Home
Patent Search
IMT Blog
REGISTER
|
SIGN IN
United States Patent
5298365
Okamoto , ; et al.
March 29, 1994
Title
Process for fabricating semiconductor integrated circuit device, and exposing system and mask inspecting method to be used in the process
Abstract
An exposure technology for a semiconductor integrated circuit device which has a pattern as fine as that of an exposure wavelength contemplates to improve the resolution characteristics of the pattern by making use of the mutual interference of exposure luminous fluxes.
Inventors:
Okamoto; Yoshihiko
(Ohme,
JP
)
, Moriuchi; Noboru
(Tokyo,
JP
)
Assignee:
Hitachi, Ltd.
(Tokyo,
JP
)
Appl. No.:
026200
Filed:
February 26, 1993
Foreign Application Priority Data
Mar 20, 1990 [JP] 2-71266
May 18, 1990 [JP] 2-126662
Sep 19, 1990 [JP] 2-247100
Current U.S. Class:
430/311
430/396
430/945
250/492.1
250/492.2
257/E21.645
257/E27.081
Field of Search:
430/311,327,396,945 250/492.1,492.2
U.S. Patent Documents
4153457
May 1979
Kellie
Foreign Patent Documents
234854
Feb., 1990
JP
298119
Apr., 1990
JP
60-107835
Jun., 1985
JP
60-109228
Jun., 1985
JP
62-59296
Dec., 1987
JP
62-67514
Mar., 1987
JP
Primary Examiner:
McCamish; Marion E.
Assistant Examiner:
Duda; Kathleen
Attorney, Agent or Firm:
Fay, Sharpe, Beall, Fagan, Minnich & McKee
Parent Case Text
This is a continuation of application Ser. No. 07/699,703, filed May 14, 1991, now abandoned which is a continuation-in-part application of U.S. Ser. No. 07/610,422, filed Nov. 7, 1990, now abandoned.
Claims
What is claimed is:
1. An exposure method for transferring a pattern of a mask onto a photoresist film located on a wafer by a size-reducing projection exposure optical system, comprising the steps of:
(a) dividing an exposure luminous flux coming from an exposing ultraviolet or far ultraviolet monochromatic light into a main luminous flux and an auxiliary luminous flux;
(b) irradiating the first principal plane of a main portion of a mask, which has a main opening pattern of minimum size corresponding to the vicinity of the resolution limit of said size-reducing projection exposure optical system, with said main luminous flux generally at a right angle to the main portion, to emit the transmitted main luminous flux from the second principal plane of said main portion;
(c) irradiating the first principal plane of an auxiliary portion of the mask, which has an auxiliary opening pattern of minimum size far smaller than the resolution limit of said size-reducing projection exposure optical system, with said auxiliary luminous flux generally at a right angle to the auxiliary portion, to emit the transmitted auxiliary luminous flux from the second principal plane of said auxiliary portion;
(d) composing the emanating main luminous flux and auxiliary luminous flux with a desired phase difference to emit a composed luminous flux; and
(e) projecting said emanated composed luminous flux on the photoresist film on the principal plane of said wafer by the size-reducing projection lens system of said size-reducing projection exposure optical system so that corresponding portions of said main luminous flux and said auxiliary luminous flux interfere with each other to focus a clear image.
2. An exposure method according to claim 1, wherein said desired phase difference is substantially an odd integral multiple of .pi..
3. An exposure method as claimed in claim 1, wherein the mask comprises first and second masks, said main portion being the first mask, and said auxiliary portion being the second mask.
4. An exposure method for transferring a pattern of a mask onto a photoresist film located on a wafer by a size-reducing projection exposure optical system, comprising:
(a) dividing an exposure luminous flux coming from an exposing ultraviolet or far ultraviolet monochromatic light into a main luminous flux and an auxiliary luminous flux;
(b) irradiating the first principal plane of a main portion of a mask, which has a main opening pattern of minimum size corresponding to the vicinity of the resolution limit of said size-reducing projection exposure optical system corresponding to an odd array, with said main luminous flux generally at a right angle to the main portion, to emit the transmitted main luminous flux from the second principal plane of said main portion, wherein said main opening pattern is periodic in at least one axial direction;
(c) irradiating the first principal plane of an auxiliary portion of the mask, which has an auxiliary opening pattern of minimum size in the vicinity of the resolution limit of said size-reducing projection exposure optical system with said auxiliary luminous flux generally at a right angle of the auxiliary portion to emit the transmitted auxiliary luminous flux from the second principal plane of said auxiliary portion, wherein said auxiliary opening pattern is periodic in at least one axial direction and corresponds to an even array;
(d) composing the emanating main luminous flux and auxiliary luminous flux with a desired phase difference to emit a composed luminous flux; and
(e) projecting said emanated composed luminous flux on the photoresist film on the principal plane of said wafer by the size-reducing projection lens system of said size-reducing projection exposure optical system so that said main luminous flux and said auxiliary luminous flux interfere with each other to focus a clear image corresponding to said periodic pattern.
5. An exposure method according to claim 4, wherein said desired phase difference is a desired value of substantially an odd integral multiple of .pi..
6. An exposure method according to claim 5, wherein said desired phase difference is shifted to and from the desired value so that the pattern on the wafer corresponding to the opening patterns in the odd array and the even array is formed on different planes.
7. An exposure method as claimed in claim 4, wherein said mask comprises first and second masks, said main portion being said first mask, and said auxiliary portion being said second mask.
Description
BACKGROUND OF THE INVENTION
The present invention relates to exposure technology and, more particularly, to a technology which is effective if applied to the photolithography process of a semiconductor integrated circuit device.
As the high integration of the semiconductor integrated circuit advances so that the circuit elements and the wiring design rules come to the sub-micron order, the photolithography process for transferring a circuit pattern on a mask onto a semiconductor wafer by making use of a beam of g- or i-line is troubled by a serious problem of a reduction in the precision of the circuit pattern to be transferred onto the wafer. In case, for example, a circuit pattern formed of transparent regions P.sub.1 and P.sub.2 and a shielding region N over a mask 20, as shown at (a) in FIG. 1J, is to be transferred onto the wafer, the phases of the light beams L just after having passed through the paired transparent regions P.sub.1 and P.sub.2 interposing the shielding region N are in phase with each other, as shown at (b) in the same Figure. As a result, the two light beams interfere with each other at that portion on the wafer, which might otherwise be intrinsically the shielding region, so that they are intensified (as shown at (c) in the same Figure). As a result, the contrast of the projected image on the wafer drops with a reduced focal depth, as shown at (d) in the same Figure, so that the pattern transfer precision is seriously degraded.
As means for solving this problem, there has been proposed a phase shift technology for preventing the drop of the contrast of the projected image by changing the phases of the light beams that transmit through the mask. In Japanese Patent Publication No. 59296/1987, for example, there is disclosed a phase shifting technology, in which one of the paired transparent regions across the shielding region is formed with a transparent film to establish a phase difference between the light beams having passed through the two transparent regions at the time of exposure so that the interfering beams may be weakened at that portion on the wafer, which might otherwise be the shielding region. When a circuit pattern formed on a mask 21, as shown at (a) in FIG. 1K, is to be transferred onto a wafer, either of the paired transparent regions P.sub.1 and P.sub.2 interposing the shielding region N is formed with a transparent film 22 having a predetermined refractive index. By adjusting the thickness of the transparent film 22, moreover, the individual light beams having passed through the transparent regions P.sub.1 and P.sub.2 go out of phase by 180 degrees, as shown at (b) in the same Figure, so that they interfere with each other in the shielding region N on the wafer and are weakened (as shown at (c) in the same Figure). As a result, the contrast of the projected image on the wafer is improved, as shown at (d) in the same Figure, to improve the resolution and the focal depth and accordingly the transfer precision of the circuit pattern formed on the mask 21.
In Japanese Patent Laid-Open No. 67514/1987, on the other hand, there is disclosed a phase shift technology in which a phase difference is established between a light beam having passed through a transparent region and a light beam having passed through a fine opening pattern, by removing the shielding region of a mask partially to form the opening pattern and by forming a transparent film in either the opening pattern or the transparent region existing in the vicinity of the opening pattern, so that the light beam having passed through the transparent region may be prevented from having its amplitude distribution expanded transversely.
The phase shifting method, by which one mask is formed thereon with an ordinary pattern (or main pattern) and a shifter pattern (or accompanying or complementary pattern) for giving a phase opposed to that of the former, will be called hereinafter the "on-mask phase shifting method" and will be called the "on-mask phase inversion shifting method" especially in case the phase shift is (2n+1).rho. (wherein n: an integer).
In Japanese Patent Laid-Open No. 109228/1985, moreover, there is disclosed a method, in which two masks are simultaneously illuminated to improve the throughput of a projecting exposure so that the portions of one wafer corresponding to different chips may be simultaneously exposed. In Japanese Patent Laid-Open No. 107835/1985, on the other hand, there is disclosed a technology, in which two masks having an identical pattern can be exposed without any trouble even if one of them is defective, by dividing one exposing line into two halves to illuminate the identical portion of the two masks and by composing them to expose the wafer.
