CN111656284A

CN111656284A - Exposure apparatus and exposure method

Info

Publication number: CN111656284A
Application number: CN201980009581.6A
Authority: CN
Inventors: 加藤正纪
Original assignee: Nikon Corp
Current assignee: Nikon Corp
Priority date: 2018-01-24
Filing date: 2019-01-15
Publication date: 2020-09-11
Anticipated expiration: 2039-01-15
Also published as: KR102604340B1; JP2022051810A; JPWO2019146448A1; CN111656284B; TW201940986A; TW202349136A; WO2019146448A1; KR20230155617A; KR20200108068A; TWI815848B

Abstract

The exposure apparatus (EX) is provided with: light sources (2A, 2B, 2C) that generate light including a plurality of bright wavelengths (g-rays, h-rays, i-rays, etc.) in order to illuminate the cover substrate M; 1 st illumination optical system having: a wavelength selection unit (6A, 6B, 6C) that receives light from the light source (2A, 2B, 2C) and extracts an illumination light beam that includes at least 1 specific bright line wavelength (i-ray) among a plurality of bright line wavelengths and is limited to a predetermined wavelength width, and a numerical aperture variable unit (8A, 8B, 8C) that adjusts the divergence angle of the illumination light beam; and a2 nd illumination optical system (ILn) including an optical integrator (fly eye lens system FEn) for irradiating the illumination light beam on the mask substrate with the same illuminance with a numerical aperture corresponding to the divergence angle; the 1 st wavelength selection element (interference filter SWb) is attached to the wavelength selection unit, and one of the elements removes a bright line on the long wavelength side and a bright line on the short wavelength side which appear near the specific bright line wavelength (i-ray), and extracts a spectral component of the specific bright line wavelength and a spectral component of low luminance which is distributed near the bottom of the specific bright line wavelength.

Description

Exposure apparatus and exposure method

Technical Field

The present invention relates to an exposure apparatus and an exposure method for transferring a pattern of a mask onto a substrate.

Background

Conventionally, in a photolithography process for manufacturing electronic devices such as liquid crystal display devices, semiconductor devices, thin film magnetic heads, and the like, an exposure apparatus has been used which irradiates a transmissive or reflective mask substrate with illumination light from a light source and projection-exposes a substrate to be exposed, such as a plate coated with a photosensitive agent such as a photoresist, by a projection optical system with transmitted light or reflected light from a device pattern (pattern for electronic devices) formed on the mask substrate. As a conventional exposure apparatus, for example, an illumination system (illumination apparatus) is known which is provided as disclosed in japanese patent laid-open publication No. 2012-049332: illumination lights from light source units such as 2 mercury lamps are combined by a bundle fiber having a circular entrance side and a rectangular (slit) exit side, and then koehler illumination is performed on a slit-shaped illumination area on a mask substrate with a uniform illuminance distribution by an integrator included in a fly-eye lens optical system or the like.

When a mercury lamp (ultra-high pressure mercury discharge lamp or the like) is used as the light source, the discharge arc of the mercury lamp contains a plurality of bright lines, and a specific bright line wavelength among the bright lines is selected as the exposure illumination light (the illumination light of the mask substrate). In the photolithography step, in consideration of the photosensitive wavelength characteristics of the resist, the optical performance (resolution, chromatic aberration characteristics) of the projection optical system, and the like, g-rays (central wavelength of 435.835nm), h-rays (central wavelength of 404.656nm), and i-rays (central wavelength of 365.015nm) in the ultraviolet wavelength region of the bright wavelength of the mercury lamp are mainly used. The resolution R represented by the minimum line width value at which projection exposure can be performed is defined by R ═ k · (λ/NAp) when the numerical aperture on the image side (the substrate side to be exposed) of the projection optical system is NAp, the wavelength of the illumination light is λ (nm), and the program constant is k (0 < k ≦ 1). Therefore, by using the i-ray having the shortest wavelength among the 3 bright wavelengths, projection exposure (high-resolution exposure) of a finer mask pattern can be performed. However, in recent years, since the number of exposure steps for a negative resist layer (photosensitive layer) having a lower sensitivity than that of a positive resist layer has increased, it is necessary to set a longer exposure time, and there is a concern that the number of processed sheets per unit time of an exposed substrate may be reduced (reduction in productivity).

Disclosure of Invention

According to the 1 st aspect of the present invention, there is provided an exposure apparatus for projection-exposing a pattern of a mask onto a photosensitive substrate, comprising: a light source that generates light including a plurality of bright wavelengths to illuminate the mask; 1 st illumination optical system having: a wavelength selecting section for receiving light from the light source and extracting an illumination light beam including at least 1 specific bright line wavelength among the plurality of bright line wavelengths and limited to a predetermined wavelength width, and a numerical aperture varying section for adjusting a divergence angle of the illumination light beam; and a2 nd illumination optical system including an optical integrator that enters the illumination light flux with the adjusted divergence angle and irradiates the illumination light flux on the mask with the same illuminance with a numerical aperture corresponding to the divergence angle; and a1 st wavelength selecting element for extracting a spectral component of the specific bright line wavelength and a spectral component of low luminance distributed in the vicinity of the bottom of the specific bright line wavelength while removing a bright line on the long wavelength side and a bright line on the short wavelength side which appear near the specific bright line wavelength.

According to the 2 nd aspect of the present invention, there is provided an exposure method for projection-exposing a pattern of a mask onto a photosensitive substrate, comprising: selecting wavelengths such that, together with a peak-like spectral component of at least 1 specific bright line wavelength in light from a light source that generates light including a plurality of bright line wavelengths, spectral components of low luminance distributed in the vicinity of the bottom of the specific bright line wavelength, which do not include bright lines on the long wavelength side and bright lines on the short wavelength side appearing beside the specific bright line wavelength, are extracted; and a projection optical system of a mirror projection method in which an illumination beam of the spectral component having the selected wavelength is irradiated onto the mask with the same illuminance, and the pattern of the mask is projected and exposed on the substrate by a mirror projection method in which no chromatic aberration occurs in the wavelength width of the spectral component having low luminance or a catadioptric method in which chromatic aberration is corrected in the wavelength width of the spectral component having low luminance.

According to the 3 rd aspect of the present invention, there is provided an exposure method for projecting and exposing an image of a pattern on a photosensitive substrate by a projection optical system into which an exposure light beam generated by a mask is incident, the exposure method including the steps of irradiating a mask carrying the pattern for electronic components with light having a spectral distribution including a specific bright line wavelength selected by a wavelength selecting unit among lights including bright line wavelengths generated by a light source device through an illumination optical system, and exposing the image of the pattern to the photosensitive substrate by the projection optical system, the exposure method including: extracting, by the wavelength selection unit, a1 st spectral distribution light and a2 nd spectral distribution light having different wavelength bands from the light generated by the light source device; and forming a1 st light source image distributed in a two-dimensional range by the 1 st spectrally distributed light and a2 nd light source image distributed in a two-dimensional range by the 2 nd spectrally distributed light on a pupil plane in the illumination optical system so as to perform kohler illumination on the mask by the illumination optical system.

According to the 4 th aspect of the present invention, there is provided an exposure method for projecting and exposing an image of a mask pattern onto a substrate by a projection optical system for projecting an image beam generated by the mask pattern onto the substrate by illuminating the mask pattern with illumination light having a predetermined wavelength distribution, the exposure method comprising: setting a width of a wavelength distribution of the illumination light including a center wavelength λ such that a ratio of a minor axis length to a major axis length of a projection image of the hole pattern deformed into an elliptical shape becomes 80% or more, preferably 90% or more, when a projection image of a square or rectangular hole pattern having a size close to a minimum resolvable line width dimension determined by a resolution R defined by k · (λ/NAp) is projected onto the substrate, assuming that the specific center wavelength in the wavelength distribution of the illumination light is λ, a numerical aperture on the substrate side of the projection optical system is NAp, and a program constant is k (0 < k ≦ 1); and illuminating the mask on which the pattern for electronic components is formed with the illumination light of the wavelength distribution of the set width, and projection-exposing the pattern for electronic components on the substrate.

Drawings

Fig. 1 is a perspective view showing a schematic configuration of a scanning type projection exposure apparatus according to embodiment 1.

Fig. 2 is a diagram showing the arrangement of optical members of a projection optical system incorporated in the projection exposure apparatus shown in fig. 1.

Fig. 3 is a perspective view showing a schematic overall configuration of an illumination device for irradiating a mask substrate loaded in the projection exposure apparatus shown in fig. 1 with exposure illumination light.

Fig. 4 is a perspective view schematically showing the configuration of the 1 st illumination optical system from the mercury lamp to the optical fiber (optical fiber bundle) in the illumination device shown in fig. 3.

Fig. 5 is a graph schematically showing an example of wavelength characteristics (spectral distribution) of light generated by arc discharge of the ultra-high pressure mercury discharge lamp.

Fig. 6 is a graph schematically showing a state where light having a narrow wavelength width including i-rays is selectively extracted from the wavelength characteristics (spectral distribution) shown in fig. 5 by the i-ray-narrow band interference filter.

Fig. 7 is a graph schematically showing a state in which light having a wide wavelength width including i-rays and the vicinity of the bottom thereof is selectively extracted from the wavelength characteristics (spectral distribution) shown in fig. 5 by the i-ray-wide band interference filter.

Fig. 8 is a graph schematically showing a state where light having a wide wavelength width including both i-rays and h-rays is selectively extracted from the wavelength characteristics (spectral distribution) shown in fig. 5 by the i-ray + h-ray interference filter.

Fig. 9 is a perspective view schematically showing the entire configuration of an optical fiber (optical fiber bundle) provided in the lighting device shown in fig. 3 and the shapes of the incident end and the exit end, respectively.

Fig. 10 is a perspective view schematically showing the configuration of the 2 nd illumination optical system that irradiates illumination light from the emission end of the optical fiber bundle to the illumination region on the mask substrate in the illumination device shown in fig. 3.

Fig. 11 is a diagram schematically showing a state of illumination light on an optical path from an emission end of the bundle fiber shown in fig. 10 to the fly eye lens system.

Fig. 12 is a view schematically showing an example of the arrangement of a plurality of spot light source images formed on each optical fiber line at the emission end of the optical fiber bundle shown in fig. 11.

Fig. 13 is a diagram showing an arrangement state of a plurality of spot light source images formed at emission ends of a plurality of lens elements constituting the fly-eye lens system shown in fig. 11.

Fig. 14 is a diagram schematically showing a state of illumination light in an optical path from the fly-eye lens system shown in fig. 10 to an illumination region on the mask substrate.

Fig. 15 is a diagram illustrating an operation of adjusting the numerical aperture (divergence angle) of the illumination light beam applied to the incident end of the optical fiber bundle by the magnification varying unit (numerical aperture varying unit) shown in fig. 4.

Fig. 16 is a diagram schematically showing states of light beams incident on 3 optical fiber bundles on the incident side of the optical fiber bundle shown in fig. 9 and illumination light beams emitted from 6 optical fiber bundles on the emission side.

Fig. 17 is a schematic diagram of an optical path from the emission end of the bundle fiber to the incident surface of the fly eye lens system when viewed from the X direction (scanning movement direction).

Fig. 18 is a diagram of a state in which circular regions CFa, CFb, and CFc in fig. 17 distributed on the incident surface of the fly-eye lens system are viewed in the XY plane.

Fig. 19 a is a diagram showing the distribution of the spot light (spot light source image) formed on the exit surface of the fly-eye lens system viewed from the X direction (scanning movement direction), and fig. 19B is a diagram showing the distribution of the spot light (spot light source image) formed on the exit surface of the fly-eye lens system viewed from the Y direction (stepping movement direction).

Fig. 20 is a diagram schematically showing alignment characteristics (characteristics of a divergence angle) of the illumination light beam Irn irradiated to the point OP on the illumination area on the mask substrate.

Fig. 21 is a table summarizing examples of combinations of the i-ray-narrow band interference filter SWa of fig. 6, the i-ray-wide band interference filter SWb of fig. 7, and the i-ray + h-ray-interference filter SWc of fig. 8, which are attached to each of the 3

wavelength selection units

6A, 6B, and 6C.

Fig. 22 is a graph schematically showing the wavelength characteristics of the illumination light beam of the mask substrate obtained by combining the interference filters represented by the combination code B2 in the table of fig. 21.

Fig. 23 is a graph showing an example of wavelength-dependent light absorption characteristics of a negative photoresist for the purpose of describing modification example 3.

Fig. 24 is a cross-sectional view schematically showing a tilt generated at an edge portion (sidewall) of a resist image of a residual film after development for the purpose of description of modification 3.

Fig. 25(a) is a view showing the configuration of modification 4 and showing the shape of the aperture plate APa on which the annular translucent portion is formed, and fig. 25(B) is a view showing the configuration of modification 4 and showing the shape of the aperture plate APb on which the 4-pole translucent portion is formed.

Fig. 26 is a diagram showing the configuration of modification 5, and shows a state in which an annular diaphragm plate is arranged in the wavelength selection portion 6A of the 1 st illumination optical system.

Fig. 27 is a view showing a schematic overall configuration of an exposure apparatus according to embodiment 2.

Fig. 28 is a graph showing detailed spectral characteristics obtained when the wavelength characteristics of the ultrahigh-pressure mercury discharge lamp shown in fig. 5 are measured by a spectrometer having a high wavelength resolution.

Fig. 29 is a graph showing a relationship between chromatic aberration characteristics of the projection optical system and a bright line wavelength of an i-ray of the mercury lamp.

Fig. 30 is a graph illustrating differences in wavelength characteristics between a high-pressure mercury discharge lamp and an ultrahigh-pressure mercury discharge lamp.

Fig. 31 is a diagram illustrating distortion in the shape of a projection image obtained when hole patterns having different sizes formed on a mask are projected onto a substrate.

Fig. 32 is a diagram illustrating a method of determining the ellipticity (ellipticity) when the projected image of the hole pattern is distorted into an elliptical shape.

Fig. 33 is a diagram illustrating the inclination of the projection image of the hole pattern distorted into an elliptical shape shown in fig. 32.

Fig. 34 is a graph showing an example of the change characteristic of the refractive index of synthetic quartz with respect to wavelength.

Fig. 35 is a diagram schematically illustrating a state of image shift caused by a parallel flat plate-shaped quartz plate provided as an image shift optical member.

Fig. 36 is a graph showing an example of a difference amount due to a relative positional shift between a projected image formed by i-rays and a projected image formed by h-rays, which varies with the inclination angle of the quartz plate of the image shift optical member.

Detailed Description

A preferred embodiment of an exposure apparatus according to an aspect of the present invention is disclosed, and the following description is made in detail with reference to the accompanying drawings. The aspect of the present invention is not limited to the embodiments, and various changes and modifications may be made. That is, the following constituent elements include those which can be easily assumed by a person having ordinary skill in the art to which the present invention pertains and substantially the same, and the following constituent elements may be appropriately combined. Various omissions, substitutions, and changes in the components may be made without departing from the spirit of the invention.

[ embodiment 1]

Fig. 1 is a perspective view showing a schematic overall configuration of a scanning type projection exposure apparatus EX according to embodiment 1, and fig. 2 is a view showing an arrangement of optical members of a partial projection optical system PLn incorporated in the projection exposure apparatus EX of fig. 1. In fig. 1 and 2, a direction in which a Z axis of XYZ, which is an orthogonal coordinate, extends represents a gravity direction, a direction in which an X axis extends represents a scanning movement direction in which a plate P as an exposure target substrate (photosensitive substrate) and a mask substrate M move for scanning exposure, and a direction in which a Y axis extends represents a direction of stepping movement of the plate P. The projection exposure apparatus EX of the present embodiment is described as a step-and-scan type exposure apparatus that transfers an image of an electronic element pattern formed on a flat mask substrate M to a flat plate P coated with a photosensitive layer (photoresist or the like) while synchronously moving the mask substrate M and the plate P in the X direction, for a projection optical system including 6 partial projection optical systems PL1 to PL6 of a catadioptric system. The projection exposure apparatus EX shown in fig. 1 and 2 has the same configuration as that disclosed in, for example, the pamphlet of international publication No. 2009/128488 or japanese patent application laid-open No. 2010-245254, and therefore the configuration of the apparatus shown in fig. 1 and 2 will be described in a simple manner.

[ constitution of projection optical System ]

The 6 illumination areas IA1 to IA6 (see fig. 1) set on the mask substrate M are each set to have a rectangular shape in which the dimension in the X direction, which is the scanning direction, is short relative to the dimension in the Y direction, which is the stepping direction. Illumination light for exposure adjusted to have a uniform illuminance distribution (for example, uniformity within ± 5%) is projected on each of illumination areas IA1 to IA6 by an illumination device described later. The 6 illumination regions IA1 to IA6 are set at positions on the object plane side of the 6 partial projection optical systems PL1 to PL6, respectively. For example, when a pattern portion of the mask substrate M appears in the illumination area IA1, the transmitted light generated by the pattern portion is reflected by the upper reflection surface of the prism PMa and enters the partial projection optical system PL 1. The partial projection optical system PL1 forms an intermediate image of the illumination area IA1 on the intermediate image plane IM1 at an equal magnification by reflecting the transmitted light (imaging light beam, exposure light beam) from the pattern portion on the reflection surface on the lower side of the prism PMa by the 1 st imaging system PL1a including the lens systems Ga1, Ga2, Ga3 and the concave mirror Ga4 arranged along the optical axis AXa as shown in fig. 2.

As shown in fig. 1, a field stop FA1 having a trapezoidal opening that inclines both end edges in the Y direction is disposed on the intermediate image plane IM 1. The imaging light beam transmitted through the opening of the field stop plate FA1 is reflected on the reflection surface on the upper side of the prism PMb, and is reflected in the direction of the plate P (-Z direction) on the reflection surface on the lower side of the prism PMb by the 2 nd imaging system PL1b including the lens systems Gb1, Gb2, Gb3, and concave mirror Gb4 arranged along the optical axis AXb, as shown in fig. 2. Thereby, the intermediate image formed in the opening of the field stop plate FA1 is re-imaged and imaged at an equal magnification within the trapezoidal projection area EA1 set on the plate P. The partial projection optical system PL1 telecentric images the image of the pattern portion in the illumination area IA1 in the projection area EA1 in an erect and erect equal-magnification relationship by the 1 st imaging system PL1a and the 2 nd imaging system PL1 b.

