JPH09326344A - Exposure method - Google Patents

Abstract

(57) Abstract: The resolution is further improved without shortening the wavelength of the illumination light for exposure while satisfying the conditions regarding the margin of the depth of focus. Under the illumination light IL from a light source, a reticle R and a wafer W are projected onto a projection optical system PL in a state where a pattern image of a reticle R is projected onto a wafer W via a projection optical system PL. On the other hand, synchronous scanning is performed to perform exposure. The focus control accuracy, the level difference of the underlying pattern, and the margin of the depth of focus that the sum of the widths obtained by multiplying the photoresist by a predetermined coefficient are smaller than the depth of focus can be satisfied even if the numerical aperture of the projection optical system is increased. First, thin the photoresist. When the photoresist is thinned, it is necessary to change the target value of the integrated exposure amount according to the uneven thickness of the photoresist. Therefore, the resist thickness measuring device 46 measures the resist thickness, and the integrated exposure amount is calculated based on the measurement result. Control.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention transfers a mask pattern onto a photosensitive substrate in a photolithography process for manufacturing, for example, a semiconductor element, an image pickup element (CCD or the like), a liquid crystal display element, a thin film magnetic head or the like. And an exposure method for doing the same.

[0002]

2. Description of the Related Art When manufacturing a semiconductor device or the like, a wafer coated with a photoresist as a photosensitive substrate is formed through a projection optical system to form an image of a reticle pattern as a mask under illumination light for exposure. Or, an exposure device for transferring onto a glass plate or the like is used. Conventionally, as the exposure apparatus, a batch exposure type projection exposure apparatus such as a stepper has been mainly used, but recently, in order to transfer a large area pattern without excessively burdening the projection optical system, A scanning exposure type projection exposure apparatus such as a step-and-scan method in which a reticle and a wafer are synchronously scanned with respect to a projection optical system is also receiving attention.

FIG. 17 shows a schematic structure of an example of a conventional projection exposure apparatus. In FIG. 17, the exposure illumination light from the illumination optical system 101 forms a pattern formation area on the pattern formation surface (lower surface) of the reticle R. Lighting up. Under the illumination light, a projection image obtained by reducing the pattern formed on the reticle R at a predetermined magnification via the projection optical system 102 is a wafer W.
The photoresist coated above is exposed. The numerical aperture of the projection optical system 102 is defined by the aperture stop 103. Then, by developing the photoresist, the projected image appears as an uneven resist pattern.

The resolution R and the depth of focus D of the image of the pattern of the reticle R image-projected on the wafer W as described above are respectively expressed by the following equations. R = k ₁ · λ / NA (1) D = k ₂ · λ / NA ² (2) where k ₁ and k ₂ are process coefficients (hereinafter,
“K factor”), λ is the wavelength of the exposure illumination light, N
A is the numerical aperture of the projection optical system 102. As shown in FIG. 17, assuming that the half aperture angle of the image-forming light beam on the wafer side (image side) of the projection optical system 102 is θ _{PL 2} and the surrounding gas is air, the numerical aperture NA is represented by sin θ _PL2 . To be done. Further, the projection magnification of the projection optical system 102 from the reticle R to the wafer W is β (β is ⅕, ¼, etc.), the projection optical system 1
_Letting θ _{PL1 be} the opening half angle on the reticle side (object side) of the image formation light flux by 02, there is the following relationship between the opening half _angles θ _PL2 and θ _PL1 . At this time, among the diffracted light from the reticle R, the diffracted light having a diffraction angle of θ _PL1 or less contributes to the formation of the projected image of the pattern of the reticle R.

Sin θ _PL1 = β · sin θ _PL2 = β · NA (3) Further, the illumination half-angle of the exposure illumination light from the illumination optical system 101 to the reticle R is θ _IL , and the illumination optical system 101
Let σ (σ value) be the coherence factor of, ie, the ratio of the numerical aperture (= sin θ _IL ) on the exit side of the illumination optical system 101 to the numerical aperture (= sin θ _PL1 ) on the incident side of the projection optical system 102. , The aperture half angle θ _IL of the illumination light is expressed as follows using the equation (3). The range of the σ value is 0 ≦ σ <1.
It is.

Sin θ _IL = σ · sin θ _PL1 = σ · β · NA (4) Then, the values of the k factors k ₁ and k ₂ that determine the resolution R and the depth of focus D are the coherence factors. It depends on the σ value. Further, the values of the k factors k ₁ and k ₂ are different from those of the normal illumination method in which the shape of the secondary light source is circular in the illumination optical system 101, and the so-called plural light sources in which the shape of the secondary light source is decentered from the optical axis. It is changed by switching to the modified illumination method or the annular illumination method, and a predetermined optical filter (so-called pupil filter) is provided on the pupil surface of the projection optical system 102 (optical Fourier transform surface with respect to the pattern formation surface of the reticle R). It also changes depending on the placement. Further, the values of the k factors k ₁ and k ₂ are, as the reticle R, a so-called phase shift reticle that gives a predetermined phase difference in a periodic pattern, or a so-called phase shift reticle that gives a predetermined transmittance distribution in the periodic pattern. It also changes when using a halftone reticle. Due to various conditions as described above,
The values of k factors k ₁ and K ₂ vary in the following ranges.

0.45 ≦ k ₁ ≦ 0.6, 0.7 ≦ k ₂ ≦ 2.0 (5) Therefore, conventionally, by making full use of the above various conditions, the depth of focus D is deep and the resolution is high. The exposure condition with a high R is used properly according to the device to be exposed. Also,
The depth of focus D is expressed by the equation (2). The focus control precision C _F which is the precision when the wafer is focused by the projection exposure apparatus by the auto focus method is C _F , and the step of the underlying pattern on the wafer (hereinafter, alignment). _Assuming that the mark step is representatively referred to as “mark step”) is D _M , the focus control accuracy C _F and the mark step D _{M are} calculated from the depth of focus D.
The width (= D−C _F −D _M ) obtained by subtracting is the width of the substantial depth of focus that can be used on the operator side of this projection exposure apparatus, that is, the usable depth of focus.

In practice, assuming that the thickness of the photoresist on the wafer (resist thickness) is T _R , in order to obtain a good resist pattern after development of the photoresist, the resist thickness T _R is set to a predetermined value. Width obtained by multiplying the following coefficient K _UD (depth of focus width required for photoresist exposure)
However, it only needs to be within the range of the depth of focus that can be used. Therefore, in the following, the width T _R · K _UD obtained by multiplying the resist thickness T _R by the coefficient K _UD is used as the usable depth of focus UD.
Called OF (Usable Depth Of Focus). From the above,
The depth of focus D in the equation (2) is the focus control accuracy C _F , the mark step D _M , and the usable depth of focus U as in the following equation.
It must be wider than the sum of DOF (= T _R · K _UD ). The condition of the following expression (6) is referred to as “condition regarding margin of depth of focus”.

C _F + D _M + T _R · K _UD ≦ D = k ₂ · λ / NA ² (6) As a general example, the illumination light for exposure is i of a mercury lamp.
The line (λ = 365 nm) is used and the k factor k ₁ is 0.
As 55, an image of a line-and-space pattern (L / S pattern) having a line width of 0.4 μm, which corresponds to the pattern of a 64 Mbit DRAM, is projected, that is, a projection for making the resolution R 0.4 μm. Numerical aperture N of optical system
When A is NA ₁ , the numerical aperture NA ₁ is as follows from the equation (1).

0.4 = 0.55 · 0.365 / NA ₁ , NA ₁ ≈0.50 (7) Further, the mark step D _M on the wafer is 0.8 μm, and the resist thickness T _R is 1 μm. When the coefficient K _UD that determines the depth of focus UDOF is 0.7 and the k factor k ₂ is 2.0, the numerical aperture NA of the projection optical system that satisfies the condition regarding the margin of the depth of focus in Expression (6) is NA. _{If 2} ,
The numerical aperture NA ₂ is as follows.

0.5 + 0.8 + 1 × 0.7 ≦ 2.0 × 0.365 / NA ₂ ² , NA ₂ ≦ 0.60 (8) From the equations (7) and (8), the i-line of the mercury lamp is obtained. The condition of the numerical aperture NA of the projection optical system for manufacturing a pattern of a 64 Mbit DRAM using is 0.5 ≦ NA ≦ 0.6. Further, in order to set the k factors k ₁ and k ₂ to the above values, it is desirable to set the σ value which is the coherence factor to about 0.7 ≦ σ <0.8. Furthermore, it has been known that the resolution can be improved without narrowing the depth of focus by applying the modified illumination method or the annular illumination method to the periodic pattern than in the past. It has also been attempted to use the modified illumination method and the annular illumination method together.

[0012]

As described above, the i-line of the mercury lamp is used as the illumination light for exposure, and the numerical aperture NA of the projection optical system is set to about 0.5≤NA≤0.6. Thus, a circuit image corresponding to a 64 Mbit DRAM, that is, an image of a pattern having a resolution R of about 0.4 μm can be transferred onto a wafer with high accuracy. Furthermore, 256 Mbit DRAM, which is a next-generation semiconductor device recently
Is about to be started, and for that purpose, it is necessary to increase the resolution R to about 0.25 μm. In order to increase the resolution R as described above, the wavelength λ of the exposure illumination light may be shortened or the numerical aperture NA of the projection optical system may be increased according to the equation (1).

However, if the wavelength λ is simply shortened,
When the numerical aperture NA is increased, the depth of focus D in the equation (2) becomes narrower, and the condition regarding the margin of the depth of focus in the equation (6) cannot be satisfied. In particular, as can be seen from the expression (2), the depth of focus D decreases in inverse proportion to the square of the numerical aperture NA, and thus it has not been conventionally considered to increase the numerical aperture NA. Therefore, conventionally, as an exposure light source, a KrF excimer laser light source (exposure wavelength λ = 248n
m) is used to obtain a resolution R of 0.25 μm, and (b) introduction of wafer flattening technology and (b) improvement of focus control technology satisfy the condition regarding the depth-of-focus margin of equation (6). Was doing.

The wafer flattening technique (a) is a technique for physically or chemically increasing the flatness of the surface of the wafer before exposure in order to reduce the warp of the wafer and the step of the circuit pattern. As a result, the mark step D _M is reduced from the conventional 0.8 μm to about 0.1 μm. Also,
Due to the improvement of the focus control technique (b), the focus control accuracy C _F is improved from the conventional 0.5 μm to about 0.4 μm. Estimating the numerical aperture NA of the projection optical system under these exposure conditions, _first , assuming that the k factor k ₁ is 0.55, the numerical aperture NA ₁ that satisfies (1) is as follows.

0.25 = 0.55 × 0.248 / NA ₁ , NA ₁ ≈0.55 (9) Further, the numerical aperture NA _{2 which} satisfies the condition regarding the depth of focus margin of the equation (6) is as follows. become. The k factor k ₂ is 1.7 because the line width is thin, and the coefficient K _UD that determines the usable depth of focus UDOF remains 0.7.

0.4 + 0.1 + 1 × 0.7 ≦ 1.7 × 0.248 / NA ₂ ² , NA ₂ ≦ 0.59 (10) That is, the numerical aperture NA range is 0.55 under these exposure conditions.
≦ NA ≦ 0.6. Further, in order to further improve the resolution without narrowing the depth of focus, it is also considered to use a modified illumination method and a halftone reticle together.

However, if the exposure light source is changed from a mercury lamp to a KrF excimer laser light source in order to manufacture a semiconductor device equivalent to a 256 Mbit DRAM, the projection exposure apparatus itself becomes expensive, and it is difficult to receive the return corresponding to the investment. There was an inconvenience that cost performance was bad. Further, under the present circumstances, ArF is used as an exposure light source as a projection exposure apparatus for forming a circuit pattern corresponding to the next-generation semiconductor device, 1 Gbit DRAM.
Development of a device using an excimer laser light source (wavelength λ is 193 nm) is planned, and next 4G bit DR
As a projection exposure apparatus for forming a circuit pattern corresponding to AM, at present, there is also a plan to develop an apparatus using exposure light of shorter wavelength, such as X-rays. That is,
Currently, in order to manufacture next-generation semiconductor devices,
The main idea is to shorten the wavelength of the illumination light for exposure to deal with it. However, when the exposure light source is switched to shorten the wavelength of the illumination light for exposure, there is a disadvantage that the manufacturing cost of the projection exposure apparatus increases significantly and the cost performance deteriorates.