However, these two disclosures are not effective in the least for improving the resolution although they are effective for preventing the defect on the mask pattern from being transferred onto the wafer or for improving the throughput.
SUMMARY OF THE INVENTION
According to our examinations, the aforementioned phase shifting technology of the prior art, in which the transmitting region of the mask is in its portion with the transparent film so that a phase difference may be established between the light having passed therethrough and the light having passed through the neighborhood transmitting region, is troubled by a problem that the manufacture of the mask takes a long time period and many process steps.
Specifically, the actual mask formed with an integrated circuit pattern is complicated by various patterns so that the mask makes it seriously difficult to select the place to be arranged with the transparent film, thereby placing serious restrictions upon the pattern design. In case the mask is formed with the transparent film, on the other hand, in addition to the step of inspecting the existence of a defect in the integrated circuit pattern, there is required a step of inspecting the existence of a defect in the transparent film so that the mask inspecting step is seriously complicated. In case, moreover, the mask is formed with the transparent film, foreign substances caught by the mask are increased, making it difficult to prepare a clean mask.
An object of the present invention is to provide a phase shifting technology which has succeeded in solving the above-specified problems.
Another object of the present invention is to provide a size-reducing projection exposure technology which is enabled to give the best image plane to the individual planes to be exposed, if stepped, by a single exposure.
A further object of the present invention is to provide a projection exposure technology which can extend the exposure limit of a fine pattern by violet or ultraviolet rays to a finer range.
A further object of the present invention is to provide a projection exposure technology which can compose and expose two master patterns.
A further object of the present invention is to provide a size-reducing projection exposure technology which can compose and interfere two mask patterns to be projected and exposed, even in case an interference distance of a light source is short.
A further object of the present invention is to provide a mask pattern layout technology which is useful for fabricating an integrated circuit by using the phase shifting method.
A further object of the present invention is to provide a projection exposure technology which is useful for fabricating an SRAM by using the phase shifting method or the like.
A further object of the present invention is to provide an exposure technology which is useful for fabricating a highly integrated semiconductor circuit such as a DRAM having a fine size as small as the exposure wavelength.
A further object of the present invention is to provide a projection exposure technology which is effective for exposing a periodic fine pattern.
A further object of the present invention is to provide a projection exposure technology which is effective if applied to an excimer laser exposure technology.
A further object of the present invention is to provide a mask inspection technology which is useful for inspecting a mask to be used in the phase shifting method.
The invention to be disclosed hereinafter will be briefly described in the following in connection with its representatives embodiments.
According to one representative embodiment of the present invention, there is provided the following exposure method. When a predetermined pattern, which is formed on a mask and composed of a shielding region and a transparent region, is to be transferred onto a specimen to be illuminated, by irradiating the mask with a light and irradiating the specimen with the light having passed through the transparent region of said mask, the light emitted from a light source is divided into two light beam, and the phases of the two light beams immediately after having passed through different portions of said mask are opposed to each other by changing the individual optical lengths for said two light beams to reach said mask. After this, said two light beams are composed to illuminate said specimen.
According to the above-specified means, the two light beams immediately after having passed through the different portions of the mask have their phases opposed to each other and are then composed to illuminate the specimen. As a result, one light beam having passed through a predetermined transparent region on the mask and the other light beam having passed through another transparent region on the mask interfere with each other and are weakened in their boundary regions at the portions, in which they are arranged close to each other, so that the projected image has its contrast improved drastically.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1A is a diagram showing the whole structure of a phase shifting mechanism which is disposed in an exposure system according to an embodiment 1-I of the present invention;
FIG. 1B is an enlarged section showing a mask according to the aforementioned embodiment of the present invention;
FIGS. 1C (a) and (b) are top plan views showing a pair of circuit patterns formed on the mask;
FIG. 1C (c) is a top plan view showing a circuit pattern which is obtained by composing the paired circuit patterns;
FIGS. 1D (a) to (e) are explanatory diagrams showing the amplitudes and intensity of the light beams which have passed through the transparent region of the circuit patterns shown at (a) and (b) in FIG. 1C;
FIGS. 1E (a) and (b) are top plan views showing a pair of registering marks formed on the mask;
FIG. 1E (c) is a top plan view showing a circuit pattern which is obtained by composing those paired registering marks;
FIGS. 1F (a) and (b) are top plan views showing another example of a pair of circuit patterns formed on a mask according to Embodiment 1-II of the present invention;
FIG. 1F (c) is a top plan view showing a circuit pattern which is obtained by composing the paired circuit patterns;
FIGS. 1G (a) to (e) are explanatory diagrams showing the amplitudes and intensities of the light beams which have passed through the transparent regions of the circuit patterns shown at (a) and (b) in FIG. 1F;
FIGS. 1H (a) and (b) are top plan views showing other examples of a pair of circuit patterns which are formed on a mask according to Embodiment 1-III of the present invention;
FIG. 1H (c) is a top plan view showing a circuit pattern which is obtained by composing the paired circuit patterns;
FIGS. 1I (a) to (e) are explanatory diagrams showing the amplitudes and intensities of the light beams which have passed through the transparent regions of the circuit patterns shown at (a) and (b) in FIG. 1H;
FIGS. 1J (a) to (d) are explanatory diagrams showing the amplitudes and intensities of the light beams which have passed through the transparent regions of the mask of the prior art; and
FIGS. 1K (a) to (d) are explanatory diagrams showing the amplitudes and intensities of the light beams which have passed through the transparent regions of the mask of the prior art having the transparent film.
FIG. 2A is a diagram showing the structure of an essential portion of the exposure optical system according to Embodiment 2 of the present invention;
FIGS. 2B (a) and (b) are top plan views showing examples of the individual pattern structures of the masks of FIG. 2A, and FIG. 2B (c) is a top plan view showing a desired pattern which is formed by those patterns;
FIGS. 2C (a) and (b) are top plan views showing essential portions of examples of the individual pattern structures of the masks of FIG. 2A, and FIG. 2C (c) is a top plan view showing a desired pattern which is formed by those patterns;
FIGS. 2D (a) and (b) are top plan views showing essential portions of examples of the individual pattern structures of the masks of FIG. 2A, and FIG. 2D (c) is a top plan view showing a desired pattern which is formed by those patterns;
FIGS. 2E (a) and (e) are explanatory diagrams showing the amplitudes and intensities of the light beams which have passed through the transparent regions of the masks of FIG. 2B;
FIGS. 2F (a) to (e) and (d') are explanatory diagrams showing the amplitudes and intensities of the light beams which have passed through the transparent regions of the masks of FIG. 2C;
FIGS. 2G (a) to (e) are explanatory diagrams showing the amplitudes and intensities of the light beams which have passed through the transparent regions of the masks shown in FIG. 2D;
FIG. 2H is a section showing the mask; and
FIGS. 2I (a) to (c) are explanatory diagrams showing a pattern registering method to be used in the system of the present invention.
FIG. 3A is a schematic front section showing the summary of an exposure optical system of a step-and-repeat type 5:1 size-reducing projection exposure system according to Embodiment 3 of the present invention;
FIG. 3B is a section showing a mask which corresponds to the periodic or pseudo-periodic line-and-space pattern of the aforementioned embodiment of the present invention;
FIG. 3C (a) is a top plan view showing a main mask pattern (or positive mask) which corresponds to a stepped periodic pattern of the aforementioned embodiment;
FIG. 3C (b) is also a top plan view showing a sub-mask pattern;
FIG. 3C (c) is a top plan view showing a composed opening pattern;
FIG. 3C (d) is a section showing a periodic step portion of the semiconductor integrated circuit device which is being fabricated on the exposed wafer;
FIG. 3D is a diagram showing the behavior of the displacements of an image plane corresponding to the main and sub-patterns in case the phase difference .PHI. of L.sub.1 and L.sub.2 of the foregoing embodiment is displaced to and from (2n+1).pi.;
FIG. 3E (a) is a top plan view showing the phase shift registering mark of the foregoing embodiment, that is formed in the main pattern portion;
FIG. 3E (b) is a top plan view showing the phase registering opening pattern which is formed in the same sub-pattern; and
FIG. 3E (C) is a top plan view showing the composed projection pattern.
FIG. 4 is a schematic section showing a stepper system of Embodiment 4 of the present invention.
FIG. 5A is a schematic front section showing the exposure projection optical system of a step-and-repeat type 5:1 size-reducing projection exposure system of Embodiment 5 of the present invention;
FIG. 5B is a schematic front section showing an exposure light source and illumination (or exposure) optical system of the same system;
FIG. 5C is an enlarged section showing the phase difference setting means of the same system; and
FIG. 5D is a top plan view showing the wafer holding portion of the same system.