As shown in fig. 2, the 1 st imaging system PL1a is a half-field-of-view type imaging system of a catadioptric system in which concave mirror Ga4 is disposed on pupil plane Epa, and the 2 nd imaging system PL1b is also a half-field-of-view type imaging system of a catadioptric system in which concave mirror Gb4 is disposed on pupil plane Epb. The pupil surfaces Epa and Epb are optically conjugate to each other, and a light source image (2-order light source image) is formed on each of the pupil surfaces Epa and Epb in the illumination device that illuminates the illumination area IA 1. In the imaging optical path of the partial projection optical system PL1, a focus adjustment optical member FC1 for finely adjusting the focus state (focal state) of the image projected on the projection area EA1 on the plate P is provided between the mask substrate M and the prism PMa. Further, an image shift optical member SC1 for fine-adjusting the position of a projection area EA1 projected on the plate P in the X direction and the Y direction independently is provided between the field stop plate FA1 and the prism PMb; and a magnification adjustment optical member MC1 for fine adjustment in a range of an image of the pattern portion projected on the projection area EA1 to a size of less than about ± several tens ppm is provided between the prism PMb and the plate P. The focus adjustment optical member FC1, the image shift optical member SC1, and the magnification adjustment optical member MC1 are disclosed in, for example, international publication No. 2013/094286 pamphlet, and therefore detailed descriptions of the configuration and functions are omitted.

In the present embodiment, as shown in fig. 2, the partial projection optical system PL1 includes: the 1 st imaging system PL1a, the 2 nd imaging system PL1b, the prisms PMa, PMb, the field stop plate FA1, the focus adjustment optical member FC1, the image shift optical member SC1, and the magnification adjustment optical member MC1, and the other partial projection optical systems PL2 to PL6 are also configured in the same manner. Therefore, the other partial projection optical systems PL2 to PL6 also form images of the pattern portion of the mask substrate M at equal magnification in the trapezoidal projection areas EA2 to EA6 set on the plate P, respectively. Thus, when the mask substrate M and the plate P are moved one-dimensionally in the X direction at the same speed to perform scanning exposure, the pattern portions exposed on the photosensitive layer of the plate P are merged in the Y direction in each of the 6 projection areas EA1 to EA 6. In addition, when it is not necessary to make a special distinction, the partial projection optical systems PL1 to PL6, the illumination areas IA1 to IA6, and the projection areas EA1 to EA6 described above are also referred to as a partial projection optical system PLn, an illumination area IAn, and a projection area EAn, respectively (n is 1 to 6).

[ constitution of Lighting device ]

Fig. 3 is a perspective view showing a schematic overall configuration of an illumination device for projecting exposure illumination light to each of 6 illumination areas IA1 to IA6 set on a mask substrate M, and the XYZ system of orthogonal coordinates is set to be the same as that of fig. 1 and 2. As disclosed in japanese patent application laid-open No. 2010-245254, the lighting device of the present embodiment includes 3 mercury lamps (short-arc type ultra-high pressure mercury discharge lamps) 2A, 2B, and 2C (light source devices) of the same specification as light sources. The number of lamps in the light source device is determined by the number of partial projection optical systems PLn so that the illumination light projected on each illumination region IAn becomes a required illuminance value, but may be 2 or more. The vapor pressure of mercury sealed in discharge tube is 10 in the ultra-high pressure mercury discharge lamp⁶Pa (Pascal) or more, and generates g-rays (wavelength 435.835nm), h-rays (wavelength 404.656nm) and i-rays (wavelength 365.015nm) which are bright rays in the ultraviolet wavelength region with high brightness. Light emitting points (arc discharge portions) of the

mercury discharge lamps

2A, 2B, and 2C are disposed at 1 st positions of the

elliptical mirrors

4A, 4B, and 4C, respectivelyThe position of the focal point is such that the light flux BM reflected by the inner reflecting surfaces of the

elliptical mirrors

4A, 4B, and 4C converges (converges) toward the 2 nd focal point of the

elliptical mirrors

4A, 4B, and 4C.

The light beams BM radiated in the-Z direction from the

elliptical mirrors

4A, 4B, 4C, respectively, are separated as follows: the spectral components in the ultraviolet wavelength region for exposure (for example, a short wavelength region of 460nm or less) are reflected in the + X direction by the dichroic mirror DM disposed in front of the 2 nd focal point, and the spectral components in a longer wavelength region are transmitted. The light flux in the ultraviolet wavelength region for exposure reflected by each of the dichroic mirrors DM has the smallest beam diameter at the position of the 2 nd focal point of each of the

elliptical mirrors

4A, 4B, 4C, and the

rotary shutters

5A, 5B, 5C are disposed at the positions of the 2 nd focal point. The light beams in the ultraviolet wavelength region for exposure that have passed through the

rotary shutters

5A, 5B, and 5C are incident on the

wavelength selection units

6A, 6B, and 6C while diverging. Each of the

wavelength selection units

6A, 6B, and 6C includes a plurality of lens elements and a wavelength selection interference filter, and transmits only a desired bright wavelength portion of the incident light beam in the ultraviolet wavelength region for exposure. The interference filters provided in each of the

wavelength selection sections

6A, 6B, and 6C are provided with: the wavelength selection characteristics of the mask substrate M can be changed (switched) among a plurality of ones having different wavelength selection characteristics according to the fineness (resolution) of the pattern to be exposed on the mask substrate M or the exposure amount (dose) to be applied to the photosensitive layer of the plate P. The difference in the wavelength selection characteristics of the interference filter will be described in detail later, but the wavelength characteristics (wavelength distribution) of the exposure illumination light projected onto the illumination region IAn on the mask substrate M can be switched to characteristics suitable for pattern exposure with higher resolution and characteristics suitable for pattern exposure with higher illuminance for improving productivity. For this purpose, the interference filter is prepared in advance to have the following characteristics: a characteristic of transmitting any bright wavelength component of g-ray (wavelength 435.835nm), h-ray (wavelength 404.656nm), and i-ray (wavelength 365.015 nm); a characteristic of transmitting 2 continuous bright wavelength components (g-ray + h-ray, or i-ray + h-ray) among g-ray, h-ray, and i-ray; or a characteristic of transmitting all the bright wavelength components of g-rays, h-rays, and i-rays.

The light beams emitted from the

wavelength selection units

6A, 6B, and 6C are incident on the

variable magnification units

8A, 8B, and 8C, and the

variable magnification units

8A, 8B, and 8C adjust the numerical apertures (maximum inclination angles of principal rays) or the dimensions (diameters) in the diameter direction of the illumination light beams BMa, BMb, and BMc incident on the 3 optical fiber bundles (optical fibers, optical transmission elements) 12A, 12B, and 12C on the incident side of the light distribution unit 10 in the subsequent stage. Each of the

magnification varying sections

8A, 8B, and 8C includes a plurality of lens elements that are movable in the optical axis direction so as to be able to continuously adjust the Numerical Aperture (NA) of the illumination light beams BMa, BMb, and BMc incident on the

optical fiber bundles

12A, 12B, and 12C within a certain range. As a result, the radial sizes of the light source images (2-time light source images) distributed on the pupil planes Epa and Epb of the partial projection optical system PLn shown in fig. 2 from the optical axes AXa and AXb can be continuously changed by the respective

magnification varying sections

8A, 8B, and 8C. That is, each of the

magnification varying units

8A, 8B, and 8C can adjust the illumination σ value (0 < σ ≦ 1) determined by NAi/NAp, which is a ratio of the numerical apertures, when the maximum numerical aperture of the partial projection optical system PLn is referred to as NAp and the numerical aperture of the illumination light beam projected to the illumination area IAn is referred to as NAi. Therefore, each of the

magnification varying units

8A, 8B, and 8C is also referred to as a numerical aperture varying unit that can continuously adjust the illumination σ value (the numerical aperture NPi of the illumination light beam). The configuration from the elliptical mirror 4A to the magnification variable section 8A, the configuration from the elliptical mirror 4B to the magnification variable section 8B, and the configuration from the elliptical mirror 4C to the magnification variable section 8C shown in fig. 3 are also collectively referred to as the 1 st illumination optical system, and the functions thereof will be described in detail later.

The light distribution section 10 divides the illumination light beams BMa, BMb, and BMc incident from the 3 incident-side

optical fiber bundles

12A, 12B, and 12C into 6 emission-side optical fiber bundles FG1 to FG6 so as to be distributed to the 2 nd illumination optical systems IL1 to IL6 arranged corresponding to the 6 illumination areas IAn, respectively. The 2 nd illumination optical systems IL1 to IL6 each use the emission ends of the optical fiber bundles FG1 to FG6 as light source images (2 nd light source images obtained by collecting a plurality of point light sources), and each illumination region IAn performs kohler illumination. In addition, when no particular distinction is required, each of the 2 nd illumination optical systems IL1 to IL6 and the fiber bundles FG1 to FG6 described above is also referred to as a2 nd illumination optical system ILn and a fiber bundle FGn (n is 1 to 6).

[ 1 st illumination optical System ]

Fig. 4 is a perspective view showing a detailed configuration of the 1 st illumination optical system arranged on an optical path from the mercury lamp 2A shown in fig. 3 to the optical fiber bundle 12A on the incident side, and the XYZ system is set to the same orthogonal coordinates as those in fig. 1 to 3. The 1 st illumination optical system from the mercury lamp 2B to the incident-side optical fiber bundle 12B and the 1 st illumination optical system from the mercury lamp 2C to the incident-side optical fiber bundle 12C are also configured in the same manner as in fig. 4. As shown in fig. 4, the light flux BM immediately after being emitted along the optical axis AX1 from the emission opening (-Z direction end) of the elliptical mirror 4A passes through the upper (+ Z direction) opening of the elliptical mirror 4A and the lower electrode portion of the mercury lamp 2A, and has a ring-shaped intensity distribution centered on the optical axis AX1, that is, a hollow distribution in which the illuminance at the center is extremely low. The light beam BM is focused toward the position PS1 of the 2 nd focal point of the elliptical mirror 4A disposed on the rotary blade of the rotary shutter 5A, but since the arc discharge portion generated between the electrodes of the mercury lamp 2A is distributed in a slender manner in the direction of the optical axis AX1, the light beam BM is not focused in a point shape at the position PS1, and becomes a beam waist having a finite size (diameter).

The wavelength selection unit 6A is provided with: a lens system (collimator lens) 6a1 that receives the light beam BM diverging from the position PS1 of the 2 nd focal point and converts the light beam BM into a substantially parallel light beam; a slide mechanism FX that holds 2-piece interference filters (wavelength selection members, wavelength selection elements, band pass filters) SWa, SWb having mutually different wavelength selection characteristics and switches either one of the interference filters SWa, SWb so as to be inserted into and removed from an optical path; and a lens system 6a2 that condenses (converges) the light beam BMa transmitted from either one of the interference filters SWa, SWb on a focal position PS2 (a position optically conjugate to the position PS 1). The slide mechanism FX has a configuration in which the interference filters SWa, SWb are easily removed or attached, respectively. When using the 3 rd interference filter (wavelength selection means, wavelength selection element, band pass filter) having a wavelength selection characteristic different from that of either one of the interference filters SWa, SWb, it is sufficient to remove either one of the interference filters SWa, SWb from the slide mechanism FX and mount the 3 rd interference filter in place of it in a state where the light beam BM from the mercury lamp 2A is shielded by the rotary shutter 5A. In addition, when the slide mechanism is not provided, attachment mechanisms that can easily attach and detach the interference filters SWa and SWb are attached.

The light beam BMa emitted from the wavelength selection unit 6A becomes a beam waist at the focal position PS2, and then enters the magnification variable unit 8A in a divergent state. At the focal position PS2, a circular light source image is formed due to a blurred image of the arc discharge portion (light emitting point) of the mercury lamp 2A. The magnification varying unit 8A has 2 lens systems 8A1, 8A2 whose positions along the optical axis AX1 can be adjusted. The light beam BMa diverging from the focal position PS2 and traveling through the lens systems 8a1 and 8a2 is projected with a predetermined beam diameter or a predetermined numerical aperture and condensed on the incident end FBi of the incident-side optical fiber bundle 12A. The incident end FBi of the optical fiber bundle 12A is basically disposed so as to optically conjugate with the focal position PS2, but the conjugate may be intentionally released by adjusting the positions of the lens systems 8A1 and 8A2 of the magnification varying unit 8A. The 2 lens systems 8a1, 8a2 function as a variable magnification relay system, and as a result, the diameter of the light beam BMa condensed at the incident end FBi of the optical fiber bundle 12A becomes smaller or larger with respect to the effective maximum diameter of the incident end FBi as the numerical aperture of the illumination light beam BMa changes.

[ wavelength selection by interference filter ]

Here, an example of wavelength selection by the interference filter attachable to the slide mechanism FX of the wavelength selection section 6A will be described with reference to fig. 5 to 8. Fig. 5 is a graph schematically showing an example of the wavelength characteristic (spectral distribution) of the luminous flux BM generated by arc discharge of a mercury lamp (ultra-high pressure mercury discharge lamp). Fig. 6 is a graph schematically showing a state where light having a narrow wavelength width including i-rays is selectively extracted from the spectral distribution of fig. 5 by the i-ray-narrow band interference filter SWa, fig. 7 is a graph schematically showing a state where light having a relatively wide wavelength width including i-rays and low-luminance portions near the bottom thereof is selectively extracted from the spectral distribution of fig. 5 by the i-ray-wide band interference filter SWb, and fig. 8 is a graph schematically showing a state where light having a wide wavelength width including both i-rays and h-rays is selectively extracted from the spectral distribution of fig. 5 by the i-ray + h-ray-interference filter SWc (3 rd interference filter). In each of the graphs of fig. 5 to 8, the horizontal axis represents the wavelength (nm) and the vertical axis represents the relative intensity (%). In the wavelength characteristics (spectral distribution) of the luminous flux BM from the ultra-high pressure mercury discharge lamp shown in fig. 5 (and fig. 6 to 8), the peak-like spectral portions of the g-ray, the h-ray, the i-ray, and the j-ray, which are main bright lines, are shown as the wavelength widths when measured by a spectrometer having a wavelength resolution that is not too high, and the actual wavelength widths of the peak-like spectral portions are about several nm to ten and several nm when defined by a full width at half maximum (a width of an intensity that is half of the peak intensity).

In the present embodiment, as shown in fig. 6 to 8, 3 types of interference filters SWa, SWb, and SWc are prepared. The i-ray narrow-band interference filter SWa has the following wavelength selection characteristics as shown in fig. 6: a transmittance of 10% or more at a wavelength of about 354nm to about 380 nm; and a transmittance of 90% or more at a wavelength of about 359nm to about 377 nm. Therefore, the full width at half maximum of the wavelength width selected by the i-ray/narrow-band interference filter SWa includes the bright line wavelength (365.015nm) of the i-ray and is about 22 nm. As shown in fig. 7, the i-ray broadband interference filter SWb has the following wavelength selection characteristics: a transmittance of 10% or more at a wavelength of from about 344nm to about 398 nm; and a transmittance of 90% or more at a wavelength of about 350nm to about 395 nm. Therefore, the full width at half maximum of the wavelength width selected by the i-ray broadband interference filter SWb includes the bright line wavelength (365.015nm) of the i-ray and is about 49 nm. Both the i-ray-narrow band interference filter SWa and the i-ray-wide band interference filter SWb select only the bright band of the i-ray as the exposure illumination light, but since the bandwidth of the wavelength selection of the i-ray-narrow band interference filter SWa is narrow, the monochromaticity of the i-ray (narrow) indicated by the hatched portion in fig. 6 is better than the i-ray (wide) indicated by the hatched portion in fig. 7 selected by the i-ray-wide band interference filter SWb, the influence due to the chromatic aberration characteristics of the partial projection optical system PLn is reduced, and a pattern exposure with higher resolution can be performed.

However, since the amount of light (area of hatched portion in fig. 6) of the i-ray (narrow) obtained by the i-ray-narrow band interference filter SWa is smaller than the amount of light (area of hatched portion in fig. 7) of the i-ray (wide) obtained by the i-ray-wide band interference filter SWb, the moving speed of the mask substrate M and the plate P during the scanning exposure must be slightly lowered, which leads to a reduction in productivity. On the other hand, the i-ray (width) obtained by the i-ray-wide band interference filter SWb contains spectral components in the vicinity of the bottom of low luminance between the bright line wavelength (365.015nm) of the i-ray and the h-ray located near the long wavelength side and between the bright line wavelength and the relatively strong peak wavelength located near the short wavelength side, and therefore the light amount can be increased by several% or more while performing high-resolution pattern exposure, and productivity can be improved. The bandwidth (about 49nm in full width at half maximum) selected by the wavelength of the i-ray broadband interference filter SWb is determined by the contrast value of the pattern projected image having the minimum line width obtained based on the chromatic aberration characteristics of the partial projection optical system PLn. The chromatic aberration of the partial projection optical system PLn includes a magnification (lateral) chromatic aberration and an on-axis (longitudinal) chromatic aberration, and is corrected to have a chromatic aberration characteristic having a tendency (tendency of a quadratic function) as follows, for example, in a projection optical system dedicated only to a bright wavelength of an i-ray: the chromatic aberration amount becomes substantially zero at the bright wavelength of the i-ray, and increases on the short wavelength side and the long wavelength side thereof. In a projection optical system which allows 2 bright wavelengths using i-rays and h-rays, the chromatic aberration characteristic is corrected to have a tendency as follows: the amount of chromatic aberration is made substantially zero at a wavelength substantially in the middle between the i-ray and the h-ray, and the rate of change in the amount of chromatic aberration is reduced between the respective bright line wavelengths of the i-ray and the h-ray.

When 2 bright wavelengths of i-rays and h-rays are used, the i-ray + h-ray interference filter SWc is attached to the slide mechanism FX, and light i-rays + h-rays having a spectral distribution shown by the hatched portion in fig. 8 are used. The i-ray + h-ray interference filter SWc has the following wavelength selection characteristics as shown in fig. 8: a transmittance of 10% or more at a wavelength of about 344nm to about 420 nm; and a transmittance of 90% or more at a wavelength of about 350nm to about 415 nm. Therefore, the full width at half maximum of the wavelength width selected by the i-ray + h-ray interference filter SWc includes the bright line wavelength (365.015nm) of the i-ray and the bright line wavelength (404.656nm) of the h-ray, and is about 70 nm. In the pattern exposure using 2 bright wavelengths of i-rays and h-rays, the minimum line width that can be resolved is larger than that in the pattern exposure using only i-rays, but the light quantity (area of the hatched portion in fig. 8) of i-rays + h-rays obtained by the i-rays + h-rays-interference filter SWc is overwhelmingly increased as compared with the case of the i-rays-narrow band interference filter SWa in fig. 6 or the i-rays-wide band interference filter SWb in fig. 7, and productivity is dramatically improved. Therefore, when the pattern of the mask substrate M projected and exposed on the photosensitive layer of the plate P does not include a pattern having a critical line width with a high fineness, the i-ray + h-ray interference filter SWc can be used to perform pattern exposure with high productivity.