In view of the above point, the present invention provides an exposure method capable of further improving the resolution without shortening the wavelength of the exposure illumination light and satisfying the condition concerning the margin of the depth of focus. To aim. In other words, the present invention can form a pattern corresponding to the next-generation 256 Mbit DRAM by using the i-line of a mercury lamp as the illumination light for exposure, and 1 G by using the KrF excimer laser light.
It is an object of the present invention to provide an exposure method capable of forming a pattern corresponding to a bit DRAM and forming a pattern corresponding to a 4G bit DRAM using ArF excimer laser light.

[0019]

According to a first exposure method of the present invention, a substrate (W) coated with a photosensitive material on a transfer pattern on a mask (R) under exposure illumination light (IL). In the exposure method of transferring and exposing the light onto the substrate (W), the exposure illumination light (I
L) controls the integrated exposure amount.

In the present invention, if the projection optical system is used, the wavelength λ of the illumination light for exposure is not shortened in order to improve the resolution of the transferred image.
The numerical aperture NA of the projection optical system in the equation (1) is increased. However, if the numerical aperture NA is simply increased, (2)
From the formula, the depth of focus D becomes narrower in inverse proportion to its square,
The condition concerning the margin of the depth of focus in the equation (6) is not satisfied. Therefore, in the present invention, in order to satisfy the condition of the expression (6), the thickness T _R of the photosensitive material (photoresist) in the expression (6) is reduced so that the usable depth of focus UDO.
Make F (= T _R · K _UD ) narrower. However, when reducing the thickness T _R of the photosensitive material, the integrated exposure amount required in order to properly expose the photosensitive material is reduced, since the proper exposure time is shortened, high-precision control of the integrated exposure amount Will be difficult.

For example, when applying a thin photosensitive material,
Temperature changes, humidity changes, storage time of photosensitive materials, etc.
Therefore, uneven thickness of the photosensitive material may occur.
As a result, the exposure time for the photosensitive material also changes.
Therefore, for example, the average thickness information from the coater of the photosensitive material
Product of exposure illumination light for individual substrates based on
The calculated exposure dose is controlled with high accuracy for each target value.
It is difficult. In order to deal with it, in the present invention, for example,
Immediately before exposure, the thickness T of the photosensitive material on the substrate (W) _RThe actual
The total exposure amount can be controlled based on this thickness.
And As a result, the numerical aperture NA of the projection optical system is increased.
The depth of focus becomes narrower as the resolution is improved.
However, the condition regarding the depth of focus margin is satisfied, and
A proper integrated exposure amount can be obtained. Also, control the integrated exposure amount
To achieve this, a stepper-type projection exposure system
Control the exposure time and exposure time.
In the de-scan type projection exposure apparatus, in addition to illuminance control,
The scanning speed of the substrate (W) and the width of the slit-shaped illumination area
The integrated exposure amount can also be controlled by controlling.

In this case, an appropriate integrated exposure amount for the photosensitive material having a predetermined reference thickness corresponding to the mask (R) is stored in advance, and the thickness of the photosensitive material is used as the predetermined reference. It is desirable to obtain a difference from the thickness and set the integrated exposure amount of the exposure illumination light (IL) based on this difference. As a result, even when the thickness of the photosensitive material changes, the proper integrated exposure amount can be obtained by a simple calculation.

Next, when the transfer pattern on the mask (R) is projected and exposed on the substrate (W) through the projection optical system (PL), the photosensitive material applied on the substrate (W). Is 0.5 μm or less, the projection optical system (P
It is desirable that the numerical aperture of L) be 0.68 or more. In the conventional exposure method, if the numerical aperture is simply set to 0.6 or more,
(6) The conditions regarding margin of the depth of focus is no longer satisfied, when the thickness T _R of photosensitive material is 0.5μm or less, the improvement and combined planarization techniques focus control techniques, numerical aperture 0 The condition of the expression (6) is satisfied from 0.68 to 0.8. As a result, the resolution of a pattern that can be transferred without shortening the wavelength of the exposure illumination light, that is, without using an expensive exposure light source, can be increased by about one generation, and the condition regarding the depth of focus margin is also satisfied. Be done.

Next, in the second exposure method according to the present invention, the transfer pattern on the mask (R) is transferred onto the substrate (R) coated with the photosensitive material under the exposure illumination light (IL). In the exposure method of exposing, the switching of the numerical aperture of the projection optical system (8) is controlled according to the thickness of the photosensitive material applied on the substrate (R). According to the present invention, for example, when the thickness of the photosensitive material coated on the substrate (W) is 0.2 μm or less, the numerical aperture of the projection optical system (PL) is 0.7 or more and the substrate (W ) When the thickness of the photosensitive material coated on the substrate is 1.0 μm or more, the projection optical system (PL)
The numerical aperture of is switched to 0.6 or less. For example, when the exposure illumination light is KrF excimer laser light (wavelength is 248
nm), 256 Mbit DRAM with numerical aperture of 0.6 or less
A considerable pattern (resolution of about 0.25 μm) can be transferred with high accuracy, and a 1 Gbit DRAM with a numerical aperture of 0.7 or more.
A considerable pattern (resolution is about 0.18 μm) can be transferred with high accuracy. As a result, a pattern corresponding to a second-generation semiconductor device can be transferred with high accuracy by one exposure apparatus.

[0025]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of an exposure method according to the present invention will be described below with reference to the drawings. In this example, the present invention is applied to the case where exposure is performed by a step-and-scan type projection exposure apparatus. FIG.
2 is a schematic configuration diagram showing a projection exposure apparatus of this example. In FIG. 1, the illumination light IL made up of a laser beam having a wavelength of 248 nm pulsed from a KrF excimer laser light source 1 as an exposure light source is a dimming unit. After being dimmed at a predetermined dimming rate at 2, the beam is expanded through a beam expander including a first lens 3, a deflection mirror 4 and a second lens 6 to
It is incident on the fly-eye lens 7 of the stage. The second lens 6 and the fly-eye lens 7 constitute an input optical system 5A for normal illumination, and this input optical system 5A can be exchanged with an input optical system 5B for annular zone illumination described later via the exchange device 8. Is configured.

The illumination light IL whose illuminance distribution is made uniform by the first-stage fly-eye lens 7 is an illuminance distribution correction plate 9 made of parallel flat glass for correcting unevenness of the illuminance distribution depending on the direction of the illumination light, and a relay lens. 10 and deflection mirror 11
Through the fly eye lens 12 for normal lighting in the second stage
Is incident on the incident surface of. The plane of incidence of the first-stage fly-eye lens 7 is conjugate with the plane of incidence of the second-stage fly-eye lens 12, and the planes of incidence of the latter fly-eye lens 12 are the lenses forming the former fly-eye lens 7. A magnified image of the entrance surface of the element is superimposed. In addition, the fly-eye lens 12 is configured to be replaceable with a fly-eye lens 13 for second-stage annular illumination, which will be described later, via an exchange device 14. Hereinafter, an optical axis of the illumination optical system after the second-stage fly-eye lens 12 and the optical axis of the projection optical system PL is the optical axis AX, the Z axis is taken in parallel with the optical axis AX, and is shown in a plane perpendicular to the Z axis. The description will be given by taking the Y axis parallel to the plane of FIG. 1 and the X axis perpendicular to the plane of FIG.

At this time, the second-stage fly-eye lens 12
Is an optical Fourier transform surface (pupil surface) with respect to the pattern formation surface of the reticle R, and a plurality of illumination system aperture stops (hereinafter referred to as “σ stop”) are provided on this exit surface. The σ diaphragm unit 15 is arranged. Figure 7
FIG. 7A shows the σ diaphragm unit 15, which is shown in FIG.
In (a), on the σ diaphragm unit 15 made of a disc rotatable about the drive shaft 15a, a σ diaphragm 75A for a large σ value and a σ diaphragm 75 for a small σ value are arranged at equal angular intervals.
B, a σ diaphragm 75C for the first annular illumination, and a σ diaphragm 75D for the second annular illumination are formed. σ diaphragm 75A
Details of the shapes of ~ 75D will be described later.

Returning to FIG. 1, in response to a main controller 29 that controls the operation of the entire apparatus instructing the illumination system control system 32 about exposure conditions and the like, the illumination system control system 32 causes the σ diaphragm unit 15 to operate. Is rotated to set a predetermined σ stop on the exit surface of the fly-eye lens 12. At the same time as the operation, the illumination system control system 32 replaces the two input optical systems 5A and 5B via the exchanging device 8 if necessary, and also the two fly-eye lenses 12 and 1 via the exchanging device 14.
Also exchange 3.

Illumination light IL emitted from the fly-eye lens 12 and passing through a predetermined σ diaphragm in the σ diaphragm unit 15.
Enters the fixed field stop (fixed reticle blind) 18 that is sequentially driven in the Z tilt direction and the movable field stop 20 that is driven in the XY directions via the split prism plate 16 and the first relay lens 17. In this example, the movable field stop 2
0 is arranged on a conjugate surface with the pattern forming surface of the reticle R, and the fixed field stop 18 is arranged on a surface slightly defocused from the conjugate surface. And fixed field stop 1
A rectangular fixed aperture for defining a slit-shaped illumination area on the reticle R is formed at 8, and the position and tilt angle (tilt angle) of the fixed field stop 18 in the Z direction are determined by the illumination system control system 32. Can be finely adjusted via the drive unit 19.

On the other hand, the movable field stop 20 is a pair of movable blades 20a and 20b which can move independently in the Y direction.
And a pair of movable blades (not shown) that can be independently driven in the X direction.
The operation such as 0b is performed by the illumination system control system 32 via the drive unit 21.
Controlled by. When scanning exposure is performed on each shot area on the wafer W, the illumination light IL is also applied to the area other than the pattern area on the reticle R immediately after the scanning exposure starts and immediately before the scanning exposure with only the fixed field stop 18. Sometimes. Therefore, the movable field stop 20 plays a role of further limiting the illumination area on the reticle R defined by the fixed field stop 18 immediately after the start of scanning exposure and immediately before the end of scanning exposure.

Fixed field diaphragm 18 and movable field diaphragm 20
The illumination light IL that has passed through the opening of the second relay lens 2
The slit-shaped illumination area 24 on the pattern formation surface (lower surface) of the reticle R is illuminated with a uniform illuminance distribution via the condenser lens 23 and the condenser lens 23. The illumination area 24 in this example is X
This is a rectangular area whose direction is the longitudinal direction, and the scanning direction of the reticle R during scanning exposure is the Y direction parallel to the paper surface of FIG. Under the illumination light IL, the pattern in the illumination area 24 on the reticle R is reduced by the projection optical system PL at the projection magnification β (β is 1/4, 1/5, etc.), and the photoresist is applied. The image is projected onto the slit-shaped exposure area 33 on the surface of the wafer W. A variable aperture stop 42 in the form of, for example, an iris diaphragm is installed on the pupil plane (optical Fourier transform plane for the pattern formation surface of the reticle R) in the projection optical system PL, and the main controller 29 is variable via the drive unit 43. By switching the aperture diameter of the aperture stop 42, the numerical aperture NA of the projection optical system PL can be switched between 0.5 and 0.8, for example.

The reticle R is held on the reticle stage 25 by vacuum suction, and the reticle stage 25 can move at a constant speed in the Y direction on a reticle base 26 provided on a column (not shown), and in the X direction, It is configured so that it can be finely moved in the Y direction and the rotation direction. The movable mirror 27 is fixed to the end of the reticle stage 25,
7 and a laser interferometer 2 arranged so as to face it
The X-coordinate, the Y-coordinate, and the rotation angle of the reticle stage 25 are constantly measured by 8, and the measurement result is supplied to the main controller 29. The operation of the reticle stage 25 is controlled.