FIG. 6A is a top plan view showing a mask pattern corresponding to an isolated band pattern according to Embodiment 6 of the present invention;
FIG. 6B is a top plan view showing a mask pattern corresponding to the isolated square pattern according to Embodiment 6 of the present invention;
FIG. 6C is a top plan view showing a mask pattern corresponding to a isolated square pattern according to a modification of FIG. 6B;
FIG. 6D is a top plan view showing a mask pattern corresponding to an "L"-shaped pattern according to Embodiment 6 of the present invention;
FIG. 6E is a top plan view showing a mask pattern corresponding to an "L"-shaped pattern according to a modification of FIG. 6D;
FIG. 6F is a top plan view showing a mask pattern corresponding to a bent isolated band pattern according to Embodiment 6 of the present invention;
FIG. 6G is a top plan view showing a mask pattern corresponding to a bent isolated band pattern according to a modification of FIG. 6F; and
FIG. 6H is a top plan view showing a mask pattern corresponding to an equal-period band pattern according to Embodiment 6 of the present invention.
FIG. 7A is a top plan view showing a wafer at an exposure step according to Embodiment 7 of the present invention;
FIG. 7B is a top plan view showing a unit exposure region in an exposure method according to Embodiment 7 of the present invention;
FIGS. 7C to 7E are flow sections showing a positive process according to Embodiment 7 of the present invention;
FIGS. 7F to 7H are flow sections showing a negative process according to Embodiment 7 of the present invention;
FIG. 7I is an overall flow chart showing a photolithography step in a twin-well SRAM according to Embodiment 7 of the present invention;
FIGS. 7J to 7P are flow sections showing the wafer step of an SRAM corresponding to FIG. 7I of the present invention; and
FIG. 7Q is a top plan layout showing a chip region of the aforementioned SRAM.
FIGS. 8A to 80 are flow sections showing the wafer step of a DRAM according to Embodiment 8 of the present invention;
FIG. 8P is a top plan layout showing a chip region of the aforementioned DRAM; and
FIG. 8Q is a top plan layout showing a unit rotation period of the memory cell region of the aforementioned DRAM.
FIG. 9A is a graph for explaining the distributions of the amplitude intensity and the energy intensity of the line in case the adjacent patterns are in phase;
FIG. 9B is a graph showing the same distribution in case the phase is shifted (relatively) by 180 degrees from FIG. 9A; and
FIG. 9C is a schematic section showing an optical system for explaining the principle of a size-reducing projection of the present invention.
FIG., 10A is a diagram showing several conditions of an exposing monochromatic light source to be used in the exposure method of the present invention.
FIG. 11A is a simplified front section showing the 5:1 size-reducing exposure system of Embodiment 11 of the present invention, in which all the projection lens systems are shared by making use of a telecentric structure at an objective side.
FIG. 12A is a simplified front section showing a mask inspecting system according to Embodiment 12 of the present invention.
FIG. 13A is a simplified front section showing a step-and-repeat type 5:1 size-reducing projection exposure system of Embodiment 13 of the present invention using two light sources which are not coherent with each other; and
FIG. 13B is a mask or wafer top plan view showing the layout of a unit exposure region to be exposed by the exposing method of FIG. 13A.
FIG. 14 is a simplified front section showing a step-and-repeat type size-reducing exposure system (using light sources which are not coherent with each other) for explaining the exposing method of Embodiment 14 of the present invention;
FIG. 14B is a top plan layout showing a unit exposure region (such as the mask or wafer) in the aforementioned method of FIG. 14A; and
FIG. 14C is a top plan pattern view showing the mask to be used in the aforementioned method of FIG. 14A.
FIG. 15A is a top plan view showing a pattern on a wafer, which corresponds to a pseudo-periodic pattern of Embodiment 15 of the present invention;
FIG. 15B is a top plan view showing a pattern on a wafer, which corresponds to another pseudo-periodic pattern of the aforementioned embodiment;
FIG. 15C is a top plan view showing a pattern on a wafer, which corresponds to still another pseudo-periodic pattern of the aforementioned embodiment;
FIG. 15D is a top plan layout or a superposed top plan layout showing either an on-mask corresponding to a pattern on the wafer of FIG. 15A or a mask in a multi-mask phase shifting method; and
FIGS. 15E and 15F are top plan layouts corresponding to FIGS. 15B and 15C, respectively.
FIG. 16A is a table enumerating photo resists to be used for practicing the present invention.
FIG. 17A is a simplified front section showing a step-and-repeat type 5:1 size-reducing projection exposure system for an exposure method in which an accompanying pattern according to Embodiment 17 of the present invention; and
FIG. 17B is a superposed mask pattern for explaining the same method.
FIG. 18A is a front section showing a simplified multi-mask-stepper according to Embodiment 18 of the present invention.
FIG. 19A is a front section showing a pair mask (or pattern) exposure system (or stepper) according to a single mask substrate of Embodiment 19 for explaining the structure of an individual illumination light source of the exposure system of each embodiment of the present invention.
FIG. 20A is a diagram showing the overall structure of a two-dimensional phase adjusting system according to Embodiment 20 of the present invention;
FIG. 20B is a top plan view showing the same two-dimensional phase shifting plate;
FIG. 20C is a section showing the same two-dimensional phase shifting plate; and
FIG. 20D is a table showing crystals having the electrochemical effect to be used in the same phase shifting plate.
DETAILED DESCRIPTION OF THE INVENTION
The following description of the embodiments of the present invention will be divided into a plurality of items for convenience, but the individual embodiments belong not to different inventions but to portions or modifications of a step relating to a single invention. Therefore, any overlapped portion will not be described unless otherwise necessary. Moreover, the reference numbers to be used in the following embodiments designate parts performing identical or similar functions, unless otherwise specified, if they have identical numerals in the lower two figures.
(1) Embodiment 1
FIG. 1A shows a phase shifting mechanism 1 of an exposure system according to Embodiment 1-I of the present invention.
The phase shifting mechanism 1 is constructed of an optical system which is interposed between a light source 2 for the exposure system and a specimen 3 to be irradiated and which includes a beam expander 4, mirrors 5, 6 and 9, half mirrors 7 and
8, a corner mirror 10, an optical path varying mechanism 11 for driving the corner mirror 10 finely, a pair of lenses 12a and 12b, and a size-reducing lens 13. In the alignment system of this optical system, there is positioned a mask 14 which is formed with the original image of a pattern to be transferred to the aforementioned irradiated specimen 3. The mask 14 (e.g., reticle) is to be used in the process for fabricating a semiconductor integrated circuit device, for example, and the irradiated specimen 3 is a semiconductor wafer which is made of a single crystal of silicon, for example.
A light beam L such as the i-line (having a wavelength of 365 nm) emitted from the light source 2 is expanded by the beam expander 4 and is then refracted through the mirror 5 in a direction normal to the principal plane of the mask 14. After this, the refracted light is divided through the half mirror 7 disposed midway of the optical path into two halves: a straight light beam L.sub.1 and a perpendicular light beam L.sub.2. This light beam L.sub.2 is refracted through the mirror 9 and the corner mirror 10 until it irradiates another portion of the mask 14 by way of a path different from that of the light beam L.sub.1. These two light beams L.sub.1 and L.sub.2 having thus passed through the different portions of the mask 14 are guided through the lenses 12a and 12b and are then composed into one light beam L' through the mirror 6 and the half mirror 8. After this, the single light beam L' is reduced by the size-reducing lens 13 and irradiates the specimen 3 which is positioned on an X-Y table 15.
Since the aforementioned phase shifting mechanism 1 has different optical paths for the two light beams L.sub.1 and L.sub.2 from the half mirror 7 to the mask 14, a desired phase difference can be established between the two beams L.sub.1 and L.sub.2 immediately after passage through the mask 14 by changing the height (or the optical path of the light beam L.sub.2) from the principal plane of the mask 14 to the corner mirror 10. For example, the phases of the two light beams L.sub.1 and L.sub.2 immediately after having passed through the mask 14 can be opposed to each other (i.e., to give a phase difference of an add integer multiple of .pi., or effectively 180 degrees) by moving the corner mirror 10 vertically from an origin, which is assumed to be the position of the corner mirror 10 when the two light beams L.sub.1 and L.sub.2 immediately after having passed through the mask 14 are in phase, by a distance (d) which is defined by the following formula:
(.lambda.: wavelength of the light beam; and m: an integer).
The vertical movement of the aforementioned corner mirror 10 can be accomplished by using the optical path varying mechanism 11 using a piezoelectric control element or the like.
FIG. 1B is an enlarged view showing a section of the aforementioned mask 14.
This mask 14 is made of transparent synthetic quartz having a refractive index of about 1.47 and has its principal plane formed with a metal layer of Cr or the like having a film thickness of about 500 to 3,000 .ANG.. For the exposure, the metal layer 16 provides shielding regions A for inhibiting transmission of light, and the remaining region provides a transparent region B for transmitting light. An integrated circuit pattern is formed of the aforementioned shielding regions A and transparent region B and has a size of five times as large as the actual one, for example.