[ light distributing section 10]

Fig. 9 is a perspective view schematically showing the entire configuration of the optical fiber bundle as the light distributing section 10 provided in the illumination device shown in fig. 3, the shape of the incident end FBi of each of the

optical fiber bundles

12A, 12B, and 12C on the incident side, and the shape of the exit end FBo of each of the optical fiber bundles FG1 to FG6 on the exit side, and the XYZ system of orthogonal coordinates is set to be the same as that of fig. 3. The incident end FBi of each of the incident-side

optical fiber bundles

12A, 12B, and 12C is formed by gathering a plurality of optical fiber strands into a circular shape having an end face with a diameter of several tens of mm or more. The plurality of optical fibers of the

optical fiber bundles

12A, 12B, and 12C are separated in the line distributing section 10a in the light distributing section 10 so that the 6 optical fiber bundles FG1 to FG6 include substantially equal numbers of lines. The shape of the emission end FBo of each of the optical fiber bundles FG1 to FG6 is a rectangle that is similar to the shape of the illumination region IAn on the mask substrate M, with a plurality of optical fiber lines being collected. The 1 fiber bundle FGn is collected so as to include approximately the same number of fiber lines from the respective incident-

side fiber bundles

12A, 12B, and 12C. For example, in the case where the

optical fiber bundles

12A, 12B, and 12C are each configured by collecting 12 ten thousand optical fiber lines (36 ten thousand in total), 1 optical fiber bundle FGn is configured by collecting 6 ten thousand optical fiber lines. Of the 6 ten thousand fiber lines of the fiber bundle FGn, about every 2 ten thousand fiber lines are formed of fiber lines from the incident-

side fiber bundles

12A, 12B, and 12C. Further, 1 fiber line is a silica fiber having an outer shape (clad) and a diameter of about 0.2 mm.

The optical fiber emits the light beam from the emission end while maintaining the numerical aperture (convergence angle or divergence angle) of the light beam irradiated to the incidence end. Therefore, the numerical aperture (convergence angle or divergence angle) of the illumination light beam BMa irradiated onto the incident end FBi of the optical fiber bundle 12A when emitted as the illumination light beam BSa from the emission end FBo of the optical fiber bundle FGn is the same as the numerical aperture of the illumination light beam BMa, the numerical aperture (convergence angle or divergence angle) of the illumination light beam BMb irradiated onto the incident end FBi of the optical fiber bundle 12B when emitted as the illumination light beam BSb from the emission end FBo of the optical fiber bundle FGn is the same as the numerical aperture of the illumination light beam BMb, and the numerical aperture (convergence angle or divergence angle) of the illumination light beam BMc irradiated onto the incident end FBi of the optical fiber bundle 12C when emitted as the illumination light beam BSc from the emission end FBo of the optical fiber bundle FGn is the same as the numerical aperture of the illumination light beam BMc. Therefore, when the numerical apertures (convergence angles) of the illumination light beams BMa, BMb, and BMc irradiated onto the incident ends FBi of the

optical fiber bundles

12A, 12B, and 12C are NAia, NAib, and NAic, and the

magnification variable sections

8A, 8B, and 8C shown in fig. 3 (fig. 4) are adjusted so that NAia is NAib is NAic, the numerical apertures (divergence angles) of the illumination light beams BSa, BSb, and BSc emitted from the emission ends FBo of the optical fiber bundles FGn are the same. When the magnification-varying

sections

8A, 8B, and 8C are adjusted so that the numerical apertures (convergence angles) NAia, NAib, and NAic of the illumination light beams BMa, BMb, and BMc are different from each other, the numerical apertures (divergence angles) of the illumination light beams BSa, BSb, and BSc emitted from the emission end FBo of each fiber bundle FGn are different from each other.

[ 2 nd illumination optical System ILn ]

Fig. 10 is a perspective view schematically showing the configuration of the 2 nd illumination optical system ILn (IL1 to IL6) that irradiates the illumination light beams (BSa, BSb, BSc) from the emission ends FBo of the 6 optical fiber bundles FGn (FG1 to FG6) shown in fig. 3 (fig. 9) onto the illumination regions IAn on the mask substrate M, and the orthogonal coordinate system XYZ is set to be the same as that in fig. 3 or fig. 9. The 2 nd illumination optical system ILn includes: a1 st condenser lens system CFn (CF1 to CF6) arranged so that the position of the front focal point coincides with the emission end FBo, in order to make a plurality of point light source images formed on the emission end FBo of the fiber bundle FGn as light source images of kohler illumination; fly-eye lens system FEn (FE1 to FE6) having incidence plane poi set at the rear focal point of capacitor lens system CFn; and a2 nd condenser lens system CPn (CP1 to CP6) that sets the position of the front focal point on the emission surface epi of the fly-eye lens system FEn and sets the illumination region IAn (IA1 to IA6) at the position of the rear focal point, in order to make the light source image (2 nd order light source image) formed on the emission surface epi of the fly-eye lens system FEn as the light source image of kohler illumination.

The condenser lens system CFn and the condenser lens system CPn are arranged along an optical axis AX2 parallel to the Z axis, and the optical axis AX2 is set so as to pass through a geometric center point of the rectangular output end FBo of the fiber bundle FGn and a geometric center point of the fly eye lens system FEn in the XY plane. The fly-eye lens system FEn is configured by joining a plurality of lens elements Le having a rectangular cross section with the Y direction as the long side and the X direction as the short side in a stacked manner in the X direction and the Y direction so that the lens elements Le have a shape similar to the rectangular illumination region IAn when viewed in the XY plane. Convex surfaces (spherical lenses) having a predetermined focal distance are formed on the incidence surface poi side and the emission surface epi side of the lens element Le, respectively. The exit surface epi of the fly-eye lens system FEn is the position of the illumination pupil of the 2 nd illumination optical system ILn, and the overall outer shape range in the XY plane of the fly-eye lens system FEn is set so as to substantially include the diameter of the illumination pupil (circular shape).

Further, the emission surface epi of the fly-eye lens system FEn is set in a relationship (imaging relationship) optically conjugate with the emission end FBo of the fiber bundle FGn, and the incidence surface poi of the fly-eye lens system FEn is set in a relationship (imaging relationship) optically conjugate with the illumination region IAn (pattern surface of the mask substrate M). Therefore, a plurality of point light source images formed at emission end FBo of fiber bundle FGn are re-imaged on emission surface epi side of each of lens elements Le of fly's eye lens system FEn, and illumination region IAn is illuminated (imaged) in a shape similar to a rectangle that is a cross-sectional shape of lens element Le.

Fig. 11(a) and 11(B) are diagrams schematically showing the state of the illumination light beam in the optical path from the emission end FBo of the fiber bundle FGn shown in fig. 10 to the fly's eye lens system FEn, and the orthogonal coordinate system XYZ is set to be the same as fig. 10. Fig. 11 a is a view of the optical path as viewed from the Y-axis direction (step movement direction), and fig. 11B is a view of the optical path as viewed from the X-axis direction (scanning movement direction). Here, fine circular light emitting points (having a diameter of 0.2mm or less) at the emission ends of the optical fiber lines as the sources of the illumination light beams BSa, BSb, and BSc emitted from the emission end FBo of the fiber bundle FGn shown in fig. 9 are point lights (point light source images) SPa, SPb, and SPc. Further, the numerical apertures of the illumination light fluxes BSa, BSb, and BSc from the spot lights SPa, SPb, and SPc, respectively, are set to be the same. Therefore, the divergence angle of illumination light fluxes BSa, BSb, and BSc parallel to optical axis AX2 from the respective central rays between emission end FBo of optical fiber bundle FGn and first capacitor lens system CFn is equal to angle θ bo in both the X direction and the Y direction.

Illumination light fluxes BSa, BSb, and BSc from the plurality of spot lights SPa, SPb, and SPc formed at the emission end FBo of the optical fiber bundle FGn pass through the 1 st condenser lens system CFn, and are all superposed on the incident surface poi of the fly eye lens system FEn as shown in fig. 11(a) and 11(B), thereby illuminating the incident surface poi with a uniform illuminance distribution. Therefore, the fiber bundle FGn and the 1 st condenser lens system CFn function as the 1 st optical integrator for the fly-eye lens system FEn.

Fig. 12 is a diagram schematically showing an example of the arrangement of a plurality of spot lights (spot light source images) SPa, SPb, SPc formed for each fiber line on the output end FB of the fiber bundle FGn shown in fig. 11, and the XYZ system of the orthogonal coordinate system is set to be the same as that of fig. 11. On the exit end FBo of the rectangle long in the Y direction, point lights (point light source images) SPa indicated by white circles, point lights (point light source images) SPb indicated by black circles, and point lights (point light source images) SPc indicated by double circles are arranged in the X direction and the Y direction in the same number and with the same distribution, respectively. In fig. 12, the 3 spot lights SPa, SPb, and SPc are shown as regularly (periodically) distributed in the XY direction, but are actually densely distributed at random. As described above, when the incident-side

optical fiber bundles

12A, 12B, and 12C include 12 ten thousand optical fiber lines, the spot lights SPa, SPb, and SPc are randomly distributed about 2 ten thousand at the emission end FBo of the optical fiber bundle FGn. As an example, the ratio of the XY-directional dimensions at output end FBo, the ratio of the XY-directional dimensions of 1 lens element Le of fly's eye lens system FEn, and the ratio of the XY-directional dimensions of illumination region IAn are all set to about 1: in case 3, if the outer diameter of the optical fiber is set to 0.2mm, about 143 optical fiber lines are arranged in the X direction of the emitting end FBo and about 420 optical fiber lines are arranged in the Y direction (the total number is 143X 420 about 6 ten thousand). In this case, the X-direction dimension of the emission end FBo is about 28.6mm (0.2mm × 143), and the Y-direction dimension is about 84mm (0.2mm × 420).

Fig. 13 is a diagram showing an arrangement state of a plurality of spot light source images (spot lights SPa ', SPb ', SPc ') formed on the emission surfaces epi of the plurality of lens elements Le constituting the fly-eye lens system FEn shown in fig. 11, and the orthogonal coordinate system XYZ is set to be the same as that of fig. 11 (or fig. 12). In fig. 13, the plurality of spot lights SPa ', SPb', SPc 'formed on the emission surface epi of each lens element Le are re-imaged by the plurality of spot lights SPa, SPb, SPc formed on the emission end FBo of the fiber bundle FGn, and about 6 ten thousand spot lights SPa', SPb ', SPc' are formed on each lens element Le. Therefore, the infinite spot lights SPa ', SPb ', SPc ' are distributed on the entire exit surface epi of the fly-eye lens system FEn to the extent of about 6 ten thousand, which is also the number of lens elements Le.

Fig. 14(a) and 14(B) are diagrams schematically showing the state of the illumination light in the optical path from the fly-eye lens system FEn shown in fig. 10 to the illumination region IAn on the mask substrate M, and the orthogonal coordinate system XYZ is set to be the same as that in fig. 10 (or fig. 11). Fig. 14 a is a view showing an optical path from the eye-return lens system FEn to the illumination area IAn viewed from the X direction (scanning movement direction), and fig. 14B is a view showing an optical path from the eye-return lens system FEn to the illumination area IAn viewed from the Y direction (stepping movement direction). Among the infinite spot lights SPa ', SPb ', SPc ' formed on the emission surface epi of the fly-eye lens system FEn, as shown in fig. 14(a), the illumination light fluxes BSa ', BSb ', BSc ' which travel by diverging from the spot lights SPa ', SPb ', SPc ' located at the distance Δ Hy farthest from the optical axis AX2 in the Y direction become parallel light fluxes by the 2 nd capacitor lens system CPn, and are projected to the entire Y direction of the illumination region IAn in a state where the central light rays (principal rays) thereof are inclined at an angle θ Hy from the optical axis AX 2. Illumination light fluxes BSa ', BSb ', BSc ' that travel while diverging from other point lights SPa ', SPb ', SPc ' arranged in the Y direction on the emission surface epi of the fly's eye lens system FEn are also collimated into parallel light fluxes with respect to the Y direction by the 2 nd capacitor lens system CPn, and are projected (superposed) on the entire Y direction of the illumination region IAn.

Among the infinite spot lights SPa ', SPb ', SPc ' formed on the emission surface epi of the fly-eye lens system FEn, as shown in fig. 14(B), the illumination light beams BSa ', BSb ', BSc ' which are diverged and travel from the spot lights SPa ', SPb ', SPc ' located at the distance Δ Hx farthest from the optical axis AX2 in the X direction become parallel light beams by the 2 nd capacitor lens system CPn, and are projected to the entire X direction of the illumination region IAn in a state where the central light ray (principal ray) thereof is inclined at the angle θ Hx from the optical axis AX 2. Illumination light fluxes BSa ', BSb ', BSc ' that travel while diverging from other point lights SPa ', SPb ', SPc ' arranged in the X direction on the emission surface epi of the fly's eye lens system FEn are also made parallel to the X direction by the 2 nd capacitor lens system CPn, and are projected (superposed) on the entire X direction of the illumination region IAn. Therefore, fly-eye lens system FEn and 2 nd condenser lens system CPn function as the 2 nd optical integrator that irradiates illumination area IAn with illumination light having a uniform illuminance distribution.

The angles θ hy, which are the maximum inclination angles in the Y direction of the illumination light beams BSa ', BSb', and BSc 'irradiated to the illumination region IAn, and the angle θ hx, which is the maximum inclination angle in the X direction, are set to substantially the same value, and the illumination light beams BSa', BSb ', and BSc' have the isotropic divergence angle θ i (θ hy — θ hx) around the principal ray parallel to the axis AX2 and perpendicular to the illumination region IAn. Accordingly, the numerical aperture NAi of the illumination light beams BSa ', BSb ', BSc ' irradiated to the illumination region IAn becomes sin (θ i). As is clear from fig. 14(a) and 14(B), when the radius from the optical axis AX2 of the circular irradiation region of the illumination light fluxes BSa, BSb, BSc irradiated on the incident surface poi of the fly-eye lens system FEn is reduced, the distances Δ Hx, Δ Hy farthest from the optical axis AX2 among the infinite spot lights SPa ', SPb', SPc 'formed on the exit surface epi of the fly-eye lens system FEn are also reduced, and therefore the divergence angle θ i (θ Hy — θ Hx) is also reduced, and as a result, the numerical apertures NAi of the illumination light fluxes BSa', BSb ', BSc' are reduced, and the illumination σ value is also reduced.

[ Functions of magnification-varying

parts

8A, 8B, and 8C ]

Fig. 15 a and 15B are diagrams illustrating a state in which the numerical aperture (divergence angle) of the illumination light beam BMa (BMb, BMc) irradiated onto the incident end FBi of the incident-side optical fiber bundle 12A (12B, 12C) is adjusted by the magnification varying unit (numerical aperture varying unit) 8A (8B, 8C) shown in fig. 4. In fig. 15 a and 15B, the focal position PS2 is a plane on which the light fluxes BMa (BMb, BMc) from the mercury lamps 2A (2B, 2C) passing through the wavelength selection sections 6A (6B, 6C) converge (converge) with the minimum diameter as shown in fig. 4, and a circular light source image LDa caused by the blur image of the arc discharge sections of the mercury lamps 2A (2B, 2C) is formed at the focal position PS 2. The light source image LDa is formed on the incident end FBi of the optical fiber bundle 12A (12B, 12C) on the incident side again as a light source image LDb by the lens system 8a1(8B1, 8C1) and the lens system 8a2(8B2, 8C 2).

When the lens system 8a1(8B1, 8C1) is set to negative refractive power (refractive power), the lens system 8a2(8B2, 8C2) is set to positive refractive power (refractive power), and the lens system 8a1(8B1, 8C1) and the lens system 8a2(8B2, 8C2) are disposed separately as shown in fig. 15 a, re-imaging is performed as follows: the numerical aperture (divergence angle) NA α of the light beam BMa (BMb, BMc) forming the light source image LDb becomes the maximum, and the light source image LDb becomes the minimum diameter on the incident end FBi of the optical fiber bundle 12A (12B, 12C). When the lens system 8a1(8B1, 8C1) and the lens system 8a2(8B2, 8C2) are disposed close to each other as shown in fig. 15B, re-imaging is performed as follows: the numerical aperture (divergence angle) NA β of the light beam BMa (BMb, BMc) forming the light source image LDb becomes minimum, and the light source image LDb becomes maximum in diameter at the incident end FBi of the optical fiber bundle 12A (12B, 12C). By appropriately adjusting the positions of the 2 lens systems 8a1(8B1, 8C1) and 8a2(8B2, 8C2) in the optical axis AX1 direction, the numerical aperture (divergence angle) of the light beam BMa (BMb, BMc) forming the light source image LDb can be adjusted between the maximum NA α and the minimum NA β. In the case of fig. 15 a, the diameter of the light source image LDb may be set to be slightly smaller than the effective diameter of the incident end FBi of the optical fiber bundle 12A (12B, 12C), and in the case of fig. 15B, the diameter of the light source image LDb may be set to be slightly larger than the effective diameter of the incident end FBi of the optical fiber bundle 12A (12B, 12C).

In the incident end FBi of the optical fiber bundle 12A (12B, 12C), the light beams BMa (BMb, BMc) incident from the respective incident ends of the plurality of optical fiber lines existing in the range where the light source image LDb is formed, that is, the range where the light beams BMa (BMb, BMc) are irradiated are emitted as the illumination light beams BSa, BSb, BSc in a state where the numerical apertures (the values in the range from the maximum numerical aperture NA α to the minimum numerical aperture NA β) on the incident side are maintained from the point lights SPa, SPb, SPc of the respective emission ends of the optical fiber lines formed on the emission end FBo of the optical fiber bundle FGn on the emission side, as described in fig. 9 or fig. 11.

When the numerical apertures (divergence angles) of the light fluxes BMa, BMb, and BMc projected on the incident ends FBi of the incident-side

optical fiber bundles

12A, 12B, and 12C are set to be the same by the respective 3

magnification varying units

8A, 8B, and 8C shown in fig. 4, the numerical apertures (corresponding to the angle θ bo shown in fig. 11) of the respective 3 illumination light fluxes BSa, BSb, and BSc emitted from the emission end FBo of the emission-side optical fiber bundle FGn are set to be the same value. However, if the numerical apertures (divergence angles) of the incident light fluxes BMa, BMb, and BMc are made different by the

magnification varying units

8A, 8B, and 8C, the numerical apertures of the 3 illumination light fluxes BSa, BSb, and BSc emitted from the emission end FBo may be made different. This will be explained with reference to fig. 16.

