherical member in a direction orthogonal to the optical axis.
To adjust the curvature of field of the projection optical system.
A projection exposure apparatus, comprising: 6. An aspherical surface of each of said pair of aspherical members.
The shape is expressed by an equation, assuming that the direction of the movement is the x direction.
And it includes a third order term of x when the, also the optical characteristics is the x direction
The field curvature related to the direction is included.
5. Projection exposure apparatus. 7. The pair of aspherical members are arranged in opposite directions to each other.
6. The projection device according to claim 5, wherein the moving member moves in the orthogonal direction.
Shadow exposure device. 8. While scanning the mask and the substrate,
Projection exposure, and the x direction is the direction of the scanning.
7. The projection exposure according to claim 6, wherein the direction is orthogonal to the direction.
apparatus. 9. A projection for projecting a mask pattern onto a substrate.
A projection optical system , wherein at least one of the projection optical systems is
A pair of aspherical members that can be displaced in a direction perpendicular to the optical axis
Has the pair of aspherical members, each aspherical surface facing
And the shape of each aspherical surface of the pair of aspherical members is
A positional relationship between the aspheric surfaces of the pair of aspheric members in the orthogonal direction.
When the relationship is changed, the projection magnification of the projection optical system is changed.
And the at least one aspherical member is oriented in a direction orthogonal to the optical axis.
To adjust the projection magnification of the projection optical system.
A projection exposure apparatus, comprising: 10. An aspherical surface of each of said pair of aspherical members.
Is expressed by an equation, assuming that the direction of the movement is the x direction.
When passed, it includes a third-order term of x, and the optical characteristic is x
The method according to claim 1, wherein the projection magnification is related to a direction.
Item 10. The projection exposure apparatus according to Item 9. 11. The pair of aspherical members have opposite directions.
And moving in the orthogonal direction.
Projection exposure equipment. 12. While scanning the mask and the substrate,
The projection exposure is performed, and the x direction is
11. The projection of claim 10, wherein the projection is orthogonal to the direction.
Exposure equipment. 13. A pattern of a mask is projected onto a substrate.
A projection optical system, wherein at least one of the projection optical systems is a projection optical system.
A pair of aspherical members that can be displaced in a direction perpendicular to the optical axis
Has the pair of aspherical members, each aspherical surface facing
And the shape of each aspherical surface of the pair of aspherical members is
A positional relationship between the aspheric surfaces of the pair of aspheric members in the orthogonal direction.
By changing the engagement, the distortion of the projection optical system changes.
The at least one aspherical member in a direction orthogonal to the optical axis.
To adjust the distortion of the projection optical system.
A projection exposure apparatus. 14. An aspherical surface of each of the pair of aspherical members.
Is expressed by an equation, assuming that the direction of the movement is the x direction.
The method according to claim 1, further comprising a fifth order term of x when passed.
13. A projection exposure apparatus. 15. An aspherical surface of each of said pair of aspherical members.
Is expressed by an equation, assuming that the direction of the movement is the x direction.
The method of claim 7, further comprising the seventh order term of x.
13. A projection exposure apparatus. 16. The pair of aspherical members have opposite directions.
14. The moving means in the orthogonal direction.
Projection exposure equipment. 17. While scanning the mask and the substrate,
The projection exposure is performed, and the x direction is
14. The method according to claim 14, wherein the direction is orthogonal to the direction.
5. Projection exposure apparatus. 18. The method according to claim 18 , further comprising the step of :
A second pair of transparent aspherical portions whose surfaces face each other
The second pair of aspherical members are orthogonal to the optical axis.
Moving in the y direction, which is a direction
And the second pair of aspherical members move.
Before projecting the mask pattern onto the substrate
The field curvature in the y direction changes, and the second pair of
The shape of each aspherical surface of the aspherical member is expressed as
7. The projection dew of claim 6 including a third order term of y.
Light device. 19. Using the two pairs of aspherical members,
Image plane bay when projecting the mask pattern onto the substrate
2. The method according to claim 1, wherein the on-axis astigmatism and the curvature are adjusted.
8. Projection exposure apparatus. 20. The second pair of aspherical members are connected to each other.
Moving in the opposite direction in the orthogonal direction.
Item 18. The projection exposure apparatus according to Item 18. 21. Each of the two pairs of aspherical members
But wherein the moving before Symbol perpendicular direction in opposite directions
19. The projection exposure apparatus according to claim 18, wherein 22. One of said two sets of aspherical members
The member is a single member with both surfaces being aspherical for both groups.
19. The projection exposure apparatus according to claim 18, wherein: 23. A pattern of a mask is projected onto a substrate.
A projection optical system, wherein at least one of the projection optical systems is a projection optical system.
A pair of aspherical members that can be displaced in a direction perpendicular to the optical axis
Has the pair of aspherical members, each aspherical surface facing
And the shape of each aspherical surface of the pair of aspherical members is
A positional relationship between the aspheric surfaces of the pair of aspheric members in the orthogonal direction.
One system of the pair of aspherical members by changing the engagement
Is determined so that the refractive power of the at least one aspherical member is changed in a direction orthogonal to the optical axis.
Adjust the refractive power of the projection optical system by displacing
And a projection exposure apparatus. 24. When the positional relationship is a predetermined relationship,
The above-mentioned system so that the refractive power is zero.
The shape of the aspherical surface is determined.
Projection exposure equipment. 25. The facing aspherical surface includes a pair of aspherical surfaces.
Shape that matches when the spherical member is in the predetermined positional relationship
25. The method according to claim 1, wherein the shape has a shape.
The projection exposure apparatus according to claim 1. 26. A according to any one of claims 1 25
Device exposure onto a substrate
Characterized in that it comprises the steps of shooting, a step of developing the substrate on which copy said transfer
Device manufacturing method.