FIGS. 1C (a) and (b) present examples of the integrated circuit patterns formed on the aforementioned mask 14. A circuit pattern P.sub.1, as shown at (a) in the same Figure, is formed of the hatched shielding region A and the L-shaped transparent regions B, for example, which are surrounded by that shielding region A. On the other hand, the transparent regions B of a circuit pattern P.sub.2, as shown at (b) in the same Figure, are patterned by arranging the shielding regions A, which have the same shape and size as those of the transparent regions B of the circuit pattern P.sub.1, in the transparent regions B, which have the same shape but an enlarged size as that of the transparent regions B of the circuit pattern P.sub.1. In other words, the transparent regions B of the circuit pattern P.sub.2 are substantially identical to the patterns of the peripheral portions of the transparent regions B of the circuit pattern P.sub.1. These two circuit patterns P.sub.1 and P.sub.2 are paired to transfer the (hatched) circuit pattern P, as indicated at (c) in FIG. 1C, highly precisely to the wafer, and are arranged in predetermined portions of the mask 14 and at a predetermined pitch.
Next, the method of preparing the aforementioned mask 14 will be briefly described in the following.
First of all, a synthetic quartz plate has its surface polished and rinsed. After this, a Cr film having a thickness of about 500 to 3,000 .ANG., for example, is deposited all over the principal plane of the quartz plate by the sputtering method. Subsequently, a photo resist is applied to the whole surface of the Cr film. Next, on the basis of the integrated circuit pattern data which are coded in advance in a magnetic tape or the like, an integrated circuit pattern is drawn on the photo resist by an electron beam exposing method. After this, the exposed portion of the photo resist is removed by a development, and the exposed Cr film is removed by a wet etching to form the integrated circuit pattern. The pattern data of the aforementioned paired circuit patterns P.sub.1 and P.sub.2 can be automatically prepared either by expanding or reducing the data of the shielding region A or the transparent regions B of one of the circuit patterns or by taking a logical product between the inverted data of one circuit pattern and the data of the other circuit pattern. For example, the pattern data of the circuit pattern P.sub.2 can be automatically prepared by taking a logical product between the enlarged data of the pattern of the transparent regions B of the circuit pattern P.sub.1 and the inverted data of the transparent regions B of the circuit pattern P.sub.1.
In order that the integrated circuit pattern prepared on the aforementioned mask 14 may be transferred onto the wafer 3, the wafer 3 having the photo resist applied to its surface is positioned at first on the X-Y table 15 of the aforementioned exposing system shown in FIG. 1, and the mask 14 is then positioned in the alignment system. Specifically, this mask 14 is so positioned that, when one light beam L.sub.1 divided by the half mirror 7 is guided to irradiate one circuit pattern P.sub.1 of the aforementioned paired circuit patterns P.sub.1 and P.sub.2, the other light beam L.sub.2 may be precisely guided to irradiate the other circuit pattern P.sub.2. Next, the corner mirror 10 is vertically moved to adjust the phase difference such that the two light beams L.sub.1 and L.sub.2 immediately after having passed through the mask 14 may be opposed to each other. In order to perform the positioning of the mask 14 and the adjustment of the phase difference between the two light beams L.sub.1
and L.sub.2 precisely, there are used, for example, a pair of positioning marks M.sub.1 and M.sub.2 which are formed on the mask 14, as shown at (a) and (b) in FIG. 1E. Each of the marks M.sub.1 and M.sub.2 is formed of a pattern, which has the hatched shielding region A and the transparent regions B having a square shape, for example, and surrounded by the shielding region A, and have absolutely identical size and shape. In case the positioning of the mask 14 and the adjustment of the phase difference between the light beams L.sub. 1 and L.sub.2 are precisely accomplished, the light beam L.sub.1 having passed through the mark M.sub.1 and the light beam L.sub.2 having passed through the mark M.sub.2 interfere with each other and completely disappear so that a projected image M of the marks M.sub.1 and M.sub.2 is not formed on the wafer 3. In other words, whether or not the positioning of the mask 14 and the adjustment of the phase difference between the light beams L.sub.1 and L.sub.2 are precisely accomplished can be easily decided by discriminating the presence of the projected image M on the wafer 3.
After the positioning of the mask 14 and the adjustment of the phase difference between the light beams L.sub.1 and L.sub.2 have been thus accomplished, the original image of the integrated circuit pattern formed on the mask 14 is optically reduced to 1/5, for example, and projected on the wafer 3. The aforementioned operations are repeated while moving the wafer 3 sequentially stepwise.
FIG. 1D (a) is a section showing the mask 14 in the region where the aforementioned circuit pattern P.sub.1 is formed, and FIG. 1D (b) is a section showing the mask 14 in the region where the aforementioned pattern P.sub.2 is formed.
The light beam L.sub.1 immediately after having passed through the transparent regions B of the circuit pattern P.sub.1 and the light beam L.sub.2 immediately after having passed through the transparent regions B of the circuit pattern P.sub.2
are opposed in phase to each other, as shown at (a') and (b') in FIG. 1D. Since, moreover, the transparent regions B of the circuit pattern P.sub.2 are identical to the patterns of the peripheral portions of the transparent regions B of the circuit pattern P.sub.1, the composed light beam L' of the two light beams L.sub.1 and L.sub.2 takes the amplitude which is shown at (c) in the same Figure. As a result, if the composed light beam L' irradiates the wafer 3, it is weakened by the interference at the boundary of the original light beams L.sub.1 and L.sub.2, as indicated at (d) in the same Figure. As a result, as shown at (e) in the same Figure, the contrast of the image projected on the wafer 3 improves the resolution and the focal depth drastically.
Thus, in the exposing system of the present Embodiment 1, the light beam emitted from the light source 2 is divided into the two light beams L.sub.1 and L.sub.2, and these two light beams L.sub.1 and L.sub.2 immediately after having passed through the mask 14 are opposed in phase to each other by changing the optical paths for the two light beams L.sub.1 and L.sub.2 to reach the mask 14. After this, these two light beams L.sub.1 and L.sub.2 are composed to irradiate the wafer 3. Moreover, the mask 14 of the present Embodiment 1 has such paired circuit patterns P.sub.1 and P.sub.2 that the transparent regions B of one circuit pattern P.sub.2 are identical to the patterns of the peripheral portions of the transparent regions B of the other circuit pattern P.sub.1. as a result, by transferring the integrated circuit pattern formed on the aforementioned mask 14 onto the wafer 3 by using the aforementioned exposing system, the light beam L', which is obtained by composing the light beam L.sub.1 having passed through the transparent regions B of the circuit pattern P.sub.1 and the light beam L.sub.2 having passed through the transparent regions B of the circuit pattern P.sub.2, is weakened by the interference at the boundary between the original light beams L.sub.1 and L.sub.2 so that the contrast of the image projected on the wafer 3 can be drastically improved to transfer the circuit pattern P highly precisely to the wafer.
As a result, the following effects can be attained in the exposing method of the present Embodiment 1-I.
(1) Since the mask need not be equipped thereon with phase shift means such as a transparent film, unlike the phase shifting technology of the prior art, there is no restriction on the pattern design. In the present Embodiment 1-I, when one circuit pattern is to be transferred onto the wafer, a pair of circuit patterns have to be formed on the mask. These paired circuit patterns can be automatically formed either by expanding or reducing the data of the shielding region or transparent regions of one circuit pattern or by taking a logical product between the inverted data of one circuit pattern and the data of the other circuit pattern.
(2) There is no necessity for that step for inspecting the existence of a defect in the transparent film, which has been indispensable for the phase shifting technology of the prior art. In the present Embodiment 1-I, the defect inspection of the paired circuit patterns can be practiced like the ordinary mask by making a comparison with the original pattern data. Moreover, the size inspection can also be practiced like the ordinary mask by the laser photometry or the like. As a result, the mask inspecting step is not complicated.
(3) Since the mask is not equipped thereon with the phase shifting means such as the transparent film or the like, it can be rinsed by a method like that of the ordinary mask. As a result, it is possible to form a mask which has no foreign obstacle as in the ordinary mask.
(4) Thanks to the aforementioned items (1) to (3), the transfer precision of the circuit pattern can be improved while requiring neither a long time nor much work for preparing the mask.
FIGS. 1F (a) and (b) show another example (i.e., Embodiment 1-II) of the paired circuit patterns which are formed on the mask of the foregoing Embodiment 1-I.