On the other hand, the wafer W is held on a Z tilt θ stage 35 via a wafer holder 34 that performs vacuum suction, and the Z tilt θ stage 35 is provided with three fulcrums 36A to 36C which are respectively projectable and retractable in the Z direction. XY stage 3
7. In this case, the Z tilt .theta. Z tilt θ stage 35 X
The tilt angle about the axis and the tilt angle about the Y axis can be controlled, and the Z tilt θ stage 35 is configured to be rotatable about the Z axis within a predetermined range. Further, the XY stage 37 is configured to move the Z tilt θ stage 35 at a constant speed in the Y direction and to move stepwise in the X direction. Wafer holder 34, Z tilt θ
The stage 35, the fulcrums 36A to 36C, and the XY stage 37 constitute a wafer stage.

A movable mirror 39 is fixed to an end portion of the Z tilt θ stage 35, and the movable mirror 39 and a laser interferometer 40 arranged so as to face the movable mirror 39 cause the X of the Z tilt θ stage 35 (wafer W). The coordinates, the Y coordinate, and the rotation angle are constantly measured, the measurement result is supplied to the main controller 29, and the main controller 39 supplies the XY stage 37, and the XY stage 37 via the wafer stage control system 41 based on the supplied measurement value. The operation of the Z tilt θ stage 35 is controlled.

Further, an oblique incidence type multi-point focus position detection system 48 for autofocusing and autoleveling is installed on the side surface of the projection optical system PL. In this example, the thickness of the photoresist coated on the wafer W may be thinner than that of the conventional one as described later, and thus higher focus control accuracy than that of the conventional one is required. Therefore, in this example, as disclosed in, for example, Japanese Patent Laid-Open No. 6-283403,
The focus position (position in the Z direction) at a plurality of measurement points on the surface of the wafer W is measured by a method that also uses the pre-reading method. That is,
The focus position detection system 48 of this example includes a plurality of slit-shaped exposure regions 33 on the wafer W, and a plurality of read-ahead regions (for example, about 10) in the pre-reading region preceding the exposure regions 33 in the scanning direction.
Of the irradiation optical system that projects the slit image at the measurement point of, and re-images the slit images by receiving the reflected light from these slit images, and A focusing optical system for generating a focus signal and a plurality of focus signals are supplied to the focus signal processing system 49.

In the focus signal processing system 49, as an example, the focus position Z ₁ of the average surface of the exposure area 33 on the surface of the wafer W and the tilt angle θ _Y1 around the Y axis are calculated from the plurality of focus signals. Then, similarly, the average focus position Z _{2 of the} surface in the prefetch area with respect to the exposure area 33 and the tilt angle θ _Y2 around the Y axis are calculated, and the calculation result is supplied to the main controller 29. In the main controller 29, in consideration of the scanning speed of the wafer W and the response speeds of the auto focus and auto leveling mechanisms, for example, the focus of the controlled object is calculated from the weighted average of the focus positions Z ₁ and Z ₂ of the exposure area 33 and the look-ahead area. The position Z ₃ is calculated. Further, the main controller 29 calculates the tilt angle θ _Y3 around the Y axis of the control target from the weighted average of the tilt angles θ _Y1 and θ _Y2 of the exposure region 33 and the pre-read region, and at the same time, calculates the exposure region 33 and the pre-read region. The tilt angle θ _X3 around the X axis to be controlled is calculated from the difference between the focal positions Z ₁ and Z ₂ .
The focus position Z ₃ and the tilt angles θ _Y3 and θ _{X3 of} these controlled objects
The auto focus and auto leveling control are performed based on.

At the time of scanning exposure, the main controller 29 causes the KrF excimer laser light source 1 via an exposure amount control system 31 described later.
Pulse emission of the reticle stage 25 is started based on the measurement values of the laser interferometers 28 and 40.
The XY stage 37 on the wafer side is scanned in the + Y direction (or −Y direction) at the speed β · VR (β is the projection magnification) in synchronization with the scanning at the constant speed VR in the direction (or the + Y direction). Further, in the main controller 29, based on the plurality of focus signals corresponding to the exposure area 33 and the pre-reading area supplied from the focus signal processing system 49, the focus position Z ₃ of the control target, and Tilt angle θ _Y3 , θ
Sequentially calculated _X3, by driving the fulcrum 36A~36C via wafer stage drive system 41, the auto-focus system to ensure that its focal position Z ₃ is converged to the focal position of the best image plane of the projection optical system PL The focus position of the wafer W is controlled by and the tilt angle of the wafer W is controlled by the auto-leveling method so that the tilt angles θ _Y3 and θ _X3 converge to the tilt angle of the exposure area 33. As a result, the wafer W
The image of the pattern of the reticle R is sequentially transferred at a high resolution to the upper shot area to be transferred. Note that K
An exposure method in which the reticle stage 25 and the XY stage 37 on the wafer side are driven in synchronization with the pulsed light emission of the rF excimer laser light source 1 is disclosed in, for example, Japanese Patent Application Laid-Open No. 6-132191.

Further, since the present example is the step-and-scan method, the width of the exposure region 33 on the wafer W in the scanning direction is narrower than that of the collective exposure method, and the wafer W is continuously exposed during the scanning exposure. Focusing is taking place. Therefore, even if the surface of the wafer W is gently undulating in the scanning direction, the exposure region 33 of the wafer W can be accurately aligned with the best imaging plane at each time point, and the transfer on the wafer W can be performed. The average value of the deviation of the focus position at the time of exposure at each point in the target shot area, that is, the focus control accuracy is higher than that of the collective exposure method (step-and-repeat method) like stepper. However, the wafer W is
If it is known that the surface undulations are small, a sufficiently high focus control accuracy can be obtained even with a projection exposure apparatus of a batch exposure system.

Next, an exposure amount control system in the projection exposure apparatus of this example will be described. First, Z tilt θ stage 3
In the vicinity of the wafer holder 34 on the wafer 5, a dose monitor 38 made of a photoelectric conversion element is installed so that the light receiving surface is at the same height as the surface of the wafer W. When the illuminance of the illumination light IL on the surface of the wafer W is directly detected, the dose monitor 3
The light receiving surface of No. 8 is installed in the exposure area 33, and the detection signal of the irradiation amount monitor 38 is supplied to the exposure amount control system 31 which controls the integrated exposure amount for each shot area of the wafer W.

However, since the exposure amount on the wafer W cannot be directly measured during the actual exposure of the wafer W, it passes through a predetermined σ stop in the σ stop unit 15 of FIG. The exposure amount on the wafer W is indirectly measured by separating a part of the illumination light IL and detecting the light amount. Therefore, the σ diaphragm unit 15 and the first
The split prism plate 16 is arranged between the relay lens 17 and the relay lens 17. That is, the split prism plate 16 is arranged near the pupil plane of the illumination optical system.

2A is a plan view of the split prism plate 16 viewed from the fly-eye lens 12 side, FIG. 2B is a right side view of FIG. 2A, and FIG. It is sectional drawing which follows the AA line when the 2 (a) division | segmentation prism board 16 is divided into two right and left. As shown in FIG. 2A, the split prism plate 16 is made of a glass substrate that transmits the illumination light IL having a predetermined thickness, and the illumination light IL is provided at the central portion and the right end portion thereof, respectively.
Prism type beam splitters 53A and 53B for taking out a part of the are embedded. Each of the beam splitters 53A and 53B has a transmittance of 95% and a reflectance of about 5% with respect to the illumination light IL.

Then, as shown in FIG. 2B, the illumination light ILF reflected by the beam splitter 53B in the illumination light incident from the fly-eye lens 12 side at the end of the split prism plate 16 is relayed. It is incident on an integrator sensor 58B composed of a photoelectric conversion element through a lens system 57B, and an output signal of the integrator sensor 58B is shown in FIG.
Is supplied to the integrated exposure amount calculation unit 56. Also, FIG.
In FIG. 2B, the reflected light reflected by the wafer W in the illumination light IL for exposure irradiated on the wafer W via the projection optical system PL in FIG. 2B passes through the projection optical system PL and the reticle R. Returning to the beam splitter 53B, the reflected light ILR reflected by the beam splitter 53B enters the reflectance monitor 55B composed of a photoelectric conversion element via the relay lens system 54B, and the output signal of the reflectance monitor 55B is shown in FIG.
Is supplied to the image formation characteristic variation amount calculation section 59. From the output signal of the reflectance monitor 55B, the imaging characteristic variation amount calculation unit 59 can obtain the reflectance of the illumination light IL on the wafer W.

Further, as shown in FIG. 2A, at the end of the split prism plate 16, a portion 16a between the beam splitter 53B and the relay lens systems 57B and 54B has a rectangular flat plate shape, The light flux reflected by the beam splitter 53B is efficiently relay lens systems 57B and 54B.
Is to be led to. Similarly, as shown in FIG. 2C, the illumination light ILF reflected by the beam splitter 53A in the illumination light incident on the central portion of the split prism plate 16 from the fly-eye lens 12 side is the relay lens system. The light enters the integrator sensor 58A via 57A. Further, the reflected light from the wafer W of FIG. 1 returns to the beam splitter 53A of FIG. 2C via the projection optical system PL and the reticle R, and the reflected light ILR reflected by the beam splitter 53A is relay lens system 54A. It enters into the reflectance monitor 55A via. The output signals of the integrator sensor 58A and the reflectance monitor 55A are supplied to the integrated exposure amount calculation unit 56 and the imaging characteristic variation amount calculation unit 59 of FIG. 1, respectively. Since the split prism plate 16 in FIG. 1 is shown in a side view, only the relay lens systems 54B and 57B, the integrator sensor 58B, and the reflectance monitor 55B appear in FIG.

In this example, the beam splitters 53A,
Since 53B is arranged, for example, when performing illumination with a small coherence factor (σ value), the reflected light from the beam splitter 53A in the central region 60A is used, and when performing annular illumination, an annular shape is used. It is assumed that the light reflected by the beam splitter 53B at the end of the area 60B is used. By properly using the beam splitters arranged at two positions in this way, the illumination light can be accurately measured by either the illumination method with a small σ value or the annular illumination method.

Further, for example, in the case of performing modified illumination using a sigma stop composed of four apertures decentered from the optical axis, it is separately provided on the optical path of the illumination light from one of the four apertures in the split prism plate 16. The beam splitter may be arranged, and a relay lens system and a photoelectric conversion element for receiving the light beam reflected by the beam splitter may be provided.
This makes it possible to reliably detect the illumination light from the fly-eye lens 12 and the reflected light from the wafer under all illumination conditions.

Returning to FIG. 1, the split prism plate 16 of this example.
The beam splitters 53A and 53B are arranged in a part, so that it is extremely thin. On the other hand, in the past, for example, immediately after the σ stop, approximately 45 ° with respect to the optical axis.
Since one large beam splitter is arranged at an inclination angle of 1 and the reflected light of this beam splitter is detected, a large space is required to arrange this beam splitter. According to this example, since it is only necessary to dispose the thin split prism plate 16, the illumination optical system can be downsized and the illumination optical system can be easily designed.

In this case, in this example, the irradiation amount monitor 3 is previously set.
8 is compared with the output signals of the integrator sensors 58A and 58B to obtain a conversion coefficient for obtaining the exposure amount on the wafer W from the output signals of the integrator sensors 58A and 58B. It is stored in the exposure amount control system 31. When the pulsed emission of the illumination light IL is started during the scanning exposure, the integrated exposure amount calculation unit 56 sets the number of exposure pulses for each point on the wafer W to N, and outputs the output signal from the integrator sensor 58B (or 58A). Is sequentially supplied to the exposure amount control system 31. The exposure amount control system 31 multiplies the integrated signal by the conversion coefficient to obtain an integrated exposure amount at each point on the wafer W, and controls the exposure amount so that the integrated exposure amount converges to a target value. .