Fig. 16 is a diagram schematically showing the states of light beams BMa, BMb, and BMc incident on the

optical fiber bundles

12A, 12B, and 12C on the incident side of the optical fiber bundle shown in fig. 9, and the states of illumination light beams BSa, BSb, and BSc emitted from the optical fiber bundle FGn (FG1 to FG6) on the emission side. In fig. 16, the numerical aperture (wide angle) of the light beam BMa projected on the incident end FBi of the optical fiber bundle 12A is NAia, the numerical aperture (wide angle) of the light beam BMb projected on the incident end FBi of the optical fiber bundle 12B is NAib, and the numerical aperture (wide angle) of the light beam BMc projected on the incident end FBi of the optical fiber bundle 12C is NAic, and thus NAia > NAib > NAic is set. In this case, the illumination light beam BSa diverged and propagated from each of the plurality of spot lights SPa formed on the emission end FBo of the emission-side optical fiber bundle FG1 becomes the numerical aperture NAia, the illumination light beam BSb diverged and propagated from each of the plurality of spot lights SPb becomes the numerical aperture NAib, and the illumination light beam BSc diverged and propagated from each of the plurality of spot lights SPc becomes the numerical aperture NAic. Similarly, the illumination light beams BSa, BSb, and BSc having different numerical apertures are simultaneously emitted from the emission ends FBo of the other optical fiber bundles FG2 to FG 6.

Fig. 17 is a schematic diagram of the optical path from the emission end FBo to the incident surface poi of the fly-eye lens system FEn viewed from the X direction (scanning movement direction) in order to explain the difference in the irradiation distribution on the incident surface poi of the fly-eye lens system FEn of the illumination light beams BSa, BSb, BSc that travel divergently from the emission end FBo of the fiber bundle FGn, and the XYZ system of orthogonal coordinates is set to be the same as that of fig. 11 (B). In fig. 17, illumination light fluxes BSa, which are diverged and travel from a plurality of spot lights SPa formed at an emission end FBo (corresponding to a pupil plane) of a fiber bundle FGn, are converted into substantially parallel light fluxes by a condenser lens system CPn, and are superimposed and irradiated on a circular region CFa centered on an optical axis AX1 in an incident plane poi of a fly eye lens system FEn. Similarly, the illumination light flux BSb that diverges and travels from each of the plurality of spot lights SPb is converted into a substantially parallel light flux by the condenser lens system CPn, and is superimposed and irradiated on a circular region CFb centered on the optical axis AX1 in the incident surface poi of the fly eye lens system FEn, and the illumination light flux BSc that diverges and travels from each of the plurality of spot lights SPc is converted into a substantially parallel light flux by the condenser lens system CPn, and is superimposed and irradiated on a circular region CFc centered on the optical axis AX1 in the incident surface poi of the fly eye lens system FEn.

Since the emission end FBo of the optical fiber bundle FGn is disposed at the position (pupil plane) of the front focal point of the condenser lens system CPn and the incidence plane poi of the fly-eye lens system FEn is disposed at the position of the rear focal point of the condenser lens system CPn, the illumination light flux BSa from the spot light SPa is irradiated to the whole of the circular region CFa, the illumination light flux BSb from the spot light SPb is irradiated to the whole of the circular region CFb, and the illumination light flux BSc from the spot light SPc is irradiated to the whole of the circular region CFc, regardless of where each of the plurality of spot lights SPa, SPb, and SPc is located on the emission end FBo.

Fig. 18 is a diagram of a state of circular areas CFa, CFb, CFc in fig. 17 distributed on the incident surface poi of the fly-eye lens system FEn viewed in the XY plane, and the XYZ orthogonal coordinate system is set to be the same as that in fig. 17. Since numerical apertures (divergence angles) NAia, NAib, and NAic of the illumination light beams BSa, BSb, and BSc have a relationship of NAia > NAib > NAic, if the radius of the region CFa centered on the optical axis AX2 is Ria, the radius of the region CFb is Rib, and the radius of the region CFc is Rib, as shown in fig. 18, the relationship of Ria > Rib is obtained. Furthermore, 3 illumination light beams BSa, BSb, and BSc are distributed so as to overlap each other in the region CFc having the radius Ric, 2 illumination light beams BSa and BSb are distributed so as to overlap each other in the annular region between the radius Ric and the radius Rib in the region CFb, and only the illumination light beam BSa is distributed in the annular region between the radius Rib and the radius Ria in the region CFa. In fig. 18, a circular area CCA indicated by a dotted line indicates a boundary range where the illumination σ value becomes 1.0(NAi ═ NAp), and the maximum value of the numerical apertures NAia, NAib, and NAic of the illumination beams BSa, BSb, and BSc is set to be equal to or smaller than the numerical aperture corresponding to the radius of the area CCA. Further, the positions separated from the optical axis AX2 in the Y direction and the X direction respectively correspond to the distances Δ Hy and Δ Hx described in fig. 14(a) and 14(B) above, in accordance with the maximum radius Ria among the 3 radii Ria, Rib, and Ric shown in fig. 18.

As described above, by adjusting each of the magnification varying sections (numerical aperture varying sections) 8A, 8B, and 8C shown in fig. 4 and 15, the radii Ria, Rib, and Ric of the respective regions CFa, CFb, and CFc of the 3 illumination light fluxes BSa, BSb, and BSc, which are distributed in a circular shape on the incident surface poi of the fly-eye lens system FEn, can be freely adjusted, so that the infinite spot lights SPa ', SPb ', and SPc ' formed on the exit surface epi of the fly-eye lens system FEn have intensity distributions according to the radial distances from the optical axis AX 2.

Fig. 19 a and 19B show an example of intensity distributions (light source images) of a plurality of spot lights SPa ', SPb ', SPc ' formed on the emission surface epi (illumination pupil plane) of the fly-eye lens system FEn corresponding to the respective areas CFa, CFb, CFc of the illumination light fluxes BSa, BSb, BSc shown in fig. 18. Fig. 19(a) is a view of fly-eye lens system FEn viewed from the X direction (scanning movement direction), and fig. 19(B) is a view of fly-eye lens system FEn viewed from the Y direction (stepping movement direction). The infinite spot light SPa ' formed on the exit surface epi of the fly-eye lens system FEn is generated on a portion of the entrance surface poi corresponding to the circular area CFa (radius Ria) irradiated by the illumination light beam BSa, the infinite spot light SPb ' formed on the exit surface epi is generated on a portion of the entrance surface poi corresponding to the circular area CFb (radius Rib) irradiated by the illumination light beam BSb, and the infinite spot light SPc ' formed on the exit surface epi is generated on a portion of the entrance surface poi corresponding to the circular area CFc (radius Rib) irradiated by the illumination light beam BSc.

When the wavelength characteristics of the 3 illumination light fluxes BSa, BSb, and BSc are the same, for example, when the interference filters attached to the

wavelength selection units

6A, 6B, and 6C shown in fig. 3 and 4 are all the i-ray narrow-band interference filters SWa shown in fig. 6, all the 3 spot lights SPa ', SPb ', and SPc ' having the i-ray (narrow) spectral distribution shown in fig. 6 are formed so as to overlap with each other at a portion corresponding to the circular region CFc having the radius Ric of the exit surface epi (pupil plane of the illumination system) of the fly-eye lens system FEn. Further, 2 point lights SPa ' and SPb ' having a spectral distribution of the i-ray (narrow) are formed in a portion of the exit surface epi (pupil plane of the illumination system) corresponding to the annular band-shaped region from the radius Ric to the radius Rib, and only 1 point light SPa ' having a spectral distribution of the i-ray (narrow) is formed in a portion of the exit surface epi (pupil plane of the illumination system) corresponding to the annular band-shaped region from the radius Rib to the radius Ria. Note that the reason why the plurality of spot lights (spot light source images) SPa ', SPb ', SPc ' formed on the exit surface epi (pupil plane of the illumination system) of the fly-eye lens system FEn shown in fig. 19 a and 19B are not distributed uniformly in the circular area CCA of fig. 18 is that the numerical apertures NAia, NAib, and NAic of the light fluxes BMa, BMb, and BMc incident on the

optical fiber bundles

12A, 12B, and 12C are intentionally set to a relationship of NAia > NAib > NAic in order to explain the functions of the magnification variable sections (numerical aperture variable sections) 8A, 8B, and 8C. In the normal pattern exposure, the numerical apertures NAia, NAib, and NAic of the beams BMa, BMb, and BMc are set to have a relationship of NAia to NAib to NAic.

As shown in fig. 19 a and 19B, when a plurality of spot lights (spot light source images) SPa ', SPb', SPc 'are distributed on an emission plane epi (pupil plane of an illumination system) of the fly eye lens system FEn and wavelength characteristics of light fluxes BMa, BMb, BMc (illumination light fluxes BSa, BSb, BSc) serving as sources of the spot lights SPa', SPb ', SPc' are the same, the illumination light system irradiated in the illumination region IAn of the mask substrate M has a characteristic of illuminance different depending on the numerical aperture as shown in fig. 20. Fig. 20 schematically shows the alignment characteristic (characteristic of the divergence angle) of the illumination light beam Irn irradiated to the point OP on the illumination region IAn, and the principal ray Lpi of the illumination light beam Irn passing through the point OP becomes perpendicular to the surface of the illumination region IAn (the pattern surface of the mask substrate M) due to the telecentric illumination condition (kohler illumination). The illumination light beam Irn is aligned so that the divergence angle θ ia from the principal ray Lpi corresponding to the numerical aperture NAia becomes the maximum numerical aperture. In the divergence angle θ ic corresponding to the numerical aperture NAic in the divergence angle θ ia, the illuminance of illumination beam Irn is the intensity of adding 3 illumination beams BSa, BSb, and BSc, the illuminance of illumination beam Irn is the intensity of adding 2 illumination beams BSa and BSb between the divergence angle θ ib corresponding to the numerical aperture NAib and the divergence angle θ ic, and the illuminance of illumination beam Irn is the intensity of only 1 illumination beam BSa between the divergence angle θ ia and the divergence angle θ ib. That is, the intensity of the divergence angle (θ ic in fig. 20) near the center of the entire divergence angle (θ ia in fig. 20) of illumination light beam Irn is high, and a distribution in which the intensity decreases as the divergence angle increases can be provided.

Note that, since each of the illumination regions IAn (IA1 to IA6) on the mask substrate M and each of the projection regions EAn (EA1 to EA6) on the board P have a conjugate relationship (image formation relationship), the exposure image forming light beam (diffracted light) projected to an arbitrary 1 point in the projection region EAn has the same alignment characteristic (characteristic of a divergence angle) as that in fig. 20.

As described above, by adjusting the numerical apertures (divergence angles) of the light fluxes BMa, BMb, and BMc projected to the incident ends FBi of the incident-side

optical fiber bundles

12A, 12B, and 12C by the magnification varying sections (numerical aperture varying sections) 8A, 8B, and 8C, the numerical aperture (NAia in fig. 20) of the entire illumination light flux Irn in the illumination region IAn projected onto the mask substrate M can be changed to change the illumination σ value, or the illumination distribution can be provided in the range of the divergence angle corresponding to the numerical aperture of the entire illumination light flux Irn. Furthermore, since the diameters of the light beams BMa, BMb, and BMc can be increased or decreased with respect to the diameter of the incident end FBi of the

optical fiber bundles

12A, 12B, and 12C by the magnification varying units (numerical aperture varying units) 8A, 8B, and 8C, the illuminance (the brightness of the spot light SPa, SPb, and SPc) of each of the 3 illumination light beams BSa, BSb, and BSc shown in fig. 9 and 16 can be adjusted.

[ Functions of

wavelength selecting parts

6A, 6B, and 6C ]

In the slide mechanism FX of each of the

wavelength selection sections

6A, 6B, and 6C shown in fig. 3 and 4, for example, any one of 3 kinds of interference filters SWa, SWb, and SWc having wavelength selection characteristics shown in fig. 6 to 8 can be interchangeably attached to perform wavelength selection. In the present embodiment, the wavelength characteristics of the illumination light to be applied to the illumination region IAn of the mask substrate M can be adjusted in accordance with the characteristics (characteristics) of the pattern to be exposed on the mask substrate M by a combination of interference filters attached to each of the 3

wavelength selection portions

6A, 6B, and 6C.

wavelength selection units

6A, 6B, and 6C. In the table of fig. 21, the left column indicates a code designating a combination of the 3 types of interference filters SWa, SWb, and SWc, and the number of o marks in the columns of the wavelength spectrum i-ray (narrow), i-ray (wide), and i-ray + h-ray in the right 3 rows indicates the number of mercury lamps generating the wavelength spectrum. In the following description using fig. 21, each of the 3 magnification-varying

sections

8A, 8B, and 8C is set so that the numerical apertures NAia, NAib, and NAic of the light fluxes BMa, BMb, and BMc projected to the incident end FBi of the incident-side

optical fiber bundles

12A, 12B, and 12C have the same value.

In fig. 21, codes a0, a1, a2, A3, a4, and T of the filter combination are a combination in which the i-ray-narrow-band interference filter SWa is always installed in any of the 3

wavelength selectors

6A, 6B, and 6C, codes B0, B1, and B2 of the filter combination are a combination in which the i-ray-narrow-band interference filter SWa is not installed but the i-ray-wide-band interference filter SWb and the i-ray + h-ray-interference filter SWc are installed in any of the 3

wavelength selectors

6A, 6B, and 6C, and code C0 of the filter combination is a combination in which the i-ray + h-ray-interference filter SWc is installed in all of the 3

wavelength selectors

6A, 6B, and 6C. In the combination, since the i-ray narrow-band interference filter SWa is mounted on all of the 3

wavelength selecting sections

6A, 6B, and 6C, the code a0 is obtained by multiplying the light quantity of the illumination light beam Irn of the illumination region IAn on the mask substrate M by about 3 times the light quantity of the i-ray (narrow) spectral distribution (see fig. 6) obtained by each of the 3

mercury lamps

2A, 2B, and 2C, and can perform high-resolution pattern exposure at relatively high illuminance. In the code a1, when the i-ray/narrow-band interference filter SWa is attached to 2 of the 3

wavelength selection units

6A, 6B, and 6C and the i-ray/wide-band interference filter SWb is attached to the remaining 1, the light quantity of the illumination beam Irn in the illumination region IAn on the mask substrate M is obtained by adding about 2 times the light quantity of the i-ray (narrow) spectral distribution (see fig. 6) from 2 of the 3

mercury lamps

2A, 2B, and 2C and the light quantity of the i-ray (wide) spectral distribution (see fig. 7) from 1 of the 3

mercury lamps

2A, 2B, and 2C, and the light quantity of the illumination beam Irn is increased by only about several% while maintaining the high-resolution pattern exposure performance as compared with the combination of the code a 0.

In fig. 21, the combination code T means that interference filters SWa, SWb, and SWc different from each other are attached to each of the 3

wavelength selectors

6A, 6B, and 6C, the combination code B0 means that only the i-ray-broadband interference filter SWb is attached to all of the 3

wavelength selectors

6A, 6B, and 6C, and the combination code C0 means that only the i-ray + h-ray-broadband interference filter SWc is attached to all of the 3

wavelength selectors

6A, 6B, and 6C. As is clear from the table of fig. 21, in any combination code, the illumination light beam Irn irradiated in the illumination area IAn contains almost 100% of the bright line component of the i-ray from each of the 3

mercury lamps

2A, 2B, and 2C. However, according to the combination of the interference filters SWa, SWb, and SWc, for example, the intensity ratio of the bright line component of the i-ray to the bright line component of the h-ray may be set to be different from the spectral distribution of the original intensity ratio of the mercury lamp (see fig. 5).

Fig. 22 is a graph schematically showing the wavelength characteristics of illumination light beam Irn obtained from combination code B2 in the table of fig. 21. In the combination code B2, an i-ray broadband interference filter SWb is attached to 1 of the 3

wavelength selection units

6A, 6B, and 6C, and an i-ray + h-ray interference filter SWc is attached to each of the remaining 2 wavelength selection units. In this case, the light quantity obtained by multiplying the spectral distribution of i-ray + h-ray shown in fig. 8 by 2 times and the light quantity of the spectral distribution of i-ray (broad) shown in fig. 7 are added to each other to obtain the wavelength spectral distribution of illumination beam Irn. Therefore, in the case of the combination code B2, the light quantity of the bright line component of the i-ray becomes 3 times the light quantity of the 1 mercury lamp, and the light quantity of the bright line component of the h-ray becomes 2 times the light quantity of the 1 mercury lamp, and the light quantity balance between the bright line component of the i-ray and the bright line component of the h-ray in the wavelength spectrum distribution of the illumination light beam Irn can be changed, that is, the spectral characteristic of the illumination light beam Irn can be changed in accordance with the tendency of the spectral characteristic of the light from the mercury lamp (see fig. 5).

In addition to the above embodiment, the following exposure method may be used: when an image of a pattern is projected and exposed on a substrate (plate P) having photo-sensitivity, as in combination codes A1 to A4 and B T, B0 to B2 of FIG. 21, from light including bright wavelengths (e.g., i-rays, h-rays, g-rays) generated by light source devices (mercury discharge lamps 2A, 2B, 2C), the light having a spectral distribution including a specific bright wavelength selected by wavelength selection units (6A, 6B, 6C) is irradiated onto an illumination area IAn (IA1 to IA6) on a mask substrate M carrying a pattern for electronic components by an illumination optical system (FIG. 3), and an image of the pattern is projected and exposed onto the substrate (plate P) by a projection optical system (partial projection optical systems PL1 to PL6) into which an exposure light beam (imaging light beam) generated by the mask substrate M (illumination area IAn) enters, the wavelength selection units Component) and at least 2 of the 2 nd spectrally distributed light (for example, the spectral component selected by the i-ray + h-ray interference filter SWc), and in order to perform kohler illumination on the mask substrate M by the illumination optical system, on the pupil plane (exit plane epi of fly's eye lens system FEn) in the illumination optical system, a1 st light source image (for example, a set of a plurality of spot light source images SPa ' in fig. 13) distributed in a two-dimensional range by the 1 st spectrally distributed light and a2 nd light source image (for example, a set of a plurality of spot light source images SPb ' in fig. 13) distributed in a two-dimensional range by the 2 nd spectrally distributed light are formed so as to overlap each other, as illustrated in fig. 20, in the angle range (incidence angle θ ia in fig. 20) corresponding to the maximum numerical aperture of the illumination light beam Irn irradiated onto the mask substrate M, the balance between the wavelength and the intensity (wavelength intensity characteristics) differs depending on the angle.