Each of the circuit pattern P.sub.1 shown at (a) in the same Figure and the circuit pattern P.sub.2 shown at (b) in the same Figure is formed with the hatched shielding region A and the transparent regions B having a rectangular shape, for example, and enclosed by the shielding region A. The paired circuit patterns P.sub.1 and P.sub.2 are provided for transferring the (hatched) circuit pattern P, as shown at (c) in the same Figure, highly precisely to the wafer and are arranged in predetermined positions of the mask 14 and at a predetermined spacing. The circuit pattern P includes four patterns P.sub.A, P.sub.B, P.sub.C and P.sub.D which are identical in size and shape to one another. The transparent region B.sub.A of the circuit pattern P.sub.1 corresponds to the pattern P.sub.A, and the transparent region B.sub.C of the circuit pattern P.sub.1 corresponds to the pattern P.sub.C. Moreover, the transparent region B.sub.B of the circuit pattern P.sub.2 corresponds to the pattern P.sub.B, and the transparent region B.sub.D of the circuit pattern P.sub.2 corresponds to the pattern P.sub.D. In short, the individual transparent regions B of the paired circuit patterns P.sub.1 and P.sub.2 are alternately arranged.
FIG. 1G (a) is a section showing a portion of the mask 14 in the region where the aforementioned circuit pattern P.sub.1 is formed, and FIG. 1G (b) is a section showing a portion of the mask 14 in the region where the aforementioned circuit pattern P.sub.2 is formed.
The light beam L.sub.1 immediately after having passed through the transparent regions B of the circuit pattern P.sub.1 and the light beam L.sub.2 immediately after having passed through the transparent regions B of the circuit pattern P.sub.2
are opposed in phase to each other, as shown at (a') and (b') in FIG. 1G. Moreover, the composed light beam L' of the two light beams L.sub.1 and L.sub.2 have their boundaries close to each other, as shown at (c) in the same Figure. As a result, if the composed light beam L' is guided to irradiate the wafer 3, it interferes and is weakened at the boundary of the original light beams L.sub.1 and L.sub.2, as shown at (d) in the same Figure. As a result, as shown at (e) in the same Figure, the contrast of the image projected on the wafer 3 improves the resolution and focal depth drastically.
FIGS. 1H (a) and (b) show another example (i.e., Embodiment 1-III) of the paired circuit patterns formed on the mask of the aforementioned Embodiment 1-I.
The circuit pattern P.sub.1, as shown at (a) in the same Figure, is formed of the hatched shielding region A and the transparent regions B which have square shapes, for example, and are enclosed by that shielding region A. On the other hand, the transparent regions B of the circuit pattern P.sub.2, as shown at (b) in the same Figure, are arranged outside of the individual sides of the transparent regions B of the circuit pattern P.sub.1. These paired circuit patterns P.sub.1 and P.sub.2 are provided for transferring the (hatched) circuit pattern P, as shown at (c) in the same Figure, highly precisely to the wafer. The two circuit patterns P.sub.1 and P.sub.2 are arranged in the predetermined positions on the mask 14 and at a predetermined spacing.
FIG. 1I (a) is a section showing a portion of the mask 14 in the region wherein the aforementioned circuit pattern P.sub.1 is formed, and FIG. 1I (b) is a section showing a portion of the mask 14 in the region wherein the aforementioned circuit pattern P.sub.2 is formed.
The light beam L.sub.1 immediately after having passed through the transparent regions B of the circuit pattern P.sub.1 and the light beam L.sub.2 immediately after having passed through the transparent regions B of the circuit pattern P.sub.2
are opposed in phase to each other, as shown at (a') and (b') in FIG. 1I. Moreover, the composed light beam L' of the two light beams L.sub.1 and L.sub.2 have their boundaries close to each other, as shown at (c) in the same Figure. As a result, if the composed light beam L' is guided to irradiate the wafer 3, it interferes and is weakened at the boundary of the original light beams L.sub.1 and L.sub.2, as shown at (d) in the same Figure. As a result, as shown at (e) in the same Figure, the contrast of the image projected on the wafer 3 improves the resolution and focal depth drastically.
Although our invention has been specifically described on the basis of its embodiments, it should not be limited to the aforementioned embodiments but can naturally be modified in various manners within the scope thereof.
The description thus far made is directed mainly to the case in which our invention is applied to the mask used in the process for fabricating a semiconductor integrated circuit device or its background field of application. Despite this description, however, the present invention should not be limited thereto but can be widely applied to the exposure technology for transferring a predetermined pattern formed on a mask by irradiating a specimen with the light having passed through said mask.
The effects to be attained by the representative embodiment of the invention disclosed herein will be briefly described in the following.
Thus, there is provided the following exposure method. When a predetermined pattern, which is formed on a mask and composed of a shielding region and a transparent region, is to be transferred onto a specimen to be illuminated, by irradiating the mask with a light and irradiating the specimen with the light having passed through the transparent region of said mask, the light emitted from a light source is divided into two light beams, and the phases of the two light beams immediately after having passed through different portions of said mask are opposed to each other by changing the individual optical lengths for said two light beams to reach said mask. After this, said two light beams are composed to illuminate said specimen. According to this exposure method, one light beam having passed through the predetermined transparent regions on the mask and the other light beam having passed through the other transparent regions on the mask interfere with each other and are weakened at the portions, in which they are arranged close to each other on the specimen, so that the contrast of the projected image can be remarkably improved.
As a result, the transfer precision of the pattern can be improved with neither long time nor much labor for fabricating the mask.
(2) Embodiment 2
Another representative embodiment of the invention to be disclosed in the present embodiment will be briefly described in the following.
According to the first mode of the present invention, there is provided a mask including first and second patterns each having a shielding region and transparent regions, so that a desired pattern may be formed on a specimen to be irradiated, by irradiating said two kinds of patterns with two light beams having a phase difference and at least a partial coherence and by composing the transmitted patterns of said light beams wherein the improvement resides in that said first pattern and said second pattern are formed on either a common substrate or two separate substrates so that the light having passed through the transparent regions of said first pattern and the light beam having passed through the transparent regions of said second pattern may interfere with each other and be weakened.
According to the second mode of the present invention, there is provided an exposure system comprising: a light source for emitting a luminous flux having at least a partial coherence; luminous flux dividing means for dividing said coherent luminous flux into two halves; an optical phase shifting member disposed in either of the optical paths for composing the luminous fluxes again from said luminous flux dividing means; an optical system for composing the luminous fluxes having passed through a first pattern and a second pattern into a single luminous flux; and an optical system for reducing and projecting said single luminous flux on a specimen to be irradiated, wherein the improvement resides in that the phases of the light having passed through the first pattern and the light having passed through the second pattern are shifted as long as 180 degrees by said optical phase shifting member to form a composed desirable pattern on said specimen.
According to the third mode of the present invention, there is provided an exposure system wherein the first and second patterns on said first mask are irradiated with two light beams having a phase difference and at least a partial coherence so that the desired pattern is formed on said specimen by composing the transmitted patterns of said light beams.
Incidentally, the luminous flux having at least a partial coherence is intended herein to mean the luminous flux having such coherency as can achieve the interfering and weakening effects.
In the present embodiment, moreover, the boundary is intended to include not only the boundary of the sections forming the aforementioned desired pattern but also the region contained by the two sections.
According to the aforementioned means, in order that the light having passed through the transparent regions of the first pattern and the light having passed through the transparent regions of the second pattern may interfere with each other and be weakened at the boundary required to have the precision of the desired pattern, the first and second patterns on the mask are irradiated with the two light beams having a phase difference and at least a partial coherence, so that the desired pattern may be formed on the specimen by composing the transparent patterns of those light beams. As a result, it is possible to improve the transfer precision of the boundary which is required to have the precision of the desired pattern.
FIG. 2A is a diagram showing the structure of an essential portion of an exposing optical system according to one embodiment of the exposure system using the mask of the present invention; FIGS. 2B to 2D are top plan views showing an essential portion of the mask of the present invention using the aforementioned exposing optical system; and FIGS. 2E to 2G are explanatory diagrams corresponding to the FIGS. 2B to 2D, respectively, and showing the amplitudes and intensities of the light beams having passed through the mask.
The exposure system of the present embodiment is roughly divided into four functional elements. The first is the (first) element for irradiating a mask 209 with two luminous fluxes having a phase difference; the second is the (second) element made of the mask 209; the third is the (third) element for composing the two light beams having passed through the mask 209 to reduce them and irradiate a specimen 215 with the reduced light beams; and the fourth is the (fourth) element including an alignment mechanism for adjusting the composition of a single luminous flux.
The first element is constructed to include: a light source 201 for emitting a partially coherent light; an expander 202 for expanding the light emitted from the light source 201; mirrors 203 and 206 for folding the optical path; a half mirror
204 for transmitting an incident light beam partially and reflecting it partially; and a phase shifting member 205 for changing the phase of the light beam. On the other hand, the third element is constructed to include: lenses 210 and 211 for arranging two light beams having passed through the mask 209 into parallel light beams; a mirror 212; a half mirror 213; a size-reducing lens 214 for reducing the light; the specimen 215; and a movable specimen table 216. The fourth element is constructed to include: an alignment mechanism 207 for moving the mirror 203, the half mirror 204, the lens 210 and the mirror 212; and a control circuit 208 for the alignment mechanism 207.