In this example, the photoresist layer on the wafer W may be thin, but if the resist thickness is thin in this way, the target value of the integrated exposure amount becomes smaller than in the conventional example. When the target value of the integrated exposure amount is low as described above, in the scanning exposure method, for example, first, the scanning speed of the stage system is increased to shorten the exposure time for each point on the wafer W. However, since the scanning speed of the stage system has an upper limit, it is necessary to reduce the illuminance of the illumination light IL when the scanning speed of the stage system reaches the upper limit. Thus, in order to reduce the illuminance of the illumination light IL, the exposure amount control system 31
Controls the extinction ratio in the extinction unit 2 of FIG. Further, in order to set the number of exposure pulses for each point on the wafer to a predetermined value, or to cope with output fluctuations of the KrF excimer laser light source 1 during exposure to each shot area on the wafer, dimming The unit 2 may control the dimming rate.

FIG. 3 shows a schematic structure of the dimming unit 2 of this example. In FIG. 3, two glass plates 61A and 61B are inclined at a symmetrical inclination angle δ on the optical path of the illumination light IL.
Is installed, and light shielding plates 62A and 62B for respectively shielding the reflected light from the glass plates 61A and 61B are arranged. The glass plates 61A and 61B maintain a symmetrically inclined state by a driving device (not shown), and the inclination angle δ is variable. In this case, when the inclination angle δ changes, the glass plates 61A and 61 with respect to the illumination light IL.
By utilizing the fact that the transmittance (= 1−extinction ratio) at B changes, the illumination light I can be controlled by controlling its inclination angle δ.
The desired extinction ratio for L is obtained. Since the two glass plates 61A and 61B are inclined symmetrically in this way,
Since the optical path of the illumination light IL does not laterally shift between when the light is incident on the light reduction unit 2 and when the light is emitted, it is not necessary to adjust other optical systems even when the light reduction rate is switched. Further, the light shielding plates 62A and 62B prevent the reflected light from the glass plates 61A and 61B from leaking out of the dimming unit 2.

Although the dimming unit 2 of FIG. 3 can continuously switch the dimming rate for the illumination light IL within a predetermined range, when it is desired to roughly switch the dimming rate for the illumination light IL in a plurality of steps. Before and after the dimming unit 2, for example, a rough energy adjustment unit may be arranged which is formed by disposing a plurality of ND filters having different transmittances on a disc. In this case, the dimming rate for the illumination light IL can be roughly switched by rotating the disk, and by combining this energy rough adjustment unit and the dimming unit 2 of FIG. The dimming rate can be continuously switched over a wide range.

Next, a mechanism for correcting the image forming characteristics of the projection optical system PL provided in the projection exposure apparatus of this example will be described. That is, the projection optical system PL is gradually heated by the irradiation of the illumination light IL that is pulsed, and the projection optical system PL is also heated by the light flux that is retransmitted through the projection optical system PL in the return light from the wafer W. As a result, the image forming characteristics such as the projection magnification β and the position of the best image forming surface change. Further, the imaging characteristics of the projection optical system PL also change due to changes in atmospheric pressure.
In order to correct the change in the image forming characteristic, the main controller 29 supplies information such as the illuminance of the illumination light IL and the exposure time so far to the image forming characteristic variation amount calculating section 59 in FIG. To do. Further, as described with reference to FIG. 2, the imaging characteristic variation amount calculation unit 59 has reflectance monitors 55B and 55, which are signals obtained by photoelectrically converting a part of the reflected light from the wafer W.
The output signal of A is also supplied. Further, a measurement value from an atmospheric pressure sensor (not shown) is also supplied to the image formation characteristic variation amount calculation unit 59.

The image formation characteristic variation calculation section 59 calculates the variation of the projection magnification β of the projection optical system PL based on the integrated values of the output signals of the reflectance monitors 55B and 55A and other information.
The amount of change in the field curvature, the amount of change in the focal position of the best image plane and the amount of change in the tilt angle thereof are predicted, and these prediction results are used as the main controller 29
Let us know. Further, the projection optical system PL is provided with a lens drive mechanism 44 that drives a part of the lens elements in the optical axis AX direction and inclines it with respect to a plane perpendicular to the optical axis AX. Lens drive control system 4
The operation of the lens drive mechanism 44 can be controlled via the control unit 5.

Then, when the predicted value of the change amount of the image forming characteristic is supplied from the image forming characteristic change amount calculating section 59 to the main controller 29, the main controller 29 firstly changes the projection magnification β and the image plane. In order to correct the amount of change in the curvature, the lens driving mechanism 44
A part of the lens elements of the projection optical system PL is driven via. Further, main controller 29 drives fulcrums 36A to 36C via wafer stage drive system 41 so as to match the amount of change in the focus position and tilt angle of the best imaging plane, thereby causing Z tilt θ stage 35 ( The height and tilt angle of the wafer W) are corrected. As a result, even if the imaging characteristics of the projection optical system PL change, the state in which the surface of the wafer W matches the best imaging surface of the projection optical system PL is maintained.

Next, the illumination optical system of this example will be described in detail. The illumination optical system of this example is configured to be able to switch between the normal illumination method in which the σ diaphragm has a circular opening and the annular illumination method. Therefore, in FIG. 1, as the second-stage fly-eye lens, a normal fly-eye lens 12 and a fly-eye lens 13 for annular illumination are provided in a replaceable manner. The fly-eye lens 13 for annular illumination is
It is used when the numerical aperture NA of the projection optical system PL is set to 0.6 or less and annular illumination is performed.

FIG. 4 (a) is a side view of the fly-eye lens 13 for annular illumination, and FIG. 4 (b) is a plan view of the fly-eye lens 13, as shown in FIG. 4 (b). The fly-eye lens 13 is formed by bundling lens elements having a rectangular cross-section around a metal member 63 having a cross-shaped cross-section. In this case, on the exit surface of the fly-eye lens 13, the annular σ diaphragms 75C, 7C in the σ diaphragm unit 15 of FIG.
Although any of 5D is arranged, if ILC is an annular light flux having a maximum cross-sectional area that passes through the lens element portion of the fly-eye lens 13 in FIG.
The apertures of the diaphragms 75C and 75D are configured to fit within the light flux ILC. Therefore, the fly-eye lens 13
It is possible to reduce the manufacturing cost by reducing the number of lens elements, and there is an advantage that the entire surfaces of the openings of the σ diaphragms 75C and 75D can be illuminated.

When the fly-eye lens 13 for annular illumination is used as described above, the illumination efficiency is better when the distribution of the illumination light IL on the incident surface of the fly-eye lens 13 is annular. Therefore, fly-eye lens 13 for annular illumination
When using, the illumination system control system 32 of FIG. 1 arranges the input optical system 5B for annular illumination between the deflection mirror 4 and the illuminance distribution correction plate 9 via the exchange device 8.

FIG. 5 shows the input optical system 5B,
In FIG. 5, the illumination light I from the deflection mirror 4 in FIG.
Let L be linearly polarized light. The illumination light IL is the lens 6
After it becomes a substantially parallel light flux by 4, the quarter wave plate 65
It is converted into circularly polarized light by and enters the polarization beam splitter 66, and is split into a P polarization component ILP and an S polarization component ILS in the polarization beam splitter 66. FIG.
In the input optical system 5A, a quarter-wave plate (not shown) is also arranged, and the illumination light IL incident on the fly-eye lens 7 is converted into circularly polarized light.

FIGS. 6A to 6D show the illuminance distribution of the cross section of the illumination light in each part of FIG. 5, and FIGS.
Is the optical axis A with reference to the optical axis AX of the illumination optical system.
The position r in the direction perpendicular to X is shown, and the vertical axis shows the illuminance at that position r. First, the illuminance distribution E _IN (r) of the illumination light IL entering the polarization beam splitter 66 of FIG. 5 has a normal distribution centered on the optical axis AX, as shown in FIG. 6A. The illuminance distribution E _IN (r) is also the illuminance distribution of the P-polarized component ILP and the S-polarized component ILS immediately after being separated by the polarization beam splitter 66 except for a predetermined proportional coefficient.

In FIG. 5, the polarization beam splitter 66 is shown.
The P-polarized component ILP transmitted through the
The optical member 71 having a shape obtained by processing a and 71b into a conical shape is transmitted. With this optical member 71, the P polarization component ILP
Of the P polarization component ILP immediately after being emitted from the optical member 71, because the inside and outside of the illuminance distribution are separated from each other and become parallel light again.
As shown in FIG. 6C, (r) has a distribution obtained by inverting and widening the distribution of FIG. 6A about the optical axis AX.

On the other hand, the S-polarized light component ILS reflected by the polarization beam splitter 66 of FIG. 5 is reflected by the mirror 67, the upper portion 68a of the cylinder is cut into a conical shape, and the lower portion 68b is processed into a conical shape. The light passes through the optical member 68 having the above shape. With this optical member 68, the S polarization component IL
Since the illuminance distribution of S is separated at approximately its center and becomes parallel light again, the illuminance distribution E1 (r) of the S-polarized component ILS immediately after being emitted from the optical member 68 is as shown in FIG. 6B. The distribution shown in FIG. 6A is widened around the optical axis AX.

The P-polarized light component ILP emitted from the optical member 71 is converted into S-polarized light by the ½ wavelength plate 72, reflected by the mirror 73, and incident on the polarization beam splitter 70 and reflected. . On the other hand, the S-polarized component ILS emitted from the optical member 68 is converted into P-polarized light by the ½ wavelength plate 69, and then the polarization beam splitter 7
After passing 0, the polarization beam splitter 70 outputs a light flux that is a combination of the polarization components emitted from the optical members 68 and 71. This converted illumination light IL ′ is converted into FIG.
It enters the illuminance distribution correction plate 9. The illuminance distribution E _OUT (r) of the illumination light IL ′ thus synthesized has a flat annular shape as shown in FIG. 6 (d). That is, the illuminance distribution is made uniform by the input optical system 5B of the present example as in the case of using the first-stage fly-eye lens 7 in FIG. Furthermore, in this example, since the sectional shape of the illumination light IL 'is shaped like a ring, the utilization efficiency of the illumination light is high when the ring illumination is performed.

Although not shown, a relay lens is arranged on the exit surface of the polarization beam splitter 70 shown in FIG.
With this relay lens and the relay lens 10 in FIG. 1, the exit surface of the polarization beam splitter 70 and the entrance surface of the fly-eye lens 13 for annular illumination in FIG. 4 are substantially conjugated. The cross-sectional shape of the illumination light IL ′ in FIG. 5 on the incident surface of the fly-eye lens 13 in FIG. It has a curved shape. As a result, the loss of illumination light when performing annular illumination is small.

Next, as already described, in FIG.
In this example, the aperture stop (σ stop) of the illumination system is switched by rotating the σ stop unit 15, and the σ stop is selected from the four types of σ stop 75A to 75D shown in FIG. 7A. . In FIG. 7A, the radius of the aperture of the σ stop 75A for large σ values (coherence factors) is rα, and the radius of the aperture of the σ stop 75B for small σ values is rα.
1. The outer radius and inner radius of the ring-shaped aperture of the first σ stop 75C for ring-shaped illumination are rβ and rβ2, respectively.
The outer and inner radii of the ring-shaped opening of the σ diaphragm 75D for ring-shaped illumination are defined as rβ and rβ1, respectively.

In this example, basically, when the numerical aperture NA of the projection optical system PL of FIG. 1 is in the range of 0.5 to 0.6, the σ diaphragms 75C and 75D for annular illumination are used, and the numerical aperture NA is Is in the range of 0.68 to 0.8, the σ diaphragm 75A having a circular aperture,
Use 75B. The σ value of the illumination optical system is usually set to 0.6 to 0.8, and when transferring the contact hole pattern, it is set to a small value of, for example, about 0.3 to 0.4. Therefore, if the projection magnification β of the projection optical system PL is used, the outer radius rβ of the apertures of the σ diaphragms 75C and 75D for annular illumination used when the numerical aperture NA is 0.6 or less is expressed in units of numerical aperture. Is set as follows.