[ Cooperation of wavelength selection part and magnification ratio variable part ]

In the above description, the numerical apertures NAia, NAib, and NAic of the light beams BMa, BMb, and BMc projected onto the incident ends FBi of the

optical fiber bundles

12A, 12B, and 12C set by the

magnification varying sections

8A, 8B, and 8C are set to the same value, but the interference filters installed in the

wavelength selecting sections

6A, 6B, and 6C may be different from each other by setting the numerical apertures NAia, NAib, and NAic of the light beams BMa, BMb, and BMc to different values, so that, as shown in fig. 20, a difference in illuminance is given according to the divergence angle (θ ia, θ ib, and θ ic) of the illumination light beam Irn, and a difference in wavelength characteristics can be given. For example, the numerical apertures NAia, NAib, and NAic shown in fig. 20 are set such that NAia is equal to NAib > NAic (the radius Rib shown in fig. 18 and 19 is set such that it is equal to the radius Ria), and as in the combination code a2 shown in fig. 21, the i-ray/narrow-band interference filter SWa is attached to each of the

wavelength selectors

6A and 6B, and the i-ray + h-ray/interference filter SWc is attached to the wavelength selector 6C. In this case, on the emission surface epi of the fly-eye lens system FEn shown in fig. 19, an infinite number of spot lights SPa ' and SPb ' having a spectral distribution of i-rays (narrow) (fig. 6) are arranged in the same manner over the entire circular region CFa having a radius Ria (═ Rib), and an infinite number of spot lights SPc ' having a spectral distribution of i-rays + h-rays (fig. 8) are arranged in the same manner over the circular region CFc having a radius Ric.

Therefore, according to this embodiment, the wavelength characteristics in the distribution range of the 2 nd light source image (the condensed image of the infinite spot lights SPa ', SPb ', SPc ') formed on the exit surface epi of the fly-eye lens system FEn which is the illumination pupil plane of the 2 nd illumination optical system ILn can be changed according to the position in the radial direction with respect to the optical axis AX 2. In this case, among the light rays in the range of the divergence angle θ ic from the principal ray Lpi (numerical aperture NAic) of the illumination light beam Irn illuminating the mask substrate M shown in fig. 20, the spectrum including the bright wavelength of both the i-ray and the h-ray is included, and among the light rays in the annular band-shaped range (numerical aperture NAic to NAia) between the divergence angle θ ic and the divergence angle θ ia (═ θ ib), only the spectrum of the i-ray is included (broad). As described above, if the wavelength characteristics of the 2 nd light source image formed on the illumination pupil plane of the 2 nd illumination optical system ILn are changed in the radial direction, the pattern formed on the mask substrate M can suppress the degradation of the quality of the projected image due to the influence of the pattern manufacturing error or the like in the case of the halftone pattern or the phase shift pattern.

In general, a halftone pattern or a phase shift pattern is produced by forming a displacement layer for controlling a film thickness on a mask substrate so that an amplitude transmittance at a specific wavelength is a predetermined condition on the premise of being used under irradiation with illumination light of the specific wavelength. However, when an error occurs in the film thickness or when the numerical aperture (illumination σ value) of the illumination light is changed, the amplitude transmittance of the displacement layer fluctuates (deteriorates) from a desired condition, and if the contrast of the pattern image of the projection exposure cannot be obtained for a target, the imaging performance cannot be reduced to obtain a target fineness. In the present embodiment, even in the case of using a mask substrate having a halftone pattern or a phase shift pattern as described above, since the wavelength characteristics (spectrum) of the light source image formed two-dimensionally on the illumination pupil plane of the illumination optical system (2 nd illumination optical system ILn) illuminating the mask substrate M are made different in the diameter direction, the deterioration of the imaging performance due to the variation (degradation) of the amplitude transmittance of the displacement layer can be suppressed even in the case where an error occurs in the film thickness of the displacement layer or in the case where the numerical aperture (illumination σ value) of the illumination light is changed.

[ modification 1]

As described above, in embodiment 1, the 3

mercury lamps

2A, 2B, and 2C are ultrahigh-pressure mercury discharge lamps of the same specification, and the bright line wavelength of i-rays and the bright line wavelength of h-rays are mainly used for pattern exposure, but the bright line wavelength of g-rays may be used for pattern exposure. In this case, the optical system is used for a projection optical system which is corrected for chromatic aberration in a wide wavelength range including 3 kinds of bright lines of i-rays, h-rays, and g-rays. As disclosed in japanese patent application laid-open No. 2012-049332, the illumination devices (the 1 st illumination optical system and the 2 nd illumination optical system ILn) of the present embodiment can also be applied to a projection exposure apparatus equipped with a projection optical system of a mirror projection system in which a large concave mirror and a small convex mirror are combined. Since the projection optical system of the mirror projection system does not use a lens element having a strong refractive power, chromatic aberration caused by a difference in wavelength of the illumination light is not substantially generated, and 3 kinds of bright line wavelengths of i-rays, h-rays, and g-rays of the mercury lamp can be easily used. Further, although 3

mercury lamps

2A, 2B, and 2C are designed as the ultra-high pressure mercury discharge lamps of the same specification, a high pressure mercury discharge lamp having different ratios of the peak intensities of the i-ray, the h-ray, and the g-ray from the ratios shown in fig. 5 may be combined with the ultra-high pressure mercury discharge lamp in terms of the wavelength characteristics of the light from the arc discharge section, or a short arc type low pressure mercury discharge lamp and the ultra-high pressure mercury discharge lamp may be combined as appropriate. The number of the 1 st illumination optical systems from the mercury lamp 2 to the optical fiber bundle 12 on the incident side may be 2 or more, and for example, in the case where the number of the partial projection optical systems PLn is 6 or more, 4 mercury lamps 2A to 2D, 41 st illumination optical systems, and 4 optical fiber bundles 12A to 12D on the incident side may be provided in order to secure the illuminance of the illumination light beam Irn.

[ modification 2]

The interference filters attached to the

wavelength selection units

6A, 6B, and 6C shown in fig. 3 and 4 are 3 types of i-ray narrow-band interference filters SWa, i-ray wide-band interference filters SWb, and i-ray + h-ray interference filters SWc having the wavelength characteristics shown in fig. 6 to 8, respectively, but when the projection optical system (part of the projection optical system PLn) can use the wavelength of g-rays, g-ray narrow-band interference filters or g-ray wide-band interference filters, i-ray + h-ray + g-ray interference filters for the ultra wide band, and the filters are attached to the slide mechanism FX. Further, an h-ray-narrow band interference filter or an h-ray-wide band interference filter may be prepared so as to include only the bright line wavelength of the h-ray. In the case of preparing an interference filter including only the bright wavelength of the h-ray, for example, one of the i-ray-narrow band interference filter SWa and the i-ray-wide band interference filter SWb is attached to each of the

wavelength selection units

6A and 6B, and one of the h-ray-narrow band interference filter and the h-ray-wide band interference filter is attached to the wavelength selection unit 6C. The numerical aperture NAia of the light beam BMa (i-ray) incident on the optical fiber 12A and the numerical aperture NAib of the light beam BMb (i-ray) incident on the optical fiber 12B are set to the same value so that the illumination σ value becomes a large value (for example, 0.7 or more) by adjusting each of the

magnification varying sections

8A, 8B, and 8C, and the numerical aperture NAic of the light beam BMc (h-ray) incident on the optical fiber 12C is set so that NAia > NAic is obtained.

In this case, the 2 nd light source image (the integrated image of the infinite spot lights SPa ', SPb ', SPc ') formed on the exit surface epi of the fly-eye lens system FEn which is the illumination pupil plane of the 2 nd illumination optical system ILn includes: an infinite number of spot lights SPa 'and SPb' each including a bright line wavelength of only the i-ray and distributed throughout an area CFa (CFb) having a radius Ria (Rib) corresponding to the numerical aperture NAia (NAib); and an infinite spot light SPc' including only the bright line wavelength of the h-ray and scattered only in the region CFc of radius Ric corresponding to the numerical aperture NAic. Therefore, the 2 nd order light source image formed on the emission surface epi of the fly-eye lens system FEn has the following wavelength distribution characteristics: the spectral components of the i-ray having the bright line wavelength distributed with substantially constant intensity over the entire region CFa (corresponding to the maximum numerical aperture NAia) having the radius Ria include the spectral components of the h-ray having the bright line wavelength distributed only in the region CFc having the radius Ric (< Ria) inside.

When an h-ray-narrow band interference filter or an h-ray-wide band interference filter is prepared, the wavelength distribution characteristics of the 2 nd light source image formed on the illumination pupil plane of the 2 nd illumination optical system ILn may be set to be opposite to the above by adjusting the

magnification varying units

8A, 8B, and 8C, respectively. That is, the entire region CFa (corresponding to the maximum numerical aperture NAia) having a radius of Ria of the 2 nd light source image formed on the illumination pupil plane (emission plane epi) of the 2 nd illumination optical system ILn may be regarded as the spectral component of the bright line wavelength of the h-ray, and only the region CFc having an inner radius of Ric (< Ria) may be regarded as the spectral component of the bright line wavelength of the i-ray. The interference filter may be a band pass filter that extracts spectral components of a predetermined wavelength width, or a low pass filter that transmits wavelength components longer than the cutoff wavelength and a high pass filter that transmits wavelength components shorter than the cutoff wavelength may be arranged in series and attached between the lens systems 6a1 and 6a 2. In this case, a low-pass filter having a cut-off wavelength set to around 350nm to 360nm, a1 st high-pass filter having a cut-off wavelength of about 375nm, and a2 nd high-pass filter having a cut-off wavelength of about 395nm are prepared, and the 1 st high-pass filter and the 2 nd high-pass filter are disposed so as to be interchangeable. Thereby, the spectral component of the i-ray (narrow) shown in fig. 6 is extracted in the combination of the low-pass filter and the 1 st high-pass filter, and the spectral component of the i-ray (wide) shown in fig. 7 is extracted in the combination of the low-pass filter and the 2 nd high-pass filter.

[ modification 3]

In a manufacturing stage of a device such as a substrate of a display panel or a circuit substrate for mounting electronic parts, or a manufacturing stage of a fine metal mask (so-called stencil mask) which is mounted in a deposition apparatus and divides a deposition portion on a substrate to be processed, a negative photoresist layer (photosensitive layer) applied on a plate P with a thickness of about several times to 10 times a normal thickness (0.5 to 1.5 μm) may be pattern-exposed. The negative resist has a smaller sensitivity than the positive resist, and the portion irradiated with the illumination beam Irn for exposure has a property of becoming insoluble in a developer and leaving a film. Further, the negative photoresist may have a large difference in sensitivity or absorption rate with respect to the wavelength of the exposure illumination beam Irn. FIG. 23 is a graph showing an example of the light absorption characteristics of a negative photoresist in which the wavelength (nm) of the illumination light beam Irn is shown on the horizontal axis and the normalized absorptance (0 to 1) is shown on the vertical axis. In the case of the photoresist of fig. 23, there is a peak of absorption at a wavelength of around 320nm, and the absorbance decreases substantially linearly between wavelengths 320nm and 450nm (wavelength dependence of absorbance), and the absorbance at a bright line wavelength of 365nm of i-ray is about 0.5, and the absorbance at a bright line wavelength of 405nm of h-ray is about 0.15. The characteristics of fig. 23 are examples, and are greatly different depending on the material of the resist. When the thickness of the negative resist layer having the characteristics as shown in fig. 23 is 10 μm or more, if pattern exposure is performed with the illumination light beam Irn including both the bright line wavelength of i-rays and the bright line wavelength of h-rays, light of the bright line wavelength of i-rays is absorbed in a large amount in the surface portion of the resist layer according to the wavelength dependency of the absorptance, and a sufficient amount of light is not given to the bottom side (plate P side) of the resist layer. On the other hand, since the absorption of light of the bright wavelength of h-rays is small in the resist layer, a sufficient amount of light is also applied to the bottom side (plate P side) of the resist layer.

Since the resist layer has a large thickness and the wavelength dependency of the absorptance exists, when the i-ray + h-ray interference filter SWc shown in fig. 8 is attached to each of the

wavelength selection units

6A, 6B, and 6C (the combination code C0 in the table of fig. 21 is selected) and the pattern of the mask substrate M is projection-exposed on the board P, the edge portion (sidewall) of the pattern (resist image) of the resist layer of the residual film after development can be inclined so as not to be perpendicular to the surface of the board P. Fig. 24 is a cross-sectional view schematically showing the inclination of the edge portion (sidewall) of the resist image of the residual film after development. In fig. 24, a negative resist layer Luv is formed with a thickness RT (10 μm or more) on the surface of a plate P on which a metal film such as nickel is formed, and after development, unexposed portions (non-irradiated portions) of the resist layer Luv are removed to form an opening HL between edge portions Ewa and Ewb. In the case of manufacturing a fine metal mask, a metal layer (nickel, copper, or the like) is deposited on the plate P exposed in the opening HL by electroplating. The side walls of the edge portions Ewa and Ewb of the resist layer Luv are formed in a state inclined toward the opening HL side, i.e., in a reverse tapered shape.

As described above, in order to control the inclination amount of the side walls of the edge portions Ewa, Ewb to be the resist images to a desired value, the balance between the light amount of the i-ray at the bright line wavelength (corresponding to the area of the shaded portion in fig. 6 or fig. 7) and the light amount of the h-ray at the bright line wavelength included in the illumination light beam Irn can be adjusted by the combination of the interference filters attached to the

wavelength selection portions

6A, 6B, 6C; alternatively, the numerical aperture of the illumination beam having the i-ray bright line wavelength and the numerical aperture of the illumination beam having the h-ray bright line wavelength included in the illumination beam Irn can be independently adjusted by each of the

magnification varying units

8A, 8B, and 8C. The sidewall of the edge portions Ewa and Ewb of the developed resist image is provided with a desired inclination amount, and the inclination amount is not limited to the negative resist, and may be applied to a positive resist.

When the resist layer Luv is used as a mask layer in a plating step during the production of a fine metal mask or the formation of a wiring layer, a resist sold under the trade name PMER P-CS series, PMER P-LA series, PMER P-HA series, PMER P-CE series, naphthoquinone type, chemically amplified PMER P-WE series, PMER P-CY series, or a negative resist sold under the trade name PMER-N-HC600PY, which is commercially available from Tokyo Utility industries, Ltd. Further, a resist for plating sold under the trade names SPR-558C-1 and SPR-530CMT-A by Shanrong chemical Co., Ltd may be used. The resist layer Luv may be formed by using an ultraviolet-curable resin having an appropriate light absorption rate in the wavelength region of the illumination light beam Irn at the time of pattern exposure, and having a composition of an ultraviolet-curable monomer/oligomer (epoxy acrylate, acrylic urethane, or polyester acrylate), a photopolymerization initiator, a photosensitizing agent, an additive, or the like as the photosensitive layer.

[ modification 4]

When only an i-ray-narrow band interference filter or an i-ray-wide band interference filter is used and the exposure illumination beam Irn is light having a bright wavelength including only i-rays, high resolution in pattern exposure can be achieved, but as the resolution increases (the wavelength of the illumination beam decreases), the Depth of Focus (DOF) also decreases. Therefore, in order to suppress the decrease in DOF in the high-resolution state, there is also a case where the shape of the light source image (2 nd order light source image) formed on the illumination pupil plane of the illumination optical system is an annular band shape, and the light source image is a 4-pole shape that is offset at a point-symmetric position (region) centered on the optical axis in the illumination pupil plane. In this case, an aperture plate (illumination aperture stop) having an annular or 4-pole light transmission portion is provided at or near the exit surface epi of the fly-eye lens system FEn.

Fig. 25(a) and 25(B) are diagrams schematically showing shapes in the XY plane of the aperture plate APa having the annular translucent portion formed thereon and the aperture plate APb having the 4-pole translucent portion formed thereon, respectively, and the XYZ orthogonal coordinate system corresponds to fig. 18. The aperture plate APa is formed by etching and removing a light-shielding layer such as chromium deposited on the surface of a parallel flat plate of quartz into an annular shape, and forming an annular light-transmitting portion TPa as shown in fig. 25 (a). Similarly, the light shielding layer on the surface of the quartz parallel plate is removed by etching to have a 4-pole shape, and as shown in fig. 25(B), fan-shaped light transmitting portions TPb are formed in each of 4 quadrants of XY coordinates with the optical axis AX2 as the origin. The aperture plate APb may be a light blocking portion in which light blocking bands extending in the X direction and the Y direction intersect at the position of the optical axis AX2 in a cross shape.

[ modification 5]

The wavelength selection unit 6A (6B, 6C) included in the 1 st illumination optical system shown in fig. 4 is provided with: a lens system (collimator lens) 6a1 that receives the light beam BM that travels while diverging from the position PS1 of the 2 nd focal point of the elliptical mirror 4A (4B, 4C) and converts the light beam BM into a substantially parallel light beam; and a lens system 6a2 that converges the substantially parallel light fluxes at the focal position PS 2. Any of the interference filters SWa, SWb, SWc, etc. is attached to the optical path between the lens systems 6a1, 6a2, but an annular diaphragm plate as shown in fig. 25(a) may be provided in combination. Fig. 26 is a diagram showing a state in which the annular diaphragm plate APa' is disposed on the wavelength selecting section 6A of the 1 st illumination optical system, and the same members as those shown in fig. 4 are denoted by the same reference numerals.

In fig. 26, the annular diaphragm plate APa' is formed by forming the following layers on a flat quartz plate: a bottom-vicinity light-shielding layer which shields the outside of an outer wheel diameter defined in accordance with the maximum diameter of the light beam BM that becomes substantially a parallel light beam by the lens system 6a 1; and a circular central light shielding layer which shields the inside of the inner diameter around the optical axis AX 1. The aperture plate APa' is attached to the slide mechanism FX and is provided to be inserted into and removed from the optical path, similarly to the interference filter SWa (or SWb, SWc, etc.). The illumination light flux BMa transmitted through the annular light transmitting portion TPa of the annular diaphragm APa' converges on the focal position PS2 by the lens system 6a2, and diverges again toward the magnification variable portion 8A at the subsequent stage. The outer diameter of the aperture plate APa 'defines the maximum numerical aperture NAd1 of the illumination beam BMa, and the inner diameter of the aperture plate APa' defines the numerical aperture NAd2 of the hollow range in which the intensity distribution becomes zero in a circle in the cross section of the illumination beam BMa.