In the aforementioned structure, the mirror 203 is provided for reducing the size of the system in its entirety but may be dispensed with by introducing the light directly from the expander 202. The half mirror 204 has a function to divide the light coming from the expander 202 into two halves and is arranged over the first pattern 209a on the mask 209. The phase shifting member 205 is interposed between the half mirror 204 and the mirror 206 or between the mirror 212 and the half mirror 213
and has an action to shift the phase to a predetermined extent. The phase shifting member 205 is made of a synthetic quartz having a refractive index of 1.47, for example. If the mask 209 is arranged so that a first luminous flux 230 coming from the mirror 212 and a second luminous flux 231 coming from the lens 211 have their phase difference reduced to 0 in the state without the phase shifting member 205, the phase shifting member used has the following thickness d:
wherein:
the light source has a wavelength .lambda.;
and the member has a refractive index n.
The reason why the phase shifting member 205 is used is to establish the phase difference of 180 degrees at the time of exposure between the lights having passed through the two transparent regions, i.e., the light having passed through the phase shifting member 205 and the light not having passed through the phase shifting member 205. In case, for example, the light to be emitted for the exposure time has a wavelength of .lambda.=0.365 .mu.m (i.e., the i-line) whereas the phase shifting member
205 has a refractive index of n=1.5, the phase shifting member 205 may be given a thickness X.sub.1 of m (i.e., an integer) times as large as 0.365 .mu.m.
The mirror 206 is one for guiding the light having passed through the half mirror 204 and the light having passed through the phase shifting member 205 in parallel. Incidentally, the mask 209 has its two patterns 209a and 209b arranged thereon at a right angle with respect to the two light beams 330 and 331.
The lenses 210 and 211 are arranged to have their optical axes aligned with the centers of the patterns 209a and 209b, respectively. The half mirror 213 is one for composing the two light beams 230 and 231. For this composition, the mirror 212
has a function to fold the light 230.
The alignment mechanism 207 for the fourth element is constructed to include a mechanism for moving that portion of the optical system of the exposure system which is necessary for the positioning, and is exemplified by a piezoelectric element. In FIG. 2A, the alignment mechanism is constructed to move the mirror 203, the half mirror 204, the lens 210 and the mirror 212. Depending upon the structure of the exposure system, however, the kinds and numbers of the optical elements to be moved are naturally varied. Incidentally, the method of controlling the movement of this alignment mechanism 207 will be described hereinafter.
Next, the structure of the mask 209 or the second element of the present invention will be described in the following.
First of all, it is assumed that the (desired) pattern to be formed on the specimen 215 be a pattern having a shape of inverted letter "L" and a two-dimensional extension, as shown at (c) in FIG. 2B. FIGS. 2B (a) and (b) are top plan views showing individual examples of the first pattern 209a and the second pattern 209b, which are formed on the mask 209 so as to form such a desired pattern. These examples are arranged while holding the relative positional relation by considering the desired pattern to be composed by the specimen 215, as shown at (c) in the same Figure.
The first pattern 209a and the second pattern 209b are individually formed by combining the shielding regions and the transparent regions. These patterns may be formed on a single substrate but separately on two glass substrates. In this case, however, even the difference between the thicknesses of the glass substrates is corrected in terms of the thickness of the aforementioned phase shifting member. Incidentally, the transparent regions are indicated by blanks, and the shielding regions are hatched, as shown in FIG. 2B.
A transparent pattern 232 at (a) of FIG. 2B has a transparent region 234' having a shape of inverted letter "L", and a transparent pattern 236 at (b) of FIG. 2B has a shielding region 234 having a shape of inverted letter "L", which is made slightly smaller and arranged in the transparent region of the inverted letter "L", so that it makes a band-shaped transparent region 236.
Next, the operations of the present invention will be described in the following.
The light emitted from the light source 201 to have at least a partial coherence is expanded by the expander 202 and has its optical path folded by the mirror 203 until it is divided into two luminous fluxes by the half mirror 204. The half mirror 204 usually has a transmittance of 50% and a reflectivity of 50% (i.e., equal reflectivity and transmittance). The phase shifting member 205 is arranged in the optical path of one of the two optical systems for the two divided luminous fluxes. The light having passed through the phase shifting member 205 is given a phase difference of 180 degrees and is then guided by the mirror 206 to irradiate the second pattern 209b of the mask 209. On the other hand, the light having passed through the half mirror 204 irradiates the first pattern 209a of the mask 209.
The two luminous fluxes having passed through the two patterns 209a and 209b on the mask 209 are guided again into parallel beams by the lenses 210 and 211 and are then composed. Specifically, the first light beam 230 having passed through the first pattern 209a has its optical path folded by the mirror 212 and is then composed into a single luminous flux with the second light beam 231 having passed through the lens 211 by the action of the half mirror 213.
After this, the specimen 215 held on the movable specimen table 216 is irradiated through the size-reducing lens 214 with the composed one of the patterns 209a and 209b at the two portions on the mask 209, to form the desired pattern on the specimen 215.
Here, when the desired pattern shown at (c) in FIG. 2B is to be projected, the transfer precision of the pattern of the mask 209 is improved if the transmitted light beams of the first pattern 209a and the second pattern 209b are composed with a phase difference of 180 degrees. The reason for that improvement will be described in the following.
First of all, as has been described hereinbefore, by considering the size-reducing ratio, the first pattern 232 of the mask 209 is formed into the pattern 232 which has a slightly larger outer circumference than that of the desired pattern 229
and having the transparent region in its inside. By considering the size reducing ratio, moreover, the second transparent pattern 236 is formed into the band-shaped transparent pattern which is prepared by subtracting the shielding pattern 234 having the same size as that of the desired pattern 229 from the first pattern 232.
With this structure, in the peripheral region 238 of the desired pattern 229 to be determined, the transmitted light coming from the second transparent pattern 236 and the transmitted light coming from the band region 236' inside of the first transparent pattern 232 can be weakened as a result of their interference to sharpen the boundary of the desired pattern 229. Moreover, the desired pattern 229 is the composition of the shield region 234 at the side of the second pattern and the transparent region 234' having the same size and formed at the side of the first pattern so that it is finally formed like the usual exposure. Incidentally, as shown at (c) in FIG. 2B, the hatched portions indicate those irradiated with the light beams, and the blank regions 238 indicate those which are weakened by the interference, as contrary to the patterns shown at (a) and (b) in FIG. 2B.
FIGS. 2E (a) and (b) are Y--Y sections showing the first pattern 209a and the second pattern 209b on the mask 209, respectively. Reference numeral 262 designates a substrate, and numeral 263 designates the shielding members. FIGS. 2E (a') and (b') show the amplitudes of the light beams immediately after having passed through the mask patterns, respectively. In the transparent regions 232 and the transparent regions 236 of the mask, it is found that the phase difference of 180 degrees is established between the light (b') having passed through the phase shifting member and the light (a') not having passed through the phase shifting member 205. FIG. 2E (c) is a diagram showing the amplitudes of the light beams immediately having passed through the first pattern and the second pattern and having been composed.
If the irradiation is made with the first pattern 232 only, the amplitude of the line on the wafer is given a gentle slope in the peripheral portion of the pattern by the diffraction of the light so that a sharp boundary cannot be achieved. In the present embodiment, however, a light beam 242 having passed through the transparent region 236 of FIG. 2B and having a phase difference of 180 degrees is arranged in the periphery of a light beam having passed through the transparent region 232 of FIG. 2B. As a result, the light beams interfere with each other and are weakened at the boundary of the desired pattern 229 to be determined, so that the amplitudes of the light beams are remarkably reduced. As a result, the blur of the contour of the image to be projected on the wafer is reduced to improve the contrast of the projected image drastically so that the resolution and the focal depth are remarkably improved (as shown at (d) in FIG. 2E). Incidentally, the optical intensity is a square of the optical amplitude so that the waveform of the optical amplitude on the wafer at the negative side is inverted to the positive side, as shown at (e) in FIG. 2E.
Thus, according to the mask of the present embodiment, when the desired pattern to be determined is one having a two-dimensional extension, the first pattern is slightly expanded from the outer periphery of the two-dimensional pattern (or the desired pattern) to form a transparent pattern having a transparent region inside thereof, and the second pattern is slightly expanded into a band-shaped transparent pattern having a slightly larger outer periphery than that of the first pattern, so that only the boundary of the desired pattern having the two-dimensional extension can be sharpened.
Incidentally, the mask 209 is formed with a positioning mark for positioning the first pattern 209a and the second pattern 209b. By this positioning mark, the drive of the aforementioned alignment mechanism 207 is controlled.