Β · 0.6 × 0.6 ≦ rβ <β · 0.8 × 0.6 (11) Further, the inner radius rβ of the opening of the σ diaphragm 75D (2/3 ring zone)
1 is set to (2/3) rβ, and the σ diaphragm 75C (1/2
The inner radius rβ2 of the opening of the annular zone is set to (1/2) rβ. On the other hand, the radius rα of the σ diaphragm 75A used when the numerical aperture NA is 0.68 or more is set as follows with the numerical aperture of 0.68 regarded as approximately 0.7 and the numerical aperture as a unit.

Β · 0.6 × 0.7 ≦ rα <β · 0.8 × 0.7 (12) Further, as shown in FIG. 7B, the radius rβ of the aperture for annular illumination is The radius is set to be smaller than the radius rα of the opening for a large σ value, for example, about 0.6 / 0.7. Also, a σ stop 7 for small σ values
The radius rα1 of 5B is set to about ½ of the radius rα, and the σ stop 75B is used when the contact hole pattern is transferred.

Returning to FIG. 1, in the projection exposure apparatus of this example, when the main controller 29 sets the numerical aperture NA of the projection optical system PL to 0.5 to 0.6, the illumination system control system 32 is replaced. Fly-eye lens 13 for annular illumination via device 14
Is selected, the σ diaphragm unit 15 is driven to set either of the σ diaphragms 75C and 75D for annular illumination on the exit surface of the fly-eye lens 13. Further, the illumination system control system 32 selects the input optical system 5B via the exchange device 8.

When the main controller 29 sets the numerical aperture NA of the projection optical system PL to 0.68 to 0.8, the normal fly-eye lens 12 is controlled by the illumination system control system 32.
Is selected, one of the circular σ stops 75A and 75B is set on the exit surface of the fly-eye lens 12, and the input optical system 5A is selected. As a result, the pattern of the reticle R is transferred onto the wafer W always under the optimum σ value and without lowering the utilization efficiency of the illumination light IL.

Next, the resolution and the like obtained when the pattern of the reticle R is transferred and exposed onto the wafer W by the projection exposure apparatus of this example will be described. First, also in this example, the resolution R of the projection image by the projection optical system PL is the illumination light I for exposure.
Using the wavelength λ of L, the k factor k ₁ , and the numerical aperture NA of the projection optical system PL, it is expressed by the following equation similar to the equation (1).

R = k ₁ · λ / NA (13) When the depth of focus of the projected image of the projection optical system PL is D, the depth of focus is expressed by equation (2) using the k factor k ₂ . Furthermore, if the focus control accuracy C _F , the mark step (step of the underlying pattern) D _M , the resist thickness T _R , and the coefficient K _UD for determining the depth of focus UDOF that can be used on the operator side are used, (6 ) It is necessary to satisfy the “condition concerning the margin of the depth of focus”, which is the same as the equation below.

C _F + D _M + T _R · K _UD ≦ D = k ₂ · λ / NA ² (14) Also in this example, (a) the mark level difference D _M is reduced to 0. 1 μm (further, 0.
(B) The focus control accuracy C _{F is set} to 0.
It is about 4 μm. Further, in this example, the projection optical system PL
Even if the numerical aperture NA of is increased and the depth of focus D is reduced, the resist thickness T _{R is set} to about 0.2 μm (1 μm or more in the conventional example) so that the condition regarding the margin of the depth of focus in Expression (14) can be satisfied. To be thin. Also,
The coefficient K _UD for determining the usable depth of focus UDOF is set to 0.7, and the k factor k _{2 is set} to 1.7 in consideration of the narrow line width.

Under these conditions, when the numerical aperture NA satisfying the equation (14) is NA ₂ , the wavelength λ is 248 nm, and the numerical aperture NA ₂ is as follows. 0.4 + 0.1 + 0.2 × 0.7 ≦ 1.7 × 0.248 / NA ₂ ² , NA ₂ ≦ 0.81 (15) That is, the numerical aperture NA of the projection optical system PL is increased to 0.81. However, the condition regarding the margin of the depth of focus is satisfied. Therefore, by taking a margin and setting the numerical aperture NA to 0.78 and substituting this value into the equation (13), the resolution R becomes as follows. However, if the k factor k ₁ is 0.
55.

R = 0.55 × 0.248 / 0.78≈0.18 (μm) (16) This resolution of 0.18 μm is the resolution of a circuit pattern corresponding to a 1 Gbit DRAM. However, the resolution R is so small, the value of k factor k ₂ after optimization to become even smaller to about 1.4, by substituting k factor k ₂ to 1.4 (14) , The upper limit of the numerical aperture NA satisfying the condition regarding the depth of focus margin is 0.74.
Becomes Further, by taking into account the margin and substituting 0.7 as the numerical aperture NA of the equation (13), the resolution R is almost 0.
It becomes 19 μm. Therefore, in the projection exposure apparatus of this example, KrF excimer laser light (wavelength 248 nm) is used as the exposure illumination light IL, the thickness of the photoresist on the wafer is set to about 0.2 μm, and the projection optical system P is used.
By setting the numerical aperture NA of L to about 0.7 to 0.74, the circuit pattern of the semiconductor device equivalent to the 1 Gbit DRAM is transferred onto the wafer with high accuracy while satisfying the condition regarding the depth of focus margin. You can do it.

Actually, the resist thickness T _{R is set} to 0.5 μm.
In some cases, the upper limit of the numerical aperture NA of the projection optical system satisfying the condition regarding the depth of focus margin of the equation (14) is about 0.7 (k ₂ is 1.7).
However, the numerical aperture NA can be set to about 0.68 with an allowance. Therefore, the resist thickness T _R is 0.5 μ
When set to m or less, numerical aperture NA is approximately 0.68
Even with the above setting, the condition regarding the margin of the depth of focus can be satisfied.

Further, in this example, when the numerical aperture NA of the projection optical system PL is set within the range of 0.5 to 0.6, for example, 0.55, the equations (9) and (10) already shown. As can be seen from the equation), even when the thickness of the photoresist is about 1 μm, the circuit pattern of the semiconductor device corresponding to the 256 Mbit DRAM can be transferred while satisfying the condition regarding the margin of the depth of focus. That is, according to the projection exposure apparatus of this example, it is possible to change the numerical aperture by changing the numerical aperture of the projection optical system.
It is possible to form a circuit pattern of a semiconductor device of a next generation. In other words, according to the projection exposure apparatus of this example,
Both a critical layer with a narrow line width and a rough layer with a wide line width can be handled by switching the numerical aperture of the projection optical system.

In this regard, conventionally, when the wafer flattening step is added, the time of the entire exposure step may be doubled. Therefore, a projection exposure apparatus for performing exposure to a critical layer requiring the wafer flattening in this way. And a projection exposure apparatus that performs exposure on other rough layers. Further, since the projection exposure apparatus is selectively used for each layer of the wafer by a so-called mix-and-match method, the process control is complicated. However, in this example, since one projection exposure apparatus can perform exposure on almost all layers of one wafer, the manufacturing line can be simplified and process control becomes easy.

In the above embodiment, the exposure illumination light is used.
The KrF excimer laser light is used as
ArF excimer laser light (wavelength 193 nm) is used
You can also. Therefore, ArF excimer laser light is used
Also try to evaluate the resolution and so on.
The projection optical system PL is newly optimized for a wavelength of 193 nm.
Shall be Even in this case, wafer flattening technology
Therefore, the mark step D_MIs as small as 0.1 μm
And resist thickness T_RTo 0.2 μm and use
Coefficient K for determining the usable depth of focus UDOF_UDTo
Set to 0.7. Further improve focus control technology
Therefore, focus control accuracy C_FTo about 0.2 μm
Shall be raised. Also, considering that the line width becomes narrower
k factor k _TwoIs set to 1.5.

Under these conditions, the numerical aperture NA satisfying the condition concerning the margin of the depth of focus in the equation (14) is set to NA.
Assuming ₂ , the wavelength λ is 193 nm, so the numerical aperture N
A ₂ is as follows. 0.2 + 0.1 + 0.2 × 0.7 ≦ 1.5 × 0.193 / NA ₂ ² , NA ₂ ≦ 0.81 (17) That is, also in this case, the numerical aperture NA of the projection optical system PL is 0.
Even if it is increased to 81, the condition regarding the margin of the depth of focus is satisfied. Therefore, the numerical aperture NA is 0.73 with a margin.
By substituting this value into equation (13),
The resolution R is as follows. However, the k factor k _{1 is set} to 0.45 in consideration of the line width.

R = 0.45 × 0.193 / 0.73≈0.12 (μm) (18) This resolution of 0.12 μm is the resolution of the circuit pattern corresponding to the 4 Gbit DRAM. Therefore, in the projection exposure apparatus of this example, ArF excimer laser light (wavelength 193 nm) is used as the exposure illumination light, the thickness of the photoresist on the wafer is set to about 0.2 μm, and the projection optical system PL By setting the numerical aperture NA to about 0.73, while satisfying the conditions regarding the depth of focus margin,
A circuit pattern corresponding to 4 Gbit DRAM can be transferred onto the wafer with high accuracy.

Similarly, when using ArF excimer laser light, the numerical aperture NA of the projection optical system PL is 0.5.
Even if the thickness of the photoresist is set to about 1 μm by setting the thickness within the range of ˜0.6, the condition for the margin of the depth of focus is satisfied and the 1 Gbit DRA is set.
The circuit pattern of a semiconductor device corresponding to M can be transferred.
That is, a single projection exposure apparatus can form a circuit pattern of a second-generation semiconductor device.

Furthermore, in the projection exposure apparatus of this example, the i-line (wavelength 365 nm) of the mercury lamp is used as the illumination light for exposure.
Is used, the numerical aperture NA of the projection optical system PL is set to about 0.7 to 0.8 and the resist thickness is 0.2 μm.
By setting the degree to about, it is possible to transfer the circuit pattern corresponding to the 256 Mbit DRAM with high accuracy while satisfying the condition regarding the depth of focus margin. On the other hand, as already described, when the numerical aperture NA is set to about 0.5 to 0.6, the circuit pattern corresponding to the 64 Mbit DRAM can be transferred with high accuracy, and as a result, the circuit of the second-generation semiconductor device is obtained. A pattern can be formed.

As the exposure illumination light, a metal vapor laser light, a harmonic of a YAG laser, or another bright line may be used. In these cases, too, the photoresist is thinned and the aperture of the projection optical system is used. By increasing the number, it is possible to satisfy the condition regarding the margin of the depth of focus and improve the resolution. Furthermore, if the mark step D _M can be reduced and the resist thickness T _R can be made thinner by the progress of the flattening technology in the future, even if the numerical aperture NA of the projection optical system is made larger than 0.7, the formula (14) is obtained. Since the condition regarding the margin of the depth of focus is satisfied, the resolution R can be further increased.

Next, in this example, as one method for satisfying the condition regarding the margin of the depth of focus in the equation (14),
The wafer planarization technique is used as described above. When the wafer is flattened in this manner, the mark step D _M, which is the step of the circuit pattern formed on the wafer W, is reduced to, for example, about 50 nm or less, and the mark step D _M is attached to each shot area on the wafer W. The step of the alignment wafer mark is also reduced. Therefore, in this example, a sensor capable of detecting a low step mark is provided as an alignment sensor for performing alignment of each of the shot areas.