The illumination beam BMa having an annular intensity distribution formed by the aperture plate APa' passes through the magnification varying section 8A at the subsequent stage to adjust the entire numerical aperture when entering the incident end FBi of the optical fiber bundle 12A, but the ratio of the maximum numerical aperture NAd1 to the numerical aperture NAd2 of the hollow range is maintained for the beam entering each optical fiber line at the incident end FBi of the optical fiber bundle 12A. As described above with reference to fig. 16, since the optical fiber lines are optically transmitted while the numerical aperture (divergence angle) of the incident light is preserved, the numerical aperture of the light beam BSa emitted from the emission end FBo of the emission-side optical fiber bundle FGn is the same as the numerical aperture of the light beam BMa incident from the incident end FBi of the optical fiber bundle 12A. Therefore, in the case of the present modification, the light beam BSa (divergent light beam from the point light SPa formed on the emission end FBo) emitted from each emission end FBo of the optical fiber bundle FGn has a ring-shaped distribution in which the ratio of the maximum numerical aperture NAd1 to the numerical aperture NAd2 of the hollow range is maintained. As described above with reference to fig. 17, the illumination light fluxes BSa that travel while diverging from the respective spot lights SPa formed at the emission ends FBo of the fiber bundles FGn overlap on the incidence plane poi of the fly's eye lens system FEn, but since the illumination light fluxes BSa themselves have an annular intensity distribution, the illumination light fluxes overlap on the incidence plane poi of the fly's eye lens system FEn with an annular distribution that maintains the ratio (annular band ratio) of the maximum numerical aperture NAd1 to the numerical aperture NAd2 in the hollow range. Similarly, the annular diaphragm plate APa' is provided in the optical paths of the other

wavelength selection portions

6B and 6C in a pluggable manner.

As described above, in the present modification, at least 1 of illumination light fluxes BSa, BSb, and BSc irradiated to be superimposed on incident surface poi of fly's eye lens system FEn can be set to an annular intensity distribution having a desired annular band ratio around optical axis AX 2. Therefore, among the infinite spot lights SPa ', SPb', SPc 'formed on the emission surface epi of the fly-eye lens system FEn by the adjustment of the

magnification varying sections

8A, 8B, 8C, for example, the spot lights SPa' and SPb 'can be distributed in an annular band-shaped range outside the region CFc with the radius Ric shown in fig. 18 or 19, and the spot light SPc' can be distributed in the region CFc with the radius Ric shown in fig. 18 or 19. At this time, by appropriately selecting the combination of the interference filters attached to the

wavelength selection portions

6A, 6B, and 6C, for example, the spot lights SPa ' and SPb ' distributed in the annular band-shaped range outside the region CFc with the radius Ric can have the spectrum of i-rays (narrow), and the spot light SPc ' distributed in the region CFc with the radius Ric can have the spectrum of h-rays (narrow). That is, the wavelength characteristics of the light source image formed on the illumination pupil plane of the 2 nd illumination optical system ILn (the exit plane epi of the fly-eye lens system FEn) can be changed to completely different wavelengths according to the distance from the optical axis AX2 (corresponding to the numerical aperture).

As shown in fig. 26, the annular diaphragm plate APa' is provided in the optical path in the wavelength selection section 6A (6B, 6C), but may be provided in the optical path in the magnification variable section 8A (8B, 8C). Further, instead of the annular diaphragm plate APa 'shown in fig. 26, a diaphragm plate APb' similar to the 4-pole diaphragm plate APb shown in fig. 25(B) may be attached. In this case, illumination light beams BSa (or BSb and BSc) irradiated onto incident surface poi of fly's eye lens system FEn overlap 4 fan-shaped regions as in light transmission portion TPb of fig. 25B. In the present modification, since the illumination light beam BSa (or BSb, BSc) irradiated on the incident surface poi of the fly's eye lens system FEn overlaps a region of a normal circular shape including the optical axis AX2, an annular band shape not including the optical axis AX2, or a 4-pole shape, there is an advantage that loss of the illumination light amount is suppressed to be small, as compared with a case where a part of the 2-time light source images (spot light SPa', SPb ', SPc') is shielded only by the diaphragm plates APa, APb shown in fig. 25 a and 25B.

[ 2 nd embodiment ]

Fig. 27 is a view showing a schematic overall configuration of the exposure apparatus according to embodiment 2, in which the Z-axis of the orthogonal coordinate system XYZ is set to the gravity direction. The detailed configuration of the exposure apparatus shown in fig. 27 is disclosed in, for example, pamphlet of international publication No. 2013/094286 and pamphlet of international publication No. 2014/073535, and therefore the following configuration of the apparatus will be briefly described. In the exposure apparatus of fig. 27, in order to scan the pattern of the exposure mask on a flexible long sheet-like substrate FS, a cylindrical mask DMM that is rotated about a center line CC1 is attached by being curved into a cylindrical surface shape with a constant radius from a center line CC1 set parallel to the Y axis and forming a reflection-type pattern on an outer peripheral surface having a predetermined length (a length corresponding to the width of the sheet-like substrate FS in the Y direction) in the Y direction. Further, the exposure apparatus of fig. 27 is provided with a drum DR having an outer peripheral surface curved in a cylindrical surface shape with a constant radius from a center line CC2 parallel to the Y axis, supporting a sheet substrate FS in close contact with the outer peripheral surface in the longitudinal direction, and rotating about a center line CC 2. Between the cylindrical mask DMM and the drum DR spaced apart in the Z direction, an odd-numbered equal-magnification image forming partial projection optical system PL1 (and the partial projection optical systems PL3 and PL5 … … not shown) having a configuration substantially similar to that shown in fig. 2 and an even-numbered equal-magnification image forming partial projection optical system PL2 (and the partial projection optical systems PL4 and PL6 … … not shown) are provided.

Further, a polarizing beam splitter PBSa for epi-illumination is provided between the outer peripheral surface of the cylindrical mask DMM and each of the odd-numbered partial projection optical systems PL1, PL3, and PL5 … …. An 1/4 wavelength plate (or film) is mounted on the surface of each polarizing beam splitter PBSa on the side of the cylindrical mask DMM. Exposure illumination light beams from odd-numbered 2 nd illumination optical systems IL1, IL3, and IL5 … … having substantially the same configuration as the 2 nd illumination optical system ILn shown in fig. 10 and from even-numbered 2 nd illumination optical systems IL2, IL4, and IL6 … … are projected through polarizing beam splitters PBSa and PBSb, respectively, in each of Y-direction elongated rectangular illumination regions IAn set on the outer peripheral surface of the cylindrical mask DMM. The polarizing beam splitters PBSa and PBSb (and 1/4 wavelength plates) separate the illumination light beam directed toward the illumination region IAn of the cylindrical mask DMM from the reflected light beam from the mask pattern appearing in the illumination region IAn according to the polarization state, but for this purpose, the illumination light beam projected onto the polarizing beam splitters PBSa and PBSb must be linearly polarized in advance. Therefore, a polarizing plate is provided at an appropriate position in the illumination optical path of the 2 nd illumination optical system ILn, for example, a position between the emission-side optical fiber bundle FGn and the 2 nd condenser lens system CPn shown in fig. 10, or in the

wavelength selection units

6A, 6B, and 6C or before and after the

magnification variable units

8A, 8B, and 8C of the 1 st illumination optical system shown in fig. 4.

When a plane parallel to the YZ plane including the center line CC1 of the cylindrical mask DMM and the center line CC2 of the drum DR is defined as the center plane CCp, the set of the odd-numbered partial projection optical systems PL1, PL3, and PL5 … … and the odd-numbered 2 nd illumination optical systems IL1, IL3, and IL5 … …, and the set of the even-numbered partial projection optical systems PL2, PL4, and PL6 … … and the even-numbered 2 nd illumination optical systems IL2, IL4, and IL6 … … are symmetrically arranged with respect to the center plane CCp when viewed in the XZ plane (in the paper plane of fig. 27). The principal ray on the cylindrical mask DMM side of the partial projection optical system PLn, into which the reflected light beam of the pattern generated by each of the illumination regions IAn on the cylindrical mask DMM is incident, is set so that the extension thereof is directed toward the center line CC 1; on the drum DR side of each of the partial projection optical systems PLn, the principal ray of the imaging beam projected on each of the projection regions EAn set on the sheet substrate FS is set so that the extension thereof is directed toward the center line CC 2.

In the present embodiment, since the projection magnification of the partial projection optical system PLn is equal to (1: 1), the radius from the center line CC1 of the outer peripheral surface (pattern forming surface) of the cylindrical mask DMM and the radius from the center line CC2 of the outer peripheral surface of the drum DR (strictly, the radius to which the thickness of the sheet substrate FS is added) are equal to each other, and the cylindrical mask DMM and the drum DR are rotated at the same rotation speed, so that the reflected light beam from the element pattern formed by the high reflection portion and the low reflection portion on the cylindrical mask DMM is scanned and exposed on the sheet substrate FS. At this time, a film body formed of a single layer or a plurality of layers is formed on the pattern formation surface of the cylindrical mask DMM, and the high reflection portion has a reflectance as high as possible and the low reflection portion has a reflectance as low as possible (ideally, a reflectance of zero) with respect to the illumination light beam from the 2 nd illumination optical system ILn. As an example of a method for forming a reflective mask pattern, there is a method comprising: after a1 st film body (a metal thin film or the like) having a high reflectance (for example, 80% or more, preferably 90% or less) in a wavelength spectrum of an exposure illumination beam is deposited on the entire pattern formation surface of the cylindrical mask DMM, a2 nd film body (a metal thin film, a dielectric multilayer film or the like) having a low reflectance (for example, 10% or more, preferably 5% or less) in a wavelength spectrum of an exposure illumination beam is laminated on the surface of the 1 st film body, and patterning by photolithography or the like leaves a portion of the 2 nd film body which is a low reflection portion, and a portion which is a high reflection portion is removed by etching to expose the 1 st film body as a base. In addition, contrary to this method, the following method is also possible: after the 2 nd film body having a low reflectance is first deposited on the entire pattern formation surface of the cylindrical mask DMM, the 1 st film body having a high reflectance is deposited on the surface of the 2 nd film body, and the portion of the 1 st film body that is to be the high reflection portion is left, and the portion that is to be the low reflection portion is removed by etching, so that the 2 nd film body of the base is exposed.

In the case of a reflective pattern, a reflective displacement pattern may be used, which is a phase difference in which the amplitude intensities of the reflected lights generated on the upper and lower surfaces of the step cancel each other out, by forming a fine step corresponding to the wavelength of the illumination light beam on the surface of the reflective film laminated on the pattern forming surface, as in the halftone method or the phase shift method used for the transmissive photomask blank. In this case, a film body having a high reflectance is uniformly formed on the entire pattern forming surface of the cylindrical mask DMM, and a diffraction grating-like or checkerboard-like concave-convex pattern formed of fine steps that give a phase difference of 180 degrees (make the amplitude reflectance zero) to the reflected light is formed on the pattern portion that reduces the reflected light on the surface of the film body. When the step structure is employed to provide a phase difference of 180 degrees or less to the reflected light, the amplitude reflectance is limited to a value other than zero, and thus an intermediate reflectance can be obtained.

In the exposure apparatus using the reflective mask as described above, there is a case where the reflectance of the reflective pattern is uneven with the exchange of the mask (cylindrical mask DMM). In particular, in the reflection type displacement pattern, a reflectance of a pattern portion intended to substantially zero the intensity of reflected light is not sufficiently reduced due to a manufacturing error of a fine step formed on the surface of the film body. In addition, even in the case of a reflection type pattern composed of only a high reflection portion and a low reflection portion, if the pattern is formed on the outer peripheral surface of the cylindrical mask DMM as shown in fig. 27, the pattern is curved in the circumferential direction of the cylindrical mask DMM, and therefore the incident angle of the principal ray of the illumination light beam slightly changes depending on the position in the circumferential direction in the illumination region IAn, and there is a possibility that a difference occurs in the reflectance in the illumination region IAn.

Therefore, in the present embodiment, as described in embodiment 1 or its modified example, by changing the combination of the interference filters attached to the

wavelength selection units

6A, 6B, and 6C, changing the diameters (numerical apertures) of the illumination light beams BSa, BSb, and BSc projected on the incident surface poi of the fly eye lens system FEn by adjusting the

magnification variable units

8A, 8B, and 8C, or changing the regions (shapes) of the illumination light beams BSa, BSb, and BSc projected on the incident surface poi of the fly eye lens system FEn, it is possible to reduce the unevenness of the reflectance due to the manufacturing error of the reflection pattern or the unevenness of the reflectance that may occur due to the curved pattern surface. As described in fig. 20 or 26, in particular, since the intensity distribution or the numerical aperture for each wavelength can be adjusted within the range of the maximum divergence angle (maximum numerical aperture) of the illumination light beam Irn projected onto the illumination region IAn, even when the reflectance is varied or varied in the reflective mask pattern, there is an advantage that the correction can be easily performed.

[ i-ray-broadband interference filter ]

As shown in fig. 6, the normal i-ray interference filter SWa is set to extract (transmit) an i-ray spectrum in a band width as narrow as possible (for example, a width of ± 10nm or less) with the bright line wavelength of the i-ray as the center. On the other hand, the i-ray-broadband interference filter SWb is set to include only the bright wavelength of the i-ray, and extracts (transmits) the i-ray spectrum in a bandwidth as wide as possible. The bandwidth of the i-ray broadband interference filter SWb is set depending on the chromatic aberration characteristics of the partial projection optical system PLn of the catadioptric system (hereinafter, simply referred to as the projection optical system) described in each of the above embodiments. Fig. 28 shows detailed spectral characteristics obtained by measuring the wavelength characteristics of light generated by the arc discharge portion of the ultra-high pressure mercury discharge lamp shown in fig. 5, using a spectrometer having a higher wavelength resolution than the spectrometer measuring the wavelength characteristics of fig. 5. The main bright lines caused by mercury in the ultra-high pressure mercury discharge lamp are g-ray having a wavelength of 435.835nm, h-ray having a wavelength of 404.656nm, i-ray having a wavelength of 365.015nm, and j-ray having a wavelength of 312.566nm, but bright lines Sxw (having a wavelength of about 330nm) are also generated between the bright line wavelength of the i-ray and the bright line wavelength of the j-ray by other substances in the lamp.

On the other hand, in the case of a projection optical system which mainly performs chromatic aberration correction on the bright line wavelength of an i-ray, the projection optical system PLn of the catadioptric system tends to have chromatic aberration characteristics as shown in fig. 29, for example. Fig. 29 is a graph in which the wavelength is taken on the horizontal axis and the chromatic aberration characteristic of the chromatic aberration amount (chromatic aberration of magnification, or chromatic aberration on the axis) is taken on the vertical axis. When chromatic aberration correction is performed on the bright line wavelength of the i-ray, the lens elements constituting the projection optical system are made of 2 or more glass materials having different dispersion or refractive index, and are optically designed so that the amount of chromatic aberration becomes substantially zero at the bright line wavelength of the i-ray. However, the chromatic aberration characteristic is that a large chromatic aberration amount occurs on the long wavelength region side and the short wavelength region side with respect to the bright line wavelength of the i-ray. Therefore, for the chromatic aberration characteristics, the following wavelength width Δ Wi is set: the chromatic aberration is within the allowable tolerance Δ CAi, and does not include any significant bright line wavelength other than the i-ray. As shown in fig. 28, although a bright line Sxw exists near the short wavelength side and an h-ray exists near the long wavelength side of the bright line wavelength of the i-ray, no bright line is evident between the wavelengths 340nm and 400 nm. Therefore, the i-ray broadband interference filter SWb has a characteristic that its transmittance is 90% or more at a wavelength of about 350nm to about 390 nm. That is, the i-ray broadband interference filter SWb has the following wavelength selection characteristics (transmission characteristics): the spectral peak of the bright wavelength of the i-ray and the spectral components of low luminance distributed in the vicinity of the bottom thereof are extracted (transmitted). In the case of using a projection optical system for correcting chromatic aberration for other bright wavelengths (h-ray or g-ray), since the optical system has more or less chromatic aberration characteristics as shown in fig. 29, an h-ray-broadband interference filter or a g-ray-broadband interference filter can be manufactured in the same manner.

[ other modifications ]

As shown in fig. 4 and 15, in order to control the maximum divergence angle (maximum numerical aperture) of each of the illumination light beams BMa, BMb, and BMc projected onto the incident end FBi of each of the

optical fiber bundles

12A, 12B, and 12C on the incident side, the

magnification varying sections

8A, 8B, and 8C of the 2-group lens systems 8A1 and 8A2 having positions in the direction of the optical axis AX1 adjustable are provided, but at least one of the lens systems 8A1 and 8A2 may be replaced with another lens system to fixedly switch the magnification (numerical aperture). Further, 2 conical prism-shaped optical members (axicon optical systems) may be provided between the front group lens system 8a1 and the rear group lens system 8a2 as disclosed in U.S. Pat. No. 5,719,704. In this case, in the case of normal illumination, the 2 conical prism-shaped optical members are closely attached in the direction of the optical axis AX1, and in the case of ring-band illumination, the interval in the direction of the optical axis AX1 of the 2 conical prism-shaped optical members may be adjusted so that the cross-sectional shape of the illumination beam BMa passing between the lens system 8a1 and the lens system 8a2 is in a ring-band shape with a variable size. In this case, it is not necessary to dispose the annular diaphragm plate APa' in the optical path of the illumination light flux as in modification 5 described with reference to fig. 26, and therefore the utilization efficiency of the illumination light in the case of performing the annular illumination can be further improved.

In the above embodiments and modifications, the fly-eye lens system FEn is used as the optical integrator, but a microlens array, a rod integrator, or the like may be used instead. Further, in the above embodiments, the mercury lamps (ultra-high pressure mercury discharge lamps) 2A, 2B, and 2C are used as the light source devices, but any other discharge type lamps may be used. Further, as the light source device, there can be used: a laser light source such as a Light Emitting Diode (LED), a solid-state laser, a gas laser, or a semiconductor laser; or a laser light source that amplifies the seed light laser light and generates a harmonic (ultraviolet wavelength range) of the seed light by the wavelength conversion element.

Further, as a laser light source for generating a harmonic wave of the seed light, for example, as a fiber amplifier laser light source disclosed in Japanese patent laid-open No. 2001-085771, a pulsed laser light having a center wavelength of 355nm can be used as an illumination light beam. In this case, the wavelength conversion element for harmonic generation incorporated in the fiber amplifier laser light source or the like functions as a wavelength selection section (wavelength selection element) similar to the interference filters SWa, SWb, and SWc. As the light source device, a mercury discharge lamp (ultra-high pressure mercury discharge lamp) may be used in combination with a laser light source. For example, light including a spectral component of i-rays (central wavelength of 365nm) extracted by the i-ray-narrow band interference filter SWa or the i-ray-wide band interference filter SWb in the light from the mercury discharge lamp may be used in combination with pulsed laser light having a central wavelength of 355nm emitted from the fiber amplifier laser light source.