FIG. 2I shows one example of the mark for positioning the two separate patterns. This mark pattern is given absolutely identical structures, relative positions and sizes at (a) and (b). The shape of the mark to be used should not be limited to the square, as shown, but may be changed into an L- or cross-shape. In order to enhance the precision, however, it is advisable to provide a plurality of marks having an identical shape in separate directions. On principle, moreover, these positioning marks are formed on the mask 209 in such dimensions as are required for positioning the alignment mechanism 207. If the two-dimensional positioning on X-Y directions are required, those marks are required in the two-dimensional directions on the X-Y axes, but one dimension is frequently sufficient for the usual system of FIG. 2A.
The transmitted light having passed through the mark is identical, as if it is wholly shielded, in case the light beams have the phase difference of 180 degrees and are correctly registered in the positions. Therefore, this shielding state is monitored by the CRT or the like so that the positioning is completed if the conditions are satisfied.
In case, on the contrary, the shielding is not complete at the initial setting, the alignment mechanism 207 may be driven to effect the shielding thereby to accomplish the positioning of (a) and (b).
Next, the process for fabricating the mask 209 according to the present embodiment will be described in the following with reference to FIG. 2H.
The mask, or rectile, 209 of the present embodiment, as shown in FIG. 2A, is used in the predetermined step of fabricating the semiconductor integrated circuit device. Incidentally, the mask 209 of the present embodiment is formed with an original of the integrated circuit pattern having a size of five times as large as the actual one and is formed with the shielding regions A and the transparent regions B.
For fabrications, the transparent substrate 262 made of quartz or the like has its surface polished and rinsed at first. After this, the surface is formed with the metal layer 263 having a thickness of about 500 to 3,000 .ANG. and made of Cr or the like by the sputtering method. Next, a photo resist (which will be referred to as the "resist") having a thickness of 0.4 to 0.8 .mu.m is applied to the upper surface of the metal layer 263. Subsequently, the resist is pre-baked and is then irradiated at its predetermined portions with an electron beam E by the electron beam exposing method on the basis of the integrated circuit pattern data of the semiconductor integrated circuit device, which are coded in advance in the magnetic tape or the like. Incidentally, the integrated circuit pattern data are recorded with the positional coordinates, shapes and so on of the patterns.
Next, the patterns (a) and (b) are transferred to the resist by the electron beam exposing method on the basis of the pattern data (a) and (b) of FIG. 2B, for example.
The pattern data (a) and (b) are automatically prepared by enlarging or reducing the pattern widths of the shielding regions A or the transparent regions B of the aforementioned integrated circuit pattern data. In the present embodiment, for example, the pattern (a) can be automatically prepared by enlarging the pattern width of the shielding regions to about 0.5 to 2.0 .mu.m, for example, and the pattern (b) can be automatically prepared by taking the logical product between the data of the pattern (a) and the inverted data of the original data.
After this, through the steps of the development, the etching of the predetermined portions, the removal of the resist, the rinsing, the inspections and so on, the mask 209 having the patterns (a) and (b) is fabricated.
The integrated circuit pattern on the mask 209 is transferred onto the specimen (which will be referred to as the "wafer") having the resist applied thereto, by using the mask 209 thus prepared, as will be described in the following.
Specifically, the size-reducing projection exposing system of FIG. 2A is arranged with the mask 209 and the wafer, and the original image of the integrated circuit pattern on the mask 209 is optically reduced in size to one-fifth and projected on the wafer. Each time the wafer is sequentially moved stepwise on the movable specimen table 216, the projection and exposure are repeated to transfer the integrated circuit pattern on the mask 209 to the whole surface of the wafer.
Next, another example of the mask according to the present embodiment will be described in the following.
FIGS. 2C (a) and (b) are sections showing the essential portions of the mask according to the present invention, respectively. The patterns (a) and (b) are the first and second patterns of the mask 209 of FIG. 2A, respectively, and show the mask patterns in sections while holding and separating the relative positional relations by taking the desired pattern into consideration. Incidentally, the pattern (c) shows the top plan view of the desired pattern composed. FIGS. 2F (a) to (e) are diagrams for explaining the amplitudes and intensities of the light beams having passed through the transparent regions of the mask shown in FIG. 2C. Incidentally, the exposure system and the method to be used are similar to those of the foregoing embodiment.
When a desired pattern 248 is a one-dimensional pattern having lines 244 to 247 arrayed transversely on line, according to the embodiment shown in FIG. 2C, the pattern on the mask is constructed to sharpen its boundary. In this case, in the relative arrangement of the mask, there are alternately arranged: transparent regions 249 and 250 of the first pattern constituting the lines 244 and 246 of the aforementioned lines 244 to 247; and transparent regions 251 and 252 of the second pattern constituting the lines 245 and 247. Then, the regions to be weakened as a result of the interferences are located at the intermediate regions 255 of the individual lines constituting the aforementioned desired pattern 248 so that the individual lines are sharpened.
With reference to (a) to (e) of FIG. 2F, the relations will be described in the following in case only the lines 244 and 245 are extracted from the desired pattern. In this case, too, the phase difference of 180 degrees is established (as shown at (a') and (b') in FIG. 2F) between the light beam 256 having passed through the transparent region 249 of the first pattern and the light beam 257 having passed through the transparent region 251 of the second pattern. As a result, these light beams have their components 259 and 260 interfering with each other and deleted at the regions 255 between the two lines 244 and 245 in the desired pattern on the wafer so that the optical amplitude has its gradient 261 enlarged, as shown at (d) in FIG. 2F. As a result, a sharp boundary can be formed in the region between the lines 244 and 245, as shown in FIG. 2C. Incidentally, FIG. 2F (d') is a schematic diagram showing the amplitude of the light on the wafer before the interference.
As a result, the contrast of the projected image of the one-dimensional pattern can be drastically improved to improve the resolution and the focal depth remarkably (as shown at (e) in FIG. 2F).
According to the present embodiment, if the desired pattern is a one-dimensional one having lines arranged transversely on line, the transparent regions of the first pattern and the second pattern forming the aforementioned lines are alternately arranged in the relative positions on the mask, and the aforementioned regions for interfering and weakening the lights are arranged at the intermediate portions of the individual lines forming the aforementioned desired pattern, so that the transfer precision can be drastically improved in case the plural lines are arranged in such a narrow region that the aforementioned two-dimensional pattern method cannot be applied.
Next, other examples of the mask according to the present invention will be described in the following.
FIGS. 2D (a) and (b) are diagrams showing the essential portions of the masks according to the present invention. The diagrams (a) and (b) are top plan views showing the first and second patterns of the mask 209 of FIG. 2A, respectively, such that the mask pattern is divided while holding the relative positions by considering the desired pattern thereof. FIGS. 2G (a) to (e) are diagrams for explaining the amplitudes and intensities of the light beams having passed through the transparent regions of the mask shown in FIG. 2D. Incidentally, the exposure system and method to be used are similar to those of the foregoing embodiments.
The desired pattern 269 of the present embodiment has square mask patterns 270 arranged therearound with fine sub-patterns 272.
The precise transfer of the fine sub-patterns 272 around the two-dimensional patterns 270 is difficult to perform by the method of applying the phase transparent film to the mask according to the prior art. According to the present invention, however, the desired excellent pattern 269 can be easily formed. In the mask of the present embodiment shown in FIG. 2D, too, in the relative positions on the mask, the first pattern is formed into a pattern 274 which is given transparent regions as wide as the two-dimensional pattern 270 by considering the size-reducing ratio, and the second pattern is formed into the aforementioned fine pattern 276, so that the phase difference (as shown at (a') and (b') in FIG. 2G) of 180 degrees is established in the individual transparent regions of the mask between a light beam 277 having passed through the phase shifting member 205 and a light beam 278 not having passed through the phase shifting member 205. These lights can interfere with each other in a region 280 between the two-dimensional pattern and the fine pattern to reduce the blur of the image to be projected on the wafer. As a result, the contrast of the projected image can be remarkably improved to improve the resolution and the focal depth drastically (as shown at (e) in FIG. 2G).
The following effects can be achieved from the mask according to these embodiments.
At the time of exposure, in the boundary required for the desired pattern to have precision, the first pattern and the second pattern are constructed such that the light having passed through the transparent regions of the first pattern and the light having passed through the transparent regions of the second pattern may interfere and be weakened. As a result, the blur of the contour of the image to be projected on the wafer can be reduced to improve the contrast of the projected image drastically thereby to improve the resolution and the focal depth remarkably. As a result, the resolution limit can be remarkably enhanced even with the same projection lenses and wavelength as those of the prior art. Even if the pattern on the mask is as complex and fine as the integrated circuit pattern, the pattern transfer precision is not partially dropped so that the transfer precision of the whole pattern formed on the mask can be remarkably improved.
Since, moreover, the two patterns are prepared to achieve the effect of the phase shift with the composed pattern, no transparent film exists on the mask surface so that no trouble arises in the inspections unlike the case of the prior art, in which the transparent film is formed on the mask.
Without the step of forming the transparent film, furthermore, the time period of fabricating the mask can be remarkably shortened from the mask in which the phase shifting means is exemplified by the transparent film on the mask substrate.