In FIG. 1, an alignment sensor 50 is arranged on the side surface of the projection optical system PL. In addition,
This alignment sensor 50 is disclosed in Japanese Patent Laid-Open No. 5-21783.
The configuration is similar to that of the alignment sensor disclosed in Japanese Patent No. 5 publication. FIG. 8 is a perspective view schematically showing the overall configuration of the alignment sensor 50. In FIG. 8, reflected light from a lattice-shaped wafer mark (not shown) formed on the wafer is reflected by the mirror M1. After being reflected and incident on the objective lens OBL, it is split by the beam splitter BS1. The light flux reflected by the beam splitter BS1 is
The imaging method, that is, the FIA (Field Image), which is reflected by the mirror M2 and is arranged on the upper stage of the alignment sensor 50.
The light enters the detection system (hereinafter referred to as “FIA system”) 51A of the Alignment method. On the other hand, the beam splitter BS1
The light flux that has passed through is a two-beam interference method arranged in the lower part, that is, LIA (Laser Interferometric Alignment)
The light enters a detection system (hereinafter, referred to as “LIA system”) 51B of the method.

FIG. 9 shows the structure of the FIA system 51A. In FIG. 9, the non-photosensitive broadband (band 270 nm or more) illumination light ALA is emitted from the optical fiber 116A to the photoresist on the wafer. The illumination light ALA illuminates the wafer illumination field diaphragm 121A with a uniform illuminance via the condenser lens 120A. The illumination light limited by the field stop 121A is reflected by the mirror 122A, passes through the lens system 123A, and is reflected by the beam splitter B.
It is incident on S2. The illumination light ALA reflected by the beam splitter BS2 is reflected by the mirror M2 and enters the beam splitter BS1. After that, the illumination light ALA is on the optical axis A.
The wafer mark on the wafer is illuminated along the Xa through the objective lens OBL and the mirror M1. In the illumination light transmission path for the wafer, the field stop plate 121A includes the lens system 123.
The illumination area for the wafer by the FIA system 51A is uniquely determined by the aperture formed in the field diaphragm 121A, which is conjugate with the wafer with respect to the combined system of A and the objective lens OBL.

Then, reflected light is generated from the wafer illuminated by the illumination light ALA from the optical fiber 116A, and this reflected light is passed through the mirror M1, the objective lens OBL, the beam splitter BS1, and the mirror M2 to the beam splitter BS2. About 1/2 of the reflected light is incident on the detection optical system through the beam splitter BS2. This detection optical system includes a lens system 124, a mirror 125, an index plate 126 on which index marks are formed, a first relay lens system 127 for imaging, a mirror 128, an interference filter 129, a second relay lens system 131, and a beam splitter. BS3
It is composed of And the beam splitter BS3
The return light from the wafer divided by is incident on the image pickup device 108X for detecting the X direction and the image pickup device 108Y for detecting the Y direction, each of which includes a two-dimensional CCD.
An image of the wafer mark and an image of the index mark on the wafer surface are formed on the image pickup surface of 108Y.

That is, the objective lens OBL and the lens system 12
For the composite system of No. 4, the wafer surface and the index plate 126 are conjugate, and for the relay lens systems 127 and 131, the index plate 126 and their imaging surfaces are conjugate. Further, a variable power optical system 130 can be inserted in the vicinity of the interference filter 129 so that the magnification of the image of the wafer mark and the index plate can be changed. In addition, the interference filter 1
Reference numeral 29 plays a role of blocking strong laser light used in the LIA system 51B described later.

Further, in FIG. 9, an illumination system for independently illuminating the index plate 126 is provided. This illumination system includes an optical fiber 116B and a condenser lens 120.
B, an illumination field diaphragm plate 121B, a mirror 122B, and a lens system 123B. Then, the illumination light ALB emitted from the optical fiber 116B passes through the condenser lens 120B to the lens system 123B, and from the side opposite to the optical path of the illumination light ALA for the wafer, the beam splitter BS.
2 and is reflected by the beam splitter BS2 to illuminate the index plate 126.

Further, the image pickup signals of the image pickup devices 108X and 108Y are supplied to the alignment signal processing system 52 of FIG.
The alignment signal processing system 52 processes the image pickup signals to detect the position of the wafer mark to be detected, and supplies the detection result to the main controller 29.
Main controller 29 aligns wafer W based on the position of each wafer mark.

Next, FIG. 10 shows the configuration of the LIA system 51B. In FIG. 10, the He--Ne laser light source 106 is shown.
The laser beam from is split into two by the beam splitter BS8 via the shutter 140. Beam splitter BS8
The laser beam LB that has passed through is incident on the heterodyne two-beam conversion unit 141X via a plurality of folding mirrors in the LIA system in the X direction. In this heterodyne two-beam conversion unit 141X, a pair of frequency shifters (acousto-optical elements) having different driving frequencies and two frequency-modulated laser beams emitted from the pair of frequency shifters are used as optical axes. And a synthesizing system for synthesizing in parallel with. The two laser beams LB1x and LB2x emitted from the synthesizing system enter the lens system 142X and uniformly illuminate the aperture plate 143X arranged on the rear focal plane of the lens system 142X. Therefore, the two laser beams LB1x and LB are arranged on the aperture plate 143X.
The 2x intersection forms a one-dimensional interference fringe. Moreover, since the drive frequencies of the pair of frequency shifters in the heterodyne two-beam conversion unit 141X are different from each other, the one-dimensional interference fringes flow in the pitch direction at a speed corresponding to the frequency difference.

Now, the two laser beams limited by the aperture plate 143X are beam splitter BS6.
Is partially reflected by, passes through the lens system 144X, the beam splitters BS4 and BS1, and then reaches the wafer through the objective lens OBL and the mirror M1. Since the two laser beams LB1x and LB2x are symmetrically inclined with respect to the X direction, when they are irradiated on the lattice-shaped X-axis wafer mark provided on the wafer, the laser beams LB1x and LB2x are ± 1 perpendicular to the wafer mark.
Next-order diffracted light is generated. The two ± 1st-order diffracted lights have the same polarization direction and thus interfere with each other, and at the same time, the interference intensity periodically changes at the frequency difference between the pair of frequency shifters, that is, the beat frequency.

The interference beat light is incident on the beam splitter BS6 via the mirror M1, the objective lens OBL, the beam splitters BS1 and BS4, and the lens system 144X, and the interference beat light transmitted through the beam splitter BS6 is received. Reach plate 145X. The light receiving aperture plate 145X includes a lens system 144X and an objective lens O.
It is arranged at a position conjugate with the wafer with respect to the combined system with BL, and has an opening through which only the interference beat light passes. The interference beat light that has passed through the aperture plate 145X is
It reaches the photoelectric sensor 148X via the mirror 146X and the lens system 147X, and the beat signal is output from the photoelectric sensor 148X.

The laser beam reflected by the beam splitter BS8 enters the LIA system for the Y direction via the mirror. This LIA system for Y direction is L for X direction.
Heterodyne 2-beam conversion unit 1 in almost symmetry with the IA system
41Y, lens system 142Y, aperture plate 143Y, beam splitter BS5, lens system 144Y, light receiving aperture plate 145Y, mirror 146Y, lens system 147.
Y and photoelectric sensor 148Y. However, the heterodyne 2-beam conversion unit 141Y
The two laser beams LB1y and LB2y from the heterodyne two-beam forming unit 141Y are arranged so as to be rotated by 90 ° with respect to the beam forming unit 141X so that they intersect the Y-axis wafer mark on the wafer in the Y direction. Is irradiated. Then, the beat signal corresponding to the wafer mark on the Y-axis is output from the photoelectric sensor 148Y.

Further, in order to obtain the reference signal of the beat signal from the photoelectric sensors 148X and 148Y, the laser beam L
The laser beam transmitted through the beam splitter BS4 among B1x and LB2x, and the laser beams LB1y and LB2
The laser beam reflected by the beam splitter BS4 in y is incident on the beam splitter BS7 via the lens system 150 and the mirror 151. Then, the laser beam reflected by the beam splitter BS7 enters the transmission type diffraction grating plate 153X via the mirror 152.
The transmission type diffraction grating plate 153X forms a diffraction grating at a position conjugate with the opening of the X-axis light receiving aperture plate 145X,
Other than that is used as a light-shielding portion, and only the laser beams LB1x and LB2x are incident on the diffraction grating, and interference beat light composed of ± first-order diffracted light generated from the diffraction grating is a lens system 154X for Fourier transform. The incident light enters the photoelectric sensor 155X via the optical sensor 155X, and the reference beat signal for the X axis is output from the photoelectric sensor 155X.

The laser beam transmitted through the beam splitter BS7 is incident on the transmission type diffraction grating plate 153Y. The transmission type diffraction grating plate 153Y forms a diffraction grating at a position conjugate with the opening of the Y-axis aperture plate 145Y for light reception, and forms the other part as a light shielding part. The laser beam LB1y is applied to the diffraction grating. , LB2y is incident, and the interference beat light composed of ± first-order diffracted light generated from the diffraction grating passes through the lens system 154Y and the photoelectric sensor 15
5Y, and the photoelectric sensor 155Y outputs a reference beat signal for the Y axis.

The beat signals from the photoelectric sensors 148X and 148Y and the reference beat signals from the photoelectric sensors 155X and 155Y are respectively the alignment signal processing system 5 of FIG.
2, the alignment signal processing system 52 detects the position of the X-axis and Y-axis wafer marks to be detected from the beat signal and the reference beat signal, and supplies the detection result to the main controller 29. Main controller 29 aligns wafer W based on the position of each wafer mark.

As described above, the alignment sensor 5 of this example is used.
0 includes a FIA system 51A and a LIA system 51B. In this case, the FIA system 51A can detect a wafer mark having a step of 100 nm or more with a high SN ratio, and since it is an image processing method, it has an asymmetric wafer mark or a wafer mark having a rough surface. The position detection can be performed with high accuracy. In contrast, the LIA system 51B can detect a low step mark (50 nm or less) with high accuracy. Therefore, by using the FIA system 51A for the high step mark and the LIA system 51B for the low step mark, it is possible to deal with any semiconductor device.

Next, in this example, the thickness of the photoresist on the wafer W (resist thickness) may be made thinner than in the conventional example. However, when the photoresist becomes thinner, the target value of the integrated exposure amount for that photoresist is set. Decreases in proportion to its thickness. Specifically, conventionally, the resist thickness is about 1 μm or more, and the variation in the resist thickness is 0.01 to
Since the thickness is 0.02 μm (10 to 20 nm), in the conventional example, the target value of the integrated exposure amount remained substantially constant even if the resist thickness varied. On the other hand, in this example, the resist thickness may be about 0.2 μm as described above, and in this case, it is necessary to correct the target value of the integrated exposure amount according to the variation in the resist thickness. For that purpose, since it is necessary to actually measure the resist thickness on the wafer W, in this example, a resist thickness measuring device 46 is provided on the side surface of the projection optical system PL as shown in FIG. The structure and operation of the resist thickness measuring device 46 are shown in FIGS.
This will be described with reference to FIG.

First, FIG. 11 is a block diagram showing a resist thickness measuring device 46. In this FIG. 11, the illumination light emitted from a light source 76 such as a lamp is reflected by an elliptic mirror 77 to be a wavelength selection filter plate. The measurement light L having a specific single wavelength which is incident on 78 and is non-photosensitive to the photoresist PR coated on the wafer W by the wavelength selection filter plate 78.
R is selected. Further, by rotating the wavelength selection filter plate 78 via the driving device 79, the measurement light L
The R wavelength can be switched to a plurality of types. The operation of the drive device 79 is controlled by the external film thickness measurement calculation device 47 via the control system 80.

The measurement light LR that has passed through the wavelength selection filter plate 78 is converted into a parallel light flux by the lens 81 and enters the half prism 82, and the measurement light LR reflected by the half prism 82 is passed through the objective lens 83 to the wafer. Thickness d on W
A spot image is formed in the photoresist PR. Measurement light L emitted by the objective lens 83 having a constant numerical aperture
R is reflected by the surface of the photoresist PR and the surface of the wafer W, and the return light LR ′ thereof is incident on the image pickup surface 84a of the image pickup device 84 including a two-dimensional CCD via the objective lens 83 and the half prism 82. To do. Image sensor 8
The image pickup surface 84a of No. 4 is the wafer W with respect to the objective lens 83.
Is the optical Fourier transform surface (pupil surface) of the surface of
Interference fringes due to the return light LR 'are formed on the imaging surface 84a. The image pickup signal of the image pickup device 84 is used as the film thickness measurement calculation device 47.
Is supplied to the photoresist PR from the information of the interference fringes obtained by processing the image pickup signal in the film thickness measurement calculation device 47.
Thickness d is calculated and the calculation result is supplied to main controller 29 in FIG.