In the case of using the laser light source as described above, in order to set the divergence angle (numerical aperture) of the illumination light beam large, for example, a zone plate diffraction grating formed of a glass material such as quartz, which is formed with fine concentric circular (zone plate-shaped) phase-type irregularities having pitches gradually decreasing in the diameter direction, may be arranged on the optical path of the laser beam emitted as a parallel light beam from the laser light source. The minimum pitch of the zone plate diffraction grating is set according to the desired divergence angle (numerical aperture) of each of the illumination beams BMa, BMb, BMc projected onto the respective incident ends FBi of the

fiber optic bundles

12A, 12B, 12C.

In the above embodiments, the exposure apparatus is described by taking a scanning exposure apparatus of a multi-lens system having a plurality of partial projection optical systems PLn as an example, but may be a step-and-repeat type exposure apparatus which exposes the pattern of the mask substrate M and moves the plate P stepwise in sequence while the mask substrate M and the plate P are stationary. The light source of the illumination device is not limited to 3 mercury lamps or 3 laser light sources, and 1, 2, or 4 or more light sources may be provided. In the above embodiment, 6 optical fiber bundles FGn having 6 emission ends FBo are used, but in the case of an exposure apparatus including 12 nd illumination optical system ILn and 1 projection optical system PLn, the number of optical fiber bundles FGn may be 1.

Further, when the illumination light beam BMa produced by 1 light source (e.g., mercury lamp 2A) and 1 st illumination optical system (including the wavelength selecting section 6A and the magnification varying section 8A) is projected onto the mask substrate M by 12 nd illumination optical system IL1 and the pattern of the mask substrate M is projected and exposed on the plate P by 1 partial projection optical system PL1, the illumination light beam BMa from the magnification varying section 8A may be directly incident on the fly's eye lens system FE1 through the 1 st condenser lens system CF1 of the 2 nd illumination optical system IL1 without providing the optical fiber bundles 12A to 12C, FGn.

According to embodiment 1 or its modification or embodiment 2 described above, the spectral distribution of a predetermined wavelength width is extracted from the luminous flux BM emitted from each of at least 21 st light sources and 2 nd light sources (2 of the mercury lamps 2A to 2C) and provided with: a1 st wavelength selection unit and a2 nd wavelength selection unit (2 of 6A to 6C) provided corresponding to each of the 1 st light source and the 2 nd light source; a mechanism (a slide mechanism FX or a mounting mechanism) which disposes wavelength selective elements (interference filters SWa, SWb, SWc, etc.) provided in each of the 1 st wavelength selective section and the 2 nd wavelength selective section and which change spectral distributions such as extracted wavelength regions or wavelength widths in an optical path in an exchangeable manner; and a light combining means (fiber bundle FGn) for combining the 1 st illumination light beam extracted by the 1 st wavelength selecting unit and the 2 nd illumination light beam extracted by the 2 nd wavelength selecting unit into a2 nd light source image on an illumination pupil plane of an illumination optical system including an optical integrator in a state of numerical apertures set by the numerical aperture varying units (2 of 8A to 8C), respectively. Therefore, different wavelength distribution characteristics (characteristics in which the intensity of each spectrum differs depending on the position in the illumination pupil plane) can be given to the 2-fold light source image distribution or the wavelength distribution can be made different at a divergence angle corresponding to the maximum numerical aperture of the illumination light beam to the mask substrate, depending on various conditions (exposure recipes) such as the difference in the type of pattern on the mask substrate (binary mask, phase shift mask, halftone mask, and the like), the fineness of the pattern to be exposed, the amount of taper tilt given to the edge portion of the resist layer after development, or the variation or unevenness in the reflectance in the case of a reflective mask pattern. Further, by changing (switching) the wavelength distribution on the illumination pupil plane, it is also possible to control (suppress) the irradiation fluctuation (projection magnification fluctuation, focus fluctuation, aberration fluctuation, and the like) of the projection optical system itself, which is generated by the energy of the image forming light beam passing through the projection optical system (partial projection optical system PLn) that performs projection exposure of the pattern of the mask substrate.

Hereinafter, referring to fig. 30, the wavelength characteristics (spectral characteristics) of light from the mercury discharge lamp are supplemented. In each embodiment or modification, short arc type ultrahigh pressure mercury discharge lamps (2A, 2B, 2C) are mainly used, but as a light source of a pattern exposure apparatus for electronic devices, a discharge tube (arc tube) in which the vapor pressure of mercury is about 10 is also used⁵Pa～10⁶And a high-pressure mercury discharge lamp of about Pa. In general, an ultrahigh-pressure mercury discharge lamp increases the mercury vapor pressure in the discharge tube to about 10⁶Pa-number of 10⁷Pa, and the spectral widths of i-rays, h-rays, and g-rays having bright wavelengths suitable for photolithography are slightly broader than those of high-pressure mercury discharge lamps, or the relative balances of the peak intensities of i-rays, h-rays, and g-rays are different from those of high-pressure mercury discharge lamps. In the discharge tube of a mercury discharge lamp, for example, as disclosed in Japanese patent laid-open publication No. 2009-193768, mercury (0.15 mg/mm) having a mercury vapor pressure of 150 to 300 atmospheres at the time of lighting is excluded³Above), a halogen such as iodine, bromine, or chlorine is enclosed in the form of a compound of argon (rare gas) and mercury or other metals at about 13 kPa. Further, in the light from the ultra-high pressure mercury discharge lamp, in the wavelength band between the i-ray, g-ray, and h-ray of each bright line wavelength, there is a spectral distribution (near bottom portion) of low luminance of about 10 to 20% or so relative to the peak of the light intensity of the bright line wavelength, as compared with the high pressure mercury discharge lamp.

Fig. 30 is a graph illustrating differences in wavelength characteristics between a high-pressure mercury discharge lamp and an ultrahigh-pressure mercury discharge lamp, where fig. 30(a) shows an example of wavelength characteristics of light from the high-pressure mercury discharge lamp, and fig. 30(B) shows an example of wavelength characteristics of light from the ultrahigh-pressure mercury discharge lamp. In each of fig. 30(a) and 30(B), the horizontal axis represents the wavelength (nm) and the vertical axis represents the relative intensity (%) of the spectrum when the peak of the intensity of i-rays having a bright line wavelength is 100%. Although the wavelength characteristics of fig. 30(a) and (B) vary somewhat depending on the difference between lamp manufacturers and the difference between rated powers of the lamps, when focusing on the spectral distribution of wavelengths 350 to 400nm including the wavelength (365nm) of i-rays, the high-pressure mercury discharge lamp does not substantially have a bottom portion near the portion where the relative intensity is several% or more, preferably 10% or more, as shown in fig. 30 (a). On the other hand, as shown in fig. 30B, the ultrahigh-pressure mercury discharge lamp has a bottom portion (spectral component of low luminance) having a relative intensity of several% or more, approximately 10% or so.

The degree of the relative intensity of the portion near the bottom of the i-ray in fig. 30(B) can be changed by the amount of mercury sealed in the discharge tube, the type or content of other rare gas or halogen, and the mercury vapor pressure, and is about several% to 20%. Further, the spectral widths of i-rays, h-rays, and g-rays at the bright wavelengths tend to be slightly expanded (thicker) in the ultra-high pressure mercury discharge lamp of fig. 30(B) than in the high pressure mercury discharge lamp of fig. 30 (a). In the wavelength characteristics shown in fig. 5, the spectral component in the range of 350 to 400nm in the vicinity of the bottom of the wavelength (365nm) of the i-ray is an intensity of about 20% relative to the peak intensity of the i-ray.

Therefore, the amount of light energy of the illumination beams (BSa, BSb, BSc) whose wavelengths are selected using the i-ray-broadband interference filter SWb shown in fig. 7 is increased as compared with the illumination beams whose wavelengths are selected using the i-ray-narrow band interference filter SWa shown in fig. 6 so as to include only i-rays in the bright-line wavelengths of mercury. Even if the relative intensity of the near-bottom portion (the range of 350nm to 365nm and 365nm to 400 nm) of the i-ray with respect to the peak intensity of the i-ray is about 10% as shown in fig. 30(B), the amount of light energy per unit time (dose) supplied to the resist layer of the plate P is increased by only a quantity determined by the product of the intensity of the near-bottom portion and the wavelength width, and therefore the exposure amount is increased by about 20% (1.1 × 1.1 ≈ 1.2 times). Therefore, when the pattern of the mask substrate M is scanned and exposed on the resist layer of the plate P by the exposure apparatus EX shown in fig. 1, the scanning speed of the mask substrate M and the plate P can be increased by about 20%, and as a result, the productivity of the step of high-resolution pattern exposure by i-ray can be improved by about 20%.

Therefore, when the wavelength width at 10% of the relative intensity of the spectral distribution of i-rays emitted from the high-pressure mercury discharge lamp shown in FIG. 30A or the wavelength selection range (bandwidth) set in the i-ray-narrow band interference filter SWa shown in FIG. 6 is BWi (nm), the outside (short-wavelength side and long-wavelength side) of the wavelength width (bandwidth) BWi set so that the wavelength reaching the peak intensity is centered in the spectral distribution of i-rays emitted from the ultra-high-pressure mercury discharge lamp, and the spectrum reaching the next bright-line wavelength is defined as the bottom vicinity portion, an ultra-high pressure mercury discharge lamp can be used in which the amount of mercury and the mercury vapor pressure, the gas pressure or component amount of rare gas, the component amount of halogen, and the like are adjusted so that the average relative intensity in the vicinity of the bottom portion becomes about several% to 10% (preferably 20% or more). In the case of using the spectral distribution of h-rays or g-rays from the ultrahigh-pressure mercury discharge lamp, the h-ray-broadband interference filter or the g-ray-broadband interference filter having wavelength selection characteristics including the bottom-near portion may be prepared, because the bottom-near portion has a relative intensity that is a few% greater than that of the h-ray or g-ray from the high-pressure mercury discharge lamp.

Then, the modified example of the mask is supplemented. In the above embodiments and the modifications thereof, it is assumed that the exposure apparatus of the transmissive or reflective mask substrate (or cylindrical mask) on which the mask pattern is fixedly formed (carried) is used, but the illumination system (fig. 3 to 20 and the like) described in the embodiments can be applied to the exposure apparatus of the variable mask system (also referred to as a mask-less exposure apparatus because a fixed mask pattern is not used) in which the pattern image is projected on the plate P by switching the angle of each Mirror at high speed based on the data (CAD data) of the pattern to be exposed by using a dmd (digital Mirror device) or the like in which a plurality of micron-sized micro mirrors are two-dimensionally arranged. In the exposure apparatus of the variable mask system, since a projection area formed on the plate P by 1 DMD is limited to a small rectangular area, similarly to the projection area EA1 shown in fig. 1, a plurality of DMDs and a plurality of projection lens systems for projecting the light reflected from each DMD on the plate P are provided. In this case, the respective reflection surfaces (surfaces on which a plurality of micro mirrors are arranged) of the plurality of DMDs carry the pattern for the electronic element in the form of the distribution of the plurality of micro mirrors that individually control the reflection direction of light based on the CAD data. The reflective surfaces of the DMDs are arranged at positions corresponding to the illumination areas IA1 to IA6 shown in fig. 1 to 3, and are irradiated with illumination light beams (corresponding to BSa ', BSb ', and BSc ' shown in fig. 14) having uniform intensity distribution within ± 2%, for example.

Therefore, the projection light beam (illumination light beam) projected on the plate P by the projection lens system can be reflected by the minute mirror whose angle is set so that the illumination light beam enters the projection lens system, among the minute mirrors of the DMD, and has the same characteristics as the alignment characteristics (characteristics of the divergence angle) shown in fig. 20. Further, the wavelength distribution may be made different within a divergence angle corresponding to the maximum numerical aperture of the projection light flux irradiated from each of the micro mirrors of the DMD onto the plate P by a combination of interference filters shown in fig. 21. In addition, instead of the DMD, a Spatial Light Modulator (SLM) may be used, which gives a phase difference to a reflected Light beam by displacing a selected micro mirror among respective reflection surfaces (all of which are usually set on the same plane) of a plurality of micro mirrors arranged two-dimensionally in a direction perpendicular to the reflection surfaces.

Next, the other modes of the projection exposure apparatus are supplemented. In the above embodiment or modification, the so-called multi-lens type exposure apparatus having the plurality of partial projection optical systems PLn (PL1 to PL6) and the plurality of 2 nd illumination optical systems ILn (IL1 to IL6) corresponding thereto is assumed, but even if an exposure apparatus having a single projection optical system and a single 2 nd illumination optical system is assumed, the same function can be easily obtained by only slightly changing the configuration in the above embodiment. Specifically, in the line distributing section 10a in the light distributing section 10 shown in fig. 9, a plurality of optical fibers included in each of the incident-side

optical fiber bundles

12A, 12B, and 12C may be collected into a single optical fiber bundle without being distributed to each of the 6 optical fiber bundles FG1 to FG6, and may be molded so as to have a rectangular shape similar to the shape of a single illumination area in which the emission end FBo of the single optical fiber bundle is set on the mask substrate M.

Next, a modification of the light source device is supplemented. In the configuration of fig. 3, a plurality of (2 or more)

mercury discharge lamps

2A, 2B, and 2C are used as the light source device, but when a single ultra-high pressure mercury discharge lamp is used, the configuration of the above embodiment may be slightly changed to easily have the same function. Specifically, a dichroic mirror is provided which transmits light in a wavelength band including the spectral components of h-rays and g-rays, but not including the spectral components of i-rays, and reflects light in a short wavelength band including the spectral components of i-rays, after the light from the single ultra-high pressure mercury discharge lamp is converted into substantially parallel light by the lens system (collimator lens) 6a1 in fig. 4. Further, as shown in fig. 4 (or fig. 3), a wavelength selection unit 6A (including an interference filter for extracting a spectral component of an h-ray, for example) and a magnification variable unit 8A are provided for light transmitted through the dichroic mirror, and a wavelength selection unit 6B (including an i-ray-narrow band interference filter SWa or an i-ray-wide band interference filter SWb) and a magnification variable unit 8B are provided for light reflected by the dichroic mirror. In this way, as in the description of fig. 22, the wavelength characteristics of the light source image (collection of point light source images) formed with a two-dimensional spread (range) on the illumination pupil plane (corresponding to the exit plane epi of the fly-eye lens system FEn) in the 2 nd illumination optical system ILn can be made variable according to the characteristics of the interference filter selected and set.

Next, the setting of the wavelength selection characteristic by the i-ray-wide band interference filter SWb is complemented. As shown in fig. 30B, the width of the near-bottom portion of the i-ray spectral component from the ultra-high pressure mercury discharge lamp (for example, a range of about 10% intensity with respect to the peak intensity at the center wavelength of the i-ray) is more than 2 times wider than the width of the near-bottom portion of the i-ray spectral component from the high pressure mercury discharge lamp shown in fig. 30 a. As shown in fig. 1 and 2, the partial projection optical system PL1(PL2 to PL6) of the exposure apparatus EX is a half field of view type image forming system of a catadioptric system in which mirrors Ga4 and Gb4 are arranged on pupil planes (stop positions) epo and Epb. Such an imaging system has an advantage that chromatic aberration correction is easy as compared with an imaging system of a total refraction system (all optical elements are constituted by only refractive elements such as lenses), and deterioration (image distortion) of a projected image due to chromatic aberration can be reduced even when a pattern of the mask M is projection-exposed on the substrate P using illumination light including a plurality of bright-line spectra (for example, i-ray spectral components and h-ray spectral components).

However, even if projection exposure is performed using only illumination light of a single i-ray spectral component that is narrowed in band by the i-ray-narrow-band interference filter SWa, if the pattern formed on the mask M becomes fine, the projected image (image intensity distribution) is distorted due to various aberrations possessed by the projection optical system PL1(PL2 to PL 6). This distortion, which appears clearly, is a pattern of fine isolated rectangles (roughly squares) called an aperture pattern.

Fig. 31 is a diagram schematically showing a relationship between a square hole pattern formed on the mask M and the shape of a projection image (light intensity distribution) obtained when the hole pattern is projected onto the substrate P using the i-ray narrow-band interference filter SWa or the shape of a resist image appearing by development of the resist layer after exposure, and the X-axis and the Y-axis correspond to the orthogonal coordinate system XYZ in fig. 1 to 3. Here, fig. 31(a) schematically shows the shape of a projected image (resist image) Ima obtained when a hole pattern CHA formed on a mask M is formed with a dimension Dx × Dy sufficiently larger than the minimum line width value analyzed by the projection optical system PL1(PL2 to PL 6); fig. 31B schematically shows the shape of a projected image (resist image) Imb obtained when a hole pattern CHB is formed in the mask M at a size of approximately 2 times the minimum line width value; fig. 31C schematically shows the shape of a projected image (resist image) Imc obtained when the hole pattern CHC is formed on the mask M in a size close to the minimum line width value. In fig. 31, the hole patterns CHA, CHB, and CHC are formed as isolated transparent portions in the peripheral light-shielding portions shown by hatching, but the opposite may be true, that is, they may be formed as isolated light-shielding portions in the peripheral transparent portions. The resolving power R, which is a minimum line width value that can be projected by the projection optical system PLn (n is 1 to 6), is generally defined by R ═ k · (λ/NAp) through the numerical aperture NAp on the image side of the projection optical system PLn, the wavelength λ (nm) of the illumination light, and the program constant k (0 < k ≦ 1).

As shown in fig. 31(a), when the minimum line width value that can be projected by the projection optical system PLn is a large square hole pattern CHA having a size Dx × Dy of several times or more, the right-angled corners of the four corners are not sufficiently resolved into circular arcs mainly by the value of the numerical aperture NAp on the image side of the projection optical system PLn, i.e., the mtf (modulation Transfer function). Such a phenomenon occurs also when an illumination light (for example, a laser having a spectral width of less than 1 nm) having an extremely narrow spread of the spectral distribution of the central wavelength λ is used. In particular, as shown in fig. 31C, in the case of a square hole pattern CHC having a size close to the minimum resolvable line width value of the projection optical system PLn, the projection image (resist image) Imc has a substantially circular shape. In the case of projection exposure of the hole pattern CHC shown in fig. 31(C) using illumination light extracted by the i-ray-wide band interference filter SWb shown in fig. 7 so as to widely include the bottom vicinity of the spectral distribution with respect to the center wavelength of the i-ray in the characteristics of the projection optical system PLn, the projected image (resist image) Imc changes from a circular shape to an elliptical shape due to the chromatic aberration characteristics of the projection optical system PLn.

Fig. 32 is an enlarged view showing the state of the projected image Imc distorted into an elliptical shape as described above, and the broken line in fig. 32 shows the projected image Imc' which is a substantially correct circular shape. The diameter of the circle of the projection image Imc' can also be theoretically estimated based on the basic optical characteristics of the projection optical system PLn. When the minor axis length in the Y axis direction of the projected image Imc distorted into an elliptical shape by the influence of chromatic aberration is CHy, the major axis length in the X axis direction is CHx, and the ellipticity (ellipticity) Δ f of the elliptical shape is Δ f which is CHy/CHx, the ellipticity (ellipticity) Δ f is preferably 80% or more, and more preferably 90% or more, in terms of the allowable range in device manufacturing. That is, the extended range of the bottom vicinity portion of the spectral distribution of the i-ray extracted by the i-ray-wide band interference filter SWb is defined as: the distortion of the shape from the circle of the projected image Imc of the square hole pattern CHC having a size close to the minimum resolvable line width value is limited to an ellipse having a flat rate (ellipticity) of 80% or more, preferably 90% or more.

In fig. 32, in the deformation of the projected image Imc into an ellipse, the major axis direction is represented as the X direction, the minor axis direction is represented as the Y direction, and each direction of the major axis and the minor axis may be directed in an arbitrary direction within the XY plane as shown in fig. 33. In fig. 33, the projection image Imc of the hole pattern deformed into an elliptical shape has its major and minor axes rotated by Δ ρ with respect to the X and Y axes. Therefore, in order to accurately determine whether the extended range of the bottom vicinity portion of the spectral distribution of the i-ray extracted by the i-ray-broadband interference filter SWb is appropriate, a projected image Imc of a square hole pattern CHC having a size close to the minimum resolvable line width value is exposed on the substrate P by test exposure or the like, a resist image corresponding to the projected image Imc after development is observed by an inspection device or the like, and the shape of the resist image corresponding to the projected image Imc is specified (determination of the directions of the major axis and the minor axis) by image software analysis, thereby measuring the major axis length CHx in the major axis direction and the minor axis length CHy in the minor axis direction. Then, it is sufficient to determine whether or not the ellipticity Δ f obtained from the measurement result is within an allowable range (80% or more, preferably 90% or more).

However, as shown in fig. 2, the image shift optical member SC1 is provided in the image space (the nearest position below the field stop FA1 disposed on the intermediate image plane IM 1) in the imaging optical path of the projection optical system PLn (n is 1 to 6). The image shift optical member SC1 is composed of a transparent parallel plate glass (quartz plate) tiltable in the XZ plane in fig. 2 and a transparent parallel plate glass (quartz plate) tiltable in a direction orthogonal thereto, as disclosed in, for example, wo 2013/094286 pamphlet. By adjusting the inclination amounts of the 2 quartz plates, the pattern image in the projection area EA1(EA2 to EA6) projected on the substrate P can be slightly displaced in any direction in the XY plane. The arrangement of the image shift optical member SC1 is not limited to the closest position below the field stop FA1 shown in fig. 2, and may be replaced with the arrangement of one of the focus adjustment optical member FC1 and the magnification adjustment optical member MC1, which are other correction optical systems, arranged in the image space.

Each of the 2 parallel flat plate-shaped quartz plates constituting the image shift optical member SC1 has a high transmittance from an ultraviolet wavelength region (about 190 nm) to a visible wavelength region, but in the case of synthetic quartz, as shown in fig. 34, for example, the refractive index tends to vary greatly depending on the wavelength from a short wavelength region having a wavelength of 500nm or less, particularly from the vicinity of the wavelength of 400 to 3000nm to the short wavelength side. In fig. 34, the horizontal axis represents wavelength (nm) and the vertical axis represents the refractive index of synthetic quartz. Therefore, for example, when the pattern of the mask M is projected by using illumination light including both an i-ray spectral component having a center wavelength of about 365nm and an h-ray spectral component having a center wavelength of about 405nm in light from the ultra-high pressure mercury discharge lamp (or the high pressure mercury discharge lamp), a phenomenon occurs in which an image projected on the substrate P by the i-ray spectral component and an image projected on the substrate P by the h-ray spectral component are slightly shifted in position in the XY plane according to the tilt amount of the quartz plate of the image shift optical member SC 1.

Fig. 35 is a view schematically showing the operation of the image beam on the quartz plate SCx for displacing the image in the X direction among the 2 quartz plates constituting the image shift optical member SC1 disposed below the field stop plate FA 1. The quartz plate SCx is configured such that an incident surface Stp on which the image beam is incident and an exit surface Sbp from which the image beam is emitted, which pass through the opening of the field stop plate FA1, are opposed to each other in parallel with a gap (thickness) Dpx therebetween, and is provided so as to be rotatable (tiltable) about a rotation center line parallel to the Y axis. Fig. 35 shows only the principal ray LPr of the imaging light beam that diverges from the image point Poc imaged as an intermediate image at the center point of the aperture of the field stop FA1 and travels, the line Lss indicates the normal line of the incident surface Stp at the point where the principal ray LPr intersects the incident surface Stp, and the line LPr' indicates the extension line of the principal ray LPr before entering the incident surface Stp. In a state of an initial posture in which the incident surface Stp of the quartz plate SCx is orthogonal to the incident principal ray LPr (a state in which the inclination of the quartz plate SCx is zero), when the quartz plate SCx is inclined at an angle Δ θ X in the XZ plane, the principal ray LPr is emitted from the emission surface Sbp in the X direction by a displacement amount X in parallel with respect to the extension line LPr'.

In general, the displacement x of the light beam due to the inclination of the parallel plate glass having the refractive index nx can be calculated as x ≈ Dpx · Δ θ x (1-1/nx) by applying the scheimpflug's law, but when the imaging beam includes an i-ray spectral component (wavelength 365nm) and an h-ray spectral component (wavelength 405nm), the refractive index of the quartz plate SCx shows slightly different values with respect to the respective wavelengths. Therefore, when the refractive index of the i-ray spectral component (wavelength 365nm) of the quartz plate SCx is ni, the refractive index of the h-ray spectral component (wavelength 405nm) is nh, the amount of displacement of the image by the i-ray spectral component (wavelength 365nm) is xi, and the amount of displacement of the image by the h-ray spectral component (wavelength 405nm) is xh, the amount of displacement xi is calculated by xi is approximately equal to dppx · Δ θ x (1-1/ni), and the amount of displacement xh is calculated by xh is approximately equal to dppx · Δ θ x (1-1/nh). Therefore, when the difference in the amount of displacement caused by the difference in wavelength (color shift) is x (i-h), the difference x (i-h) is defined as x (i-h)

x(i－h)≒Dpx·Δθx[(1－1/ni)－(1－1/nh)]。

For example, when the change of the difference amount x (i-h) with respect to the angle Δ θ x (0 ° to 10 °) is obtained by setting the thickness Dpx of the quartz plate SCx to 10mm, the refractive index ni of the i-ray spectral component (wavelength 365nm) to 1.4746, and the refractive index nh of the h-ray spectral component (wavelength 405nm) to 1.4696, the linear characteristic is as shown in the graph of fig. 36. In the graph of fig. 36, the horizontal axis represents the inclination angle Δ θ x [ deg. ]ofthe quartz plate SCx, and the vertical axis represents the difference amount x (i-h) [ μm ]. Since the quartz plate SCx is disposed in the image space where the intermediate image is formed and the projection magnification of the projection optical system PLn is equal (x 1), the difference x (i-h) in fig. 36 directly becomes the relative positional displacement between the pattern image due to the i-ray spectral component and the pattern image due to the h-ray spectral component, which are projected onto the substrate P. For example, when the inclination angle Δ θ x of the quartz plate SCx is 5 °, the difference x (i-h) due to color shift becomes about 2 μm in the x direction, and the pattern subjected to projection exposure is distorted or the line width is varied.

The above-described influence of the color shift by the quartz plate SCx is similarly generated in the other quartz plate (SCy) that slightly shifts the projection image in the Y direction, and when the quartz plate SCy is inclined at the inclination angle Δ θ from the horizontal initial state around the rotation center line parallel to the X axis, the difference Y (i-h) due to the color shift occurs in the Y direction. As described above, when the illuminance is to be improved by the illumination light including both the i-ray spectral component (wavelength 365nm) and the h-ray spectral component (wavelength 405nm), the range of image shift by the image shift optical member SC1, which is necessary to maintain good continuity accuracy between pattern images projected on the substrate P by each of the projection optical systems PLn (n is 1 to 6), may be limited depending on the degree of the difference amounts x (i-h) and y (i-h) due to color shift.

On the other hand, as in the above-described embodiments and modifications, the use of the i-ray-broadband interference filter SWb can extract the vicinity of the bottom of the i-ray spectral components from the ultra-high pressure mercury discharge lamp as shown in fig. 7 widely in a range not including the bright line component (h-ray) on the side of the long wavelength side or the bright line component on the side of the short wavelength side, and can suppress the color shift error [ the error corresponding to the difference amounts x (i-h) and y (i-h) ] generated by the inclination of the quartz plates SCx and SCy of the image shift optical member SC1 to be small while increasing the illuminance by about several% to several tens% with respect to the illuminance when the i-ray-broadband interference filter SWa is used.

The ellipticity (ellipticity) Δ f when the projected image Imc of the hole pattern CHC described in fig. 32 and 33 is distorted into an elliptical shape, or the directivity in the XY plane of the major axis/minor axis also changes depending on the degree of the inclination angle of the quartz plates SCx and SCy of the image shift optical member SC 1. Therefore, as in the case of the projection optical system PLn shown in fig. 2, when the image shift optical means SC1 using tiltable parallel flat glasses (quartz plates SCx, SCy) is provided, the wavelength selection range of the i-ray-broadband interference filter SWb may be set so that the ellipticity Δ f of the projection image Imc of the hole pattern CHC close to the minimum resolvable line width value (resolution R) becomes 80% or more (preferably 90% or more) theoretically or actually in the maximum inclination angles Δ θ x, Δ θ y of the parallel flat glasses (quartz plates SCx, SCy) corresponding to the maximum nominal image shift amount by the image shift optical means SC 1.

As described above, in the embodiments described with reference to fig. 31 to 36, in the projection exposure method for projecting and exposing the image of the mask pattern onto the substrate by the projection optical system for illuminating the mask pattern (transmissive or reflective) with the illumination light having a predetermined wavelength distribution (for example, the light from the ultra-high pressure mercury discharge lamp) and projecting the image of the mask pattern onto the substrate by injecting the imaging light beam generated by the mask pattern, the projection image of the square or rectangular hole pattern having a size close to the minimum resolvable dimension determined by the resolution R defined by k · (λ/NAp) is projected onto the substrate with the specific center wavelength in the wavelength distribution of the illumination light being λ (for example, the center wavelength of i-ray), the numerical aperture on the substrate side of the projection optical system being NAp, and the program constant being k (0 < k ≦ 1), by setting the width of the wavelength distribution of the illumination light including the center wavelength λ (for example, the wavelength width selected by the interference filter) such that the ratio (CHy/CHx) of the minor axis length (CHy) to the major axis length (CHx) of the projection image of the hole pattern deformed into an elliptical shape becomes 80% (0.8) or more, and preferably 90% (0.9) or more, it is possible to perform high-resolution pattern exposure while increasing the illuminance of the illumination light irradiated to the mask pattern. The size of the hole pattern is set to a size that is larger than a size determined by the analysis force R and smaller than a size 2 times the analysis force R, in terms of the size of an image projected on the substrate side.

In other words, from another viewpoint, an interference filter for filtering light from a light source (such as a mercury discharge lamp) that emits light including a plurality of bright lines into illumination light having a wavelength width suitable for projection exposure of a mask pattern is incorporated in an illumination system of an exposure apparatus, the wavelength width of the interference filter being set as follows: when a projection image of a square or rectangular hole pattern having a size close to a minimum resolvable line width dimension determined by a resolving power R defined by k · (λ/NAp) is projected onto a substrate, a ratio CHy/CHx of a minor axis length CHy of the projection image of the hole pattern deformed into an elliptical shape to a major axis length CHx is 80% (0.8) or more, preferably 90% (0.9) or more.

Claims

1. An exposure apparatus for projection-exposing a pattern of a mask onto a photosensitive substrate, comprising:

a light source that generates light including a plurality of bright wavelengths to illuminate the mask;

1 st illumination optical system having: a wavelength selecting section for receiving light from the light source and extracting an illumination light beam including at least 1 specific bright line wavelength among the plurality of bright line wavelengths and limited to a predetermined wavelength width, and a numerical aperture varying section for adjusting a divergence angle of the illumination light beam; and

a2 nd illumination optical system including an optical integrator that receives the illumination light beam whose divergence angle is adjusted, and irradiates the illumination light beam onto the mask with the same illuminance according to a numerical aperture corresponding to the divergence angle; and is

The 1 st wavelength selecting element is attached to the wavelength selecting section, and extracts spectral components of the specific bright line wavelength and spectral components of low luminance distributed in the vicinity of the bottom of the specific bright line wavelength while removing bright lines on the long wavelength side and bright lines on the short wavelength side which appear beside the specific bright line wavelength.

2. The exposure apparatus according to claim 1, wherein

The vicinity of the bottom portion where the low-luminance spectral component is distributed is set in a range where the relative intensity of the low-luminance spectral component is several% or more, preferably 10% or more on average, with respect to the peak intensity of the spectral component of the specific bright line wavelength.

3. The exposure apparatus according to claim 2, wherein

The wavelength selecting section is provided with a mechanism for mounting a2 nd wavelength selecting element so as to be exchangeable with the 1 st wavelength selecting element, and the 2 nd wavelength selecting element extracts a peak-like spectral component of the specific bright line wavelength other than the spectral component of the low luminance distributed in the vicinity of the bottom of the specific bright line wavelength.

4. The exposure apparatus according to claim 3, wherein

The wavelength selection unit is provided with a3 rd wavelength selection element which is extracted so as to include both a peak-shaped spectral component of at least 1 bright line wavelength appearing beside the specific bright line wavelength and a peak-shaped spectral component of the specific bright line wavelength, and which is mounted so as to be exchangeable.

5. An exposure method for projection exposure of a pattern of a mask onto a photosensitive substrate, comprising:

selecting wavelengths such that, together with a peak-like spectral component of at least 1 specific bright line wavelength in light from a light source that generates light including a plurality of bright line wavelengths, spectral components of low luminance distributed in the vicinity of the bottom of the specific bright line wavelength, which do not include bright lines on the long wavelength side and bright lines on the short wavelength side appearing beside the specific bright line wavelength, are extracted; and

the mask is irradiated with the illumination light beam of the selected spectral component at the same illuminance, and the pattern of the mask is projected and exposed on the substrate by a projection optical system of a mirror projection method in which no chromatic aberration occurs in the wavelength width including the spectral component of low luminance or a catadioptric method in which chromatic aberration is corrected in the wavelength width of the spectral component of low luminance.

6. The exposure method according to claim 5, wherein

7. The exposure method according to claim 6, wherein

The light source is an ultra-high pressure mercury discharge lamp, and the specific bright line wavelength is any one of i-ray, h-ray, and g-ray.

8. An exposure method, which is to irradiate a light beam containing a specific bright line wavelength selected by a wavelength selection part in the light beam containing the bright line wavelength generated by a light source device onto a mask carrying a pattern for an electronic component through an illumination optical system, and to project and expose an image of the pattern onto a photosensitive substrate by a projection optical system for receiving an exposure light beam generated by the mask; which comprises

Extracting, by the wavelength selection unit, a1 st spectral distribution light and a2 nd spectral distribution light having different wavelength bands from the light generated by the light source device; and

in order to perform kohler illumination on the mask by the illumination optical system, a1 st light source image distributed in a two-dimensional range by the 1 st spectral distribution light and a2 nd light source image distributed in a two-dimensional range by the 2 nd spectral distribution light are formed to overlap each other on a pupil plane in the illumination optical system.

9. The exposure method according to claim 8, wherein

The two-dimensional range of the 1 st light source image formed on the pupil plane is set within a circular area of the 1 st radius from the center of the pupil plane, and the two-dimensional range of the 2 nd light source image formed on the pupil plane is set within a circular area of the 2 nd radius from the center of the pupil plane.

10. The exposure method according to claim 9, wherein

The 1 st radius of the circular region forming the 1 st light source image and the 2 nd radius of the circular region forming the 2 nd light source image may be adjusted to be the same value or different values.

11. The exposure method according to any one of claims 8 to 10, wherein

The light source device includes: a1 st mercury discharge lamp that generates the 1 st spectral distribution of light extracted by the wavelength selection unit; and a2 nd mercury discharge lamp that generates the 2 nd spectral distribution of light extracted by the wavelength selection unit.

12. The exposure method according to claim 11, wherein

The 1 st mercury discharge lamp and the 2 nd mercury discharge lamp are respectively set such that the mercury vapor pressure in the discharge tube is 10⁶An extra-high pressure mercury discharge lamp of Pa or more.

13. The exposure method according to claim 12, wherein

Providing an i-ray-broadband interference filter in the wavelength selection unit, the i-ray-broadband interference filter extracting a spectral distribution in a bottom vicinity portion including a spectral component of the i-ray from a plurality of bright-line wavelengths included in light generated by the ultra-high pressure mercury discharge lamp and including an intensity of a few% or more, preferably 10% or more, with respect to a peak intensity of the spectral component of the i-ray;

either one of the light of the 1 st spectral distribution and the light of the 2 nd spectral distribution is extracted by the i-ray broadband interference filter.

14. The exposure method according to any one of claims 8 to 10, wherein

The light source device includes: a mercury discharge lamp for obtaining the light of the 1 st spectral distribution extracted by the wavelength selection unit; and a harmonic laser light source for obtaining the light of the 2 nd spectral distribution extracted by the wavelength selection unit.

15. An exposure method for illuminating a mask pattern with illumination light having a predetermined wavelength distribution, wherein an image of the mask pattern is projected and exposed onto a substrate by a projection optical system for injecting and projecting an imaging beam generated by the mask pattern onto the substrate; it includes:

setting a width of a wavelength distribution of the illumination light including a center wavelength λ such that a ratio of a minor axis length to a major axis length of a projection image of the hole pattern deformed into an elliptical shape becomes 80% or more, preferably 90% or more, when a projection image of a square or rectangular hole pattern having a size close to a minimum resolvable line width dimension determined by a resolution R defined by k · (λ/NAp) is projected onto the substrate, assuming that the specific center wavelength in the wavelength distribution of the illumination light is λ, a numerical aperture on the substrate side of the projection optical system is NAp, and a program constant is k (0 < k ≦ 1); and

the mask on which the pattern for electronic components is formed is illuminated with the illumination light having the wavelength distribution of the set width, and the pattern for electronic components is projection-exposed on the substrate.