FIG. 12 is an explanatory view of the measurement principle of the thickness d of the photoresist PR. In FIG.
(A) shows reflection of the measurement lights LR1 and LR2 on the respective surfaces of the wafer W and the photoresist PR, and (b) shows an imaging surface 8 of the return lights LR1 ′ and LR2 ′ corresponding to the measurement lights.
The state of incidence on 4a is shown. In this case, the measurement light LR in FIG. 11 is a light beam having a predetermined numerical aperture, and the measurement light LR1 in FIG. 12A is the measurement light that is vertically incident on the surface of the wafer W. LR2 represents the measurement light incident on the surface at the incident angle θ. In addition, the return light LR2 ′ reflected by the wafer W or the photoresist PR and traveling in the same direction is reflected by the objective lens 83 on the imaging surface 84a at a predetermined centered on the optical axis, as shown in FIG. It is focused on the circumference.

In FIG. 12A, the photoresist P
When the refractive index of R is n, the refraction angle θ ′ of the measurement light LR2 in the photoresist is as follows. sin θ ′ = sin θ / n (19) In addition, the case where the measurement light LR2 is directly reflected by the surface A ′ of the photoresist PR, the case where the measurement light LR2 is directly reflected by the surface A of the photoresist PR,
Surface B of wafer W and surface A ′ of photoresist PR
The optical path difference Δ with respect to the case of being reflected through the optical path connecting the points is as follows. Note that, for example, the length of the optical path between the front surface A and the front surface B is represented by AB, and the intersection point of the perpendicular line drawn from the front surface A to the measurement light LR2 toward the front surface A ′ and the measurement light LR2 is P. .

Δ = n (AB + BA ′) − A′P = 2nd / cos θ′−2dtan θ′sin θ = 2nd / cos θ′−2ndtan θ′sin θ ′ = 2ndcos θ (20) Further, refraction of air From the index, the refractive index n of the photoresist PR
Is larger, the optical path connecting the measurement light LR2 directly reflected by the surface A ′ and the surface A, the surface B, and the surface A ′ in consideration of the phase change π when the photoresist PR reflects on the surface A. Phase difference δ with the measurement light LR2 reflected through
Is as follows. The wavelength of the measurement light LR is λ _LR .

Δ = (2π / λ _LR ) Δ + π (21) Substituting the equation (20) into this gives the following. δ = (2π / λ _LR ) 2ndcos θ ′ + π (22) The phase difference δ determines the brightness of the return light (interference light) LR2 ′ corresponding to the measurement light LR2. That is, an integer N of 1 or more
When the phase difference δ is (2N + 1) π using, the return light LR2 ′ becomes a dark interference fringe. That is, in the Nth dark interference fringe, the following expression holds for the refraction angle θ ′ of the measurement light LR2.

(2N + 1) π = (2π / λ _LR ) 2ndcos θ ′ + π (23) Substituting the equation (19) into this gives the following. θ ² = (n / 2d) (4nd−2Nλ _LR ) (24) Here, the radius from the optical axis at the center of the Nth dark interference fringe on the image pickup surface 84a of the image pickup device 84 is r _N , If the focal length of the objective lens 83 is f, then θ = r _N / f, and therefore the equation (24) is as follows.

R _N ² = (nf ² / d) (2nd−Nλ _LR ) (25) When this equation is modified, the photoresist thickness is calculated based on the radius r _N of the _Nth dark interference fringe as follows. d is required. ^{_{r N 2 d = 2n 2 f}} 2 d-Nλ LR, d = (- nf 2 Nλ LR) / (r N 2 -2n 2 f 2) (26) FIG. 13 (b), the wafer W and the photoresist PR 13B shows the appearance of interference fringes on the image pickup surface 84a of the image pickup element 84 due to the return light from the image pickup surface 84a.
Dark interference fringes 85 are formed concentrically on a.
In this example, a rectangular detection area 84b centered on the optical axis is set on the image pickup surface 84a, and image pickup signals corresponding to the image intensity in the detection area 84b are sequentially added in the short side direction.
In this case, the detection area 84b passes through the optical axis on the imaging surface 84a.
The axis along a straight line parallel to the longitudinal direction of is the y-axis, the origin of the y-axis is on the optical axis, and the image pickup signal ST thus added is represented as a function of the y-axis.

FIG. 13A shows the image pickup thus added.
FIG. 13A shows the signal ST, and in FIG.
Position y on the y-axis where ST is depressed₀, Y₁, Y₁',
y_Two, Y_Two', ... correspond to the positions of the dark interference fringes described above.
So, from the position of this dark interference fringe,
..., N'th (N 'is the most observed in the detection region 84b)
Radius r of dark fringes up to (indicating the order of distant fringes)₁,
r_Two, ..., r_N′ Is calculated, and these radii are converted into the equation (26).
Substituting the thickness d of the photoresist₁, d _Two, ...,
d_N'And calculate the average value of these N'thicknesses at the end.
Let the final thickness d. This gives an averaging effect
And to obtain the thickness d of the photoresist PR with high accuracy.
Can be.

Returning to FIG. 11, in this example, since the photoresist PR to be measured is thin, it is necessary to select, as the measurement light LR, light of a short wavelength that the photoresist PR is not exposed to. When KrF excimer laser light with a wavelength of 248 nm is used as the illumination light IL for exposure as in this example, the measurement light LR is, for example, a wavelength of 365 nm from a mercury lamp.
It is conceivable to use the i-line or the like. However, the measurement light L
As R, He-Ne laser light (wavelength 633 nm) or the like may be used. In addition, it is necessary to increase the numerical aperture of the objective lens 83. Furthermore, although dark interference fringes are detected in this example, when the photoresist PR is thick, it may be necessary to detect bright interference fringes.

Next, an example of the operation of controlling the exposure amount by measuring the resist thickness on the wafer using the resist thickness measuring device 46 of this example will be described. The thickness unevenness (unevenness of coating) of the photoresist on the wafer varies depending on the management status of the photoresist, and will be described in different cases as follows. First, FIG. 14 shows a case where there is uneven thickness (uneven coating) of the photoresist PR on one wafer W. Particularly, when the photoresist PR is thin, this FIG.
As shown in FIG. 4, the ratio of the thickness unevenness to the average value of the thickness of the photoresist PR becomes large, and it becomes necessary to partially change the target value of the integrated exposure amount. Therefore, when the photoresist is thinly applied on one wafer W and there is loose thickness unevenness, the resist thickness measuring device 46 is used to measure the shot area on each wafer W. It is necessary to measure the resist thickness.

Specifically, for example, as shown in FIG. 15B, the measurement point AF2 at the center of the shot area SA to be exposed.
When measuring the focus position at the position using the focus position detection system 48 of FIG. 1, the resist thickness measuring device 46 is used to measure the resist thickness at the film thickness measuring point RD2 near the measuring point AF2. It may be supplied to the main controller 29. In this case, the main controller 29 of FIG.
A target value E ₀ of the integrated exposure amount at ₀ is obtained in advance, and a reference value d ₀ of the resist thickness d _i measured in the i-th shot area is obtained.
Letting Δd _i be the difference from, the target value E _i of the integrated exposure amount for the i-th shot area can be obtained from the following equation.

E _i = E ₀ + E ₀ (Δd _i / d ₀ ) (27) The target value E _i of the integrated exposure amount is supplied to the exposure amount control system 31, and the target value E _{i is} set in the exposure amount control system 31. The light reduction rate or the like in the light reduction unit 2 is controlled in accordance with the above. Next, for example, when there is a difference in resist thickness between manufacturing lots, the resist thickness is measured at a plurality of points on the first wafer of the manufacturing lot, and the measurement results are averaged to obtain an average of the wafers of the lot. Then, the resist thickness d is calculated. Next, the target value of the integrated exposure amount for the wafers of the lot may be determined according to the difference between the resist thickness d thus obtained and the reference resist thickness.

When there is a difference in resist thickness between wafers in the same lot, the resist thickness may be measured during search alignment for roughly aligning the wafers, for example. At the time of the search alignment, the focus position detection system 48 of FIG. 1 is used to measure the focus position at a predetermined measurement point on the wafer, and the focus position at the measurement point is adjusted to the image plane by the autofocus method. ing.

FIG. 15A shows an example of resist thickness measurement points when such search alignment is performed.
In FIG. 15A, the measurement point AF1 of the focus position for autofocus is set in the center of the wafer W,
Measurement points ALG for search alignment at two points on the surface of the
1 and ALG2 are set, and search alignment marks are formed at these measurement points ALG1 and ALG2. For example, the position of the mark for search alignment can be detected through the FIA system 51A of FIG. Further, at three film thickness measurement points RD1 to RD3 surrounding the center of the wafer W, the resist thickness is measured by using the resist thickness measuring device 46, and the average value of the measured values is set as the resist thickness of the wafer W. The exposure amount is controlled based on the resist thickness.

When the variation in the resist thickness is large, it is desirable to control the exposure amount according to the resist thickness even within one shot area on the wafer. It is difficult to partially control the exposure amount in one shot area in the batch exposure method, but it is 1 in the scanning exposure method like this example.
By pre-reading the distribution of resist thickness within one shot area, it becomes possible to perform exposure while correcting the exposure amount. For that purpose, in FIG. 1, a plurality (for example, three) of resist thickness measuring devices that are the same as the resist thickness measuring device 46 may be provided, and the resist thickness may be pre-read at a plurality of rows of measurement points in each shot area.

FIG. 15C shows an example of film thickness measurement points in the case of pre-reading the resist thickness during such scanning exposure. On the other hand, the shot area S on the wafer
A is scanned in the + Y direction, for example. Then, the exposure area 3
The focus position is measured by the focus position detection system 48 of FIG. 1 at a plurality of measurement points including the measurement point AF3 at the center of the three measurement points, and three film thickness measurement points R preceding the exposure area 33 are measured.
At D5, RD6, and RD7, the resist thickness measuring device similar to the resist thickness measuring device 46 of FIG. 1 measures the resist thickness at a predetermined sampling cycle. In this case, as an example, three film thickness measurement points RD5 to RD7
The average value of the resist thickness measured in
The exposure amount is obtained as a function of the position of A in the Y direction, and the exposure amount is controlled by the exposure amount control system 31 of FIG. 1 according to the resist thickness at the position Y as the shot area SA moves in the Y direction.

Further, in order to deal with the variation of the resist thickness within the shot area SA accurately, in addition to controlling the exposure amount according to the position of the shot area SA in the Y direction, the film thickness at three locations is also controlled. The illuminance distribution in the non-scanning direction (X direction) of the illumination light IL applied to the reticle R in FIG. 1 may be controlled based on the distribution of the resist thickness at the measurement points RD5 to RD7. That is, three film thickness measurement points RD5, RD6,
In RD7, for example, when the resist thickness at the film thickness measuring point RD5 at the end is thicker than the other measuring points (the required exposure amount is large), the illumination around the film thickness measuring point RD5 is performed. It is desirable to increase the illuminance of light. When the illuminance distribution of the illumination light IL in the non-scanning direction is controlled to some extent in this way, the illuminance distribution correction plate 9 of FIG. 1 is used. Since this illuminance distribution correction plate 9 is arranged in the vicinity of the pupil plane with respect to the pattern forming surface of the reticle R, the illuminance distribution correction plate 9
By controlling the illuminance distribution in accordance with the inclination angle of the illumination light in the non-scanning direction at the stage, it is possible to control the illuminance distribution in the non-scanning direction on the reticle R (and thus on the wafer W).

FIG. 16A shows the illuminance distribution correction plate 9 placed in the optical path of the illumination light emitted from the fly-eye lens 7 (or the input optical system 5B) shown in FIG.
Assuming that the vertical direction (x direction) along the paper surface of 6 (a) corresponds to the non-scanning direction (X direction) of the wafer, the illuminance distribution correction plate 9 made of parallel flat glass is shown in FIG.
It is rotatably supported about an axis perpendicular to the plane of FIG. Then, the rotation angle Φ around the axis is set to 0 (when parallel to the x direction) through a rotation device (not shown).
The illumination system control system 32 of FIG.

In FIG. 16A, the optical axis of the illumination optical system
Illumination light along₀With respect to the x direction
φ₀And φ₁(φ₁> Φ₀) Illumination light IL inclined only₁as well as
IL _TwoIs assumed. Also, the angle φ₁Is equal to the rotation angle Φ
I.e., illumination light IL_TwoIs vertical to the illuminance distribution correction plate 9
It is assumed to be incident on. Generally parallel flat glass
Has a smaller transmittance as the angle of incidence increases, so
Meiko IL₀~ IL_TwoBy the illuminance distribution correction plate 9 for
The transmittance changes.

FIG. 16B shows the transmittance of the illumination light by the illuminance distribution correction plate 9 with respect to the illumination light. In FIG. 16B, the horizontal axis indicates the non-illuminance with respect to the optical axis of the illumination light incident on the illumination distribution correction plate 9. The inclination angle φ in the scanning direction (x direction) is represented, and the vertical axis represents the transmittance T (%) of the illumination light having the inclination angle φ. In FIG. 16B, the tilt angles φ are 0, φ ₀ ,
The illumination light of φ ₁ is the illumination light I of FIG.
It corresponds to L ₀ , IL ₁ and IL ₂ . In this case, the illumination light I
Since L ₂ is vertically incident on the illuminance distribution correction plate 9, the transmittance T for the illumination light IL ₂ is the highest, and the illuminance of the illumination light IL ₂ after passing through the illuminance distribution correction plate 9 is the highest. growing. Therefore, by rotating the illuminance distribution correction plate 9 by the tilt angle in the direction in which the illuminance of the illumination light is desired to be maximized in the non-scanning direction, the illuminance of the illumination light in that direction can be maximized.

Since the wafer is scanned in the Y direction with respect to the exposure area, variations in the illuminance distribution of the illumination light in the scanning direction (Y direction) do not pose a problem. However, if necessary, the illuminance distribution correction plate 9 may be tilted in the direction corresponding to the Y direction to control the illuminance distribution in the scanning direction on the wafer. As described above, in this example, the thickness of the photoresist on the wafer or the distribution thereof is actually measured, and the exposure amount for the wafer is controlled according to the thickness. However, it is possible to always give a proper integrated exposure amount to the photoresist. The resist thickness on the wafer is measured during search alignment measurement, autofocus measurement, or scanning exposure, and there is no special time for measuring the resist thickness. Throughput (productivity) of the exposure process due to measurement does not decrease.

In the above example, the resist thickness measuring device 46 for measuring the resist thickness is arranged in the projection exposure apparatus. However, for example, the wafer is transferred from the resist coater to the projection exposure apparatus on the way. The resist thickness on the wafer is measured by the arranged resist thickness measuring device, and the data is stored in the main controller 29 in the projection exposure apparatus as a data file online or by manual input, and the integrated exposure amount is controlled based on the information. It is also possible to do so.

In the above embodiment, the illuminance of the illumination light for exposure (integrated exposure amount) is optically corrected using the dimming unit 2. However, the number of exposure pulses for each point on the wafer is Alternatively, the pulse energy of the KrF excimer laser light source 1 as the light source for exposure may be changed and corrected.
Alternatively, the integrated exposure amount may be controlled by driving the fixed field diaphragm 18 or the movable field diaphragm 20 to control the width of the slit-shaped illumination area 33 on the reticle R in the scanning direction. Further, the integrated exposure amount can be similarly controlled by finely adjusting the scanning speed.

Further, it goes without saying that the present invention can be applied not only to the case of performing the exposure by the step-and-scan type projection exposure apparatus but also to the case of performing the exposure by the collective exposure method such as the stepper. That is, even in the collective exposure method, the condition regarding the margin of the depth of focus can be satisfied even if the numerical aperture of the projection optical system is increased by reducing the resist thickness, and the resolution can be improved.
An appropriate integrated exposure amount can be obtained by measuring the resist thickness.

As described above, the present invention is not limited to the above-mentioned embodiments, and various configurations can be adopted without departing from the gist of the present invention.

[0125]

According to the first exposure method of the present invention, since the integrated exposure amount is controlled according to the thickness of the photosensitive material coated on the substrate, when the photosensitive material is thinned, the substrate Even if there is variation in the thickness of the photosensitive material among the substrates, between substrates, or between lots, it is possible to give an appropriate integrated exposure amount to the photosensitive material based on the measured thickness. Further, when the projection optical system is used and the photosensitive material is made thin, the condition regarding the margin of the depth of focus is satisfied even if the numerical aperture of the projection optical system is increased. Therefore, by increasing the numerical aperture of the projection optical system and thinning the photosensitive material, there is an advantage that the resolution of the transferred pattern can be increased without shortening the wavelength of the illumination light for exposure. Moreover, even when the thickness of the photosensitive material varies, the present invention can provide an appropriate integrated exposure amount even if the thickness of the photosensitive material varies.

Further, an appropriate integrated exposure amount for the photosensitive material having a predetermined reference thickness corresponding to the mask is stored in advance, and the thickness of the photosensitive material is calculated from the predetermined reference thickness. When the integrated exposure amount of the exposure illumination light is set based on this difference, there is an advantage that an appropriate integrated exposure amount can be easily calculated even if the thickness of the photosensitive material varies. .

When the transfer pattern on the mask is projected and exposed on the substrate through the projection optical system, the thickness of the photosensitive material coated on the substrate is set to 0.5 μm or less and the projection optical system is set. If the numerical aperture is set to 0.68 or more, the circuit pattern of the semiconductor device of almost one generation can be transferred with high accuracy without shortening the wavelength of the exposure illumination light. That is, a pattern corresponding to the next-generation 256 Mbit DRAM can be formed using the i-line of a mercury lamp as the exposure illumination light, and a pattern corresponding to the 1 Gbit DRAM can be formed using the KrF excimer laser light.
By using the rF excimer laser light, it is possible to form a pattern corresponding to a 4 Gbit DRAM.

Next, according to the second exposure method of the present invention,
Since the switching of the numerical aperture of the projection optical system is controlled according to the thickness of the photosensitive material coated on the substrate, when the photosensitive material is thin, the numerical aperture can be increased to easily provide a margin for the depth of focus. Meet the conditions for
The resolution can be increased. In addition, when the photosensitive material is thick, the numerical aperture is made small and transfer is performed at a low resolution, so that there is an advantage that patterns having different resolutions can be dealt with. Specifically, when the thickness of the photosensitive material is 0.2 μm or less, the numerical aperture of the shadow optical system is 0.7 or more, and when the thickness of the photosensitive material coated on the substrate is 1.0 μm or more, the shadow optical system is By switching the numerical aperture of the system to 0.6 or less, it is possible to deal with the second generation semiconductor device. That is, since it becomes possible to form a pattern corresponding to a second-generation semiconductor device with one exposure apparatus without changing the wavelength of the exposure illumination light, the process control for the exposure apparatus, which was conventionally performed on another line, can be performed. Will be enough for one car. Therefore, the process control is easy, the process control time is shortened, and the throughput of the exposure process is improved.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram showing a projection exposure apparatus used in an example of an embodiment of an exposure method according to the present invention.

2A is a plan view showing a configuration of a split prism plate 16 and a detection system in FIG. 1, FIG. 2B is a right side view of FIG. 2A, and FIG. 2C is a view of FIG. 2A. It is sectional drawing which follows the AA line.

3 is a schematic configuration diagram showing a configuration of a dimming unit 2 in FIG.

4A is a side view showing a fly-eye lens 13 for annular illumination in FIG. 1, and FIG. 4B is a fly-eye lens 1 thereof.
FIG.

5 is an optical path diagram showing a configuration of an input optical system 5B for annular illumination in FIG.

FIG. 6 is a diagram showing a light amount distribution of illumination light at each point of the input optical system 5B of FIG.

FIG. 7 is a diagram showing a configuration of a σ diaphragm unit 15 in FIG. 1.

8 is a perspective view showing a schematic configuration of an alignment sensor 50 in FIG.

9 is a perspective view showing the FIA system 51A of FIG. 8. FIG.

10 is a perspective view showing the LIA system 51B of FIG. 8. FIG.

11 is a schematic configuration diagram showing a resist thickness measuring device 46 in FIG.

12A is a partially enlarged view showing an optical path of measurement light irradiated on the wafer and the photoresist from the resist thickness measuring device 46 of FIG. 11, and FIG. FIG. 6 is an optical path diagram showing how interference fringes are formed on the image pickup surface 84a of FIG.

FIG. 13 is a diagram showing interference fringes formed on an image pickup surface and an image pickup signal corresponding to the interference fringes.

14A is a plan view showing a wafer coated with a photoresist, and FIG. 14B is a side view of FIG. 14A.

15A is a plan view showing measurement points RD1 to RD3 of resist thickness on one wafer, FIG. 15B is a plan view showing measurement points RD4 of resist thickness in one shot area SA, FIG. (C) is a plan view showing measurement points RD5 to RD7 when the resist thickness in the shot area SA is pre-read.

16A is a plan view showing an illuminance distribution correction plate 9 in FIG. 1, and FIG. 16B is a diagram showing a transmittance T with respect to an inclination angle of illumination light incident on the illuminance distribution correction plate 9.

FIG. 17 is a schematic configuration diagram showing a conventional projection exposure apparatus.

[Explanation of symbols]

 R reticle PL projection optical system W wafer 1 KrF excimer laser light source 2 dimming unit 5A, 5B input optical system 7, 12 fly-eye lens 13 fly-eye lens for annular illumination 15 σ diaphragm unit 16 split prism plate 18 fixed field diaphragm 20 Movable Field Stop 25 Reticle Stage 29 Main Controller 31 Exposure Control System 32 Illumination Control System 35 Z Tilt θ Stage 42 Variable Aperture Stop 44 Lens Drive Mechanism 46 Resist Thickness Measuring Device 47 Film Thickness Measurement Calculator 48 Focus Position Detection System 50 Alignment sensor 51A FIA system 51B LIA system 53A, 53B Beam splitter 55A, 55B Reflectance monitor 56 Integrated exposure amount calculation unit 58A, 58B Integrator sensor 59 Imaging characteristic fluctuation amount calculation unit 84 Image sensor

Claims (4)

[Claims] 1. An exposure method of transferring and exposing a transfer pattern on a mask onto a substrate coated with a photosensitive material under exposure illumination light, wherein the thickness of the photosensitive material coated on the substrate is adjusted. An exposure method, wherein the integrated exposure amount of the exposure illumination light is controlled in accordance with the above. 2. The exposure method according to claim 1, wherein an appropriate integrated exposure amount for the photosensitive material having a predetermined reference thickness corresponding to the mask is stored in advance, An exposure method, wherein a difference in thickness from the predetermined reference thickness is obtained, and an integrated exposure amount of the exposure illumination light is set based on the difference. 3. The exposure method according to claim 1, wherein the transfer pattern on the mask is projected and exposed on the substrate via a projection optical system, and the photosensitive material applied on the substrate. Material thickness is 0.5μ
m or less, and the numerical aperture of the projection optical system is 0.68 or more. 4. An exposure method in which a transfer pattern on a mask is transferred and exposed on a substrate coated with a photosensitive material under exposure illumination light, wherein the thickness of the photosensitive material coated on the substrate is adjusted. An exposure method, wherein the switching of the numerical aperture of the projection optical system is controlled accordingly.