Although our invention has been specifically described in connection with the embodiments thereof, it should not be limited to those embodiments but can naturally be modified in various manners within the scope of the gist thereof.
According to the exposure method using the mask of the present invention, for example, there is neither restriction upon the specific structure of the system nor the structure of the aforementioned embodiment using the two split luminous fluxes, but the means may divide the luminous flux into a plurality and give them individual phase differences so that the patterns of the plural masks may be composed and exposed.
In the description thus far made, our invention has been described with respect to the technology of fabricating a semiconductor device according to the background field of application thereof but should not be limited thereto. The present invention can naturally be widely applied to the technological field of exposure, to which the form improving effect according to the phase shifting method can be applied.
The effects to be attained by the representative mode of the invention disclosed in the present embodiment will be briefly described in the following.
In order that the light having passed through the transparent regions of the first pattern and the light having passed through the transparent regions of the second pattern may interfere with each other and be weakened at the boundary in which the precision of the desired pattern is required, the first and second patterns on the mask are irradiated with two light beams having a phase difference and at least a partial coherence, and the transparent patterns of those light beams are composed to form the desired pattern on the specimen to be irradiated, so that the transfer precision of the boundary requiring the precision of the desired pattern can be improved.
The method of composing and exposing the ordinary main pattern, the main pattern for giving a phase shift of .lambda. or an equivalent phase shift or a fine shift (or accompanying) pattern on the two masks, as has been described in connection with Embodiments 1 and 2, will be called hereinafter the "multi-mask phase shifting method" or the "multi-mask phase inversion shifting method".
(Embodiment 3)
FIG. 3A shows a phase shifting mechanism 301 of the exposure system (of 1:5 size-reducing projection/step-and-repeat type) of Embodiment 3 of the present invention.
In the same Figure, the phase shifting mechanism 301 is constructed of an optical system which is interposed between a light source 302 of the exposure system and a specimen 303 (e.g., wafer) to be irradiated and which includes: a beam expander
304; mirrors 305, 307 and 308; half mirrors 306 and 313; an optical axis shifter 309; a corner mirror 310; an optical path varying mechanism 311 for driving the corner mirror 310 finely; a pair of relay lenses 312a and 312b; and a size-reducing lens system 315. In the alignment system of this optical system, there is positioned a mask 314 (or rectile) which is formed with the original image of a pattern to be transferred to the aforementioned specimen 303. The mask 314 is to be used in the process for fabricating the semiconductor integrated circuit device, for example, and the specimen 303 is a semiconductor wafer made of a single crystal of silicon, for example.
The light L such as the i-line (having a wavelength of 365 nm) emitted from the light source 302 is expanded by the beam expander 304 and is then refracted through the mirror 305 in a direction normal to the principal plane of the mask 314. After this, the refracted light is divided through the half mirror 306 disposed midway of the optical path into a straight light beam L.sub.1 and a light beam L.sub.2 advancing at a right angle with respect to the former. The light beam L.sub.2 is refracted through the mirror 307 and the corner mirror 310 so that it passes through a path different from that of the light beam L.sub.1 to irradiate another portion of the mask 314. The two light beams L.sub.1 and L.sub.2 thus having passed through the different portions of the mask 314 pass through the lenses 312a and 312b and are then composed into one light beam L' through the mirror 308 and the half mirror 313. After this, the light beam L' has its size reduced by the size-reducing lens 315
and is focused to irradiate the specimen 303 which is positioned on an X-Y table 316.
In the aforementioned phase shifting mechanism 301, the light beams L.sub.1 and L.sub.2 having passed through the half mirror 306 have different optical paths so that a desired phase difference can be established between the light beams L.sub.1
and L.sub.2 having reached the wafer 303 by changing the height (i.e., the optical path of the light beam L.sub.2) from the principal plane of the mask 314 to the corner mirror 310. The vertical movement of the aforementioned corner mirror 310 is accomplished by using the optical path varying mechanism 311, which may be a piezoelectric control element, for example.
FIG. 3A is an enlarged section showing the aforementioned mask 314. This mask 314 is made of transparent synthetic quartz 322 having a refractive index of about 1.47, for example, and has its principal plane formed with a metal layer 323 of Cr (chromium) having a thickness of about 500 to 3,000 .ANG.. At the time of exposure, the metal layer 323 provides the shielding regions A allowing no optical transmission, whereas the remaining regions provide transparent regions B allowing optical transmissions. The integrated circuit pattern is formed of the aforementioned shielding regions B and has a size of five times as large as the actual size (i.e., the size on the wafer), for example.
FIGS. 3C (a) and (b) show examples of the integrated circuit pattern formed on the aforementioned mask 314. The circuit pattern P.sub.1, as shown at (a) in the same Figure, is a portion of the composed pattern (c) after transfer and is extracted from the lower portion of the step of the specimen surface. The circuit pattern P.sub.2, as shown at (c) in the same Figure, is a portion of the composed pattern (c) after the transfer and is extracted from the higher portion of the step of the specimen surface. The patterns P.sub.1 and P.sub.2 are arranged in predetermined portions of the mask 314 and at a predetermined spacing. In FIGS. 3C (a) to (d): reference numeral 331 designates either a substrate of single crystal of Si or a semiconductor substrate of epitaxial (Si) layer; numeral 332 designates a SiO.sub.2 film; numerals 334a and 334b designate gate electrodes or wiring lines which are made of poly-Si, polycide, silicide or refractory metal; numeral 333 designates a positive type resist film applied to the film 332; letters B.sub.A and B.sub.C designate opening patterns over the main mask 314a; letters B.sub.B and B.sub.D designate opening patterns on the sub-mask 314b; letters P.sub.A and P.sub.C designate the positions on the resist film corresponding to the lower pattern; and P.sub.B and P.sub.D designate the positions on the resist film corresponding to the higher pattern.
Next, the method of preparing the aforementioned masks 314a and 314b will be briefly described in the following. First of all, synthetic quartz has its surface polished and rinsed, and a Cr film having a thickness of about 500 to 3,000 .ANG. is then deposited on the whole surface of the principal plane by the sputtering method. Subsequently, an electron beam resist is applied to the whole surface of the Cr film. Next, on the basis of the integrated circuit pattern data coded in advance in a magnetic tape or the like, the integrated circuit pattern is drawn on the electron beam resist by the electron beam exposure method. After this, the exposed portion of the electron beam resist is removed by the development, and the exposed Cr film is removed by the wet etching to form the integrated circuit pattern. The pattern data of the aforementioned paired circuit patterns P.sub.1 and P.sub.2 can be automatically prepared by expanding or reducing the data of the shielding regions A or the transparent regions B of one circuit pattern or by taking a logical product between the inverted data of one circuit pattern and the data of the other circuit pattern. For example, the pattern data of the circuit pattern P.sub.2 can be automatically prepared by taking a logical product between the data enlarged from the pattern of the transparent regions B of the circuit pattern P.sub.1 and the inverted data of the transparent regions B of the circuit pattern P.sub.1.
In order that the integrated circuit pattern formed on the aforementioned mask 314 may be transferred onto the wafer 303 (as shown in FIG. 3A), the wafer 303 having the photo resist applied to its surface is positioned at first on the X-Y table
316 of the exposure system shown in FIG. 3A, and the mask 314 (e.g., 314a and 314b) is positioned in the alignment system. The mask 314 is positioned such that, when one light beam L.sub.1 divided by the half mirror 306 is guided to irradiate one circuit pattern P.sub.1 of the aforementioned paired circuit patterns P.sub.1 and P.sub.2, the other light beam L.sub.2 is precisely guided to irradiate the other circuit pattern P.sub.2. Next, the corner mirror 310 is vertically moved to adjust the phase difference such that the two light beams L.sub.1 and L.sub.2 may have their phases reversed when composed again. At this time, the difference of the two optical paths is minimized by considering the interference distance of the light source. In order accomplish the positioning of the mask 314 and the adjusting of the phase difference of the two light beams L.sub.1 and L.sub.2 precisely, use is made of a pair of positioning marks M.sub.11, M.sub.12, M.sub.21 and M.sub.22 (all of which will be referred to "Mln"), which are formed on the mask 314, for example, as shown at (a) and (b) in FIG. 3E. These marks Mln are formed of openings which are arranged equidistantly in the shielding region, as hatched, and which have identical shapes and arrangements. In short, all the marks M.sub.11, M.sub.12, M.sub.21 and M.sub.22 have identical gaps and sizes. In case the positioning of the mask 314 (or 314a and 314b0 and the adjustment of the phase difference between the light beams L.sub.1 and L.sub.2 are precisely accomplished, the light beam L.sub.1 having passed through the mark M.sub.1 n and the light beam L.sub.2 having passed through the mark M.sub.2 n interfere with each other and completely disappear so that the mark images M.sub.1 and M.sub.2 are not formed on the wafer 303. In other words, by discriminating the existence of the projected images M.sub.1 and M.sub.2 on the wafer 303, it is possible to easily decide whether or not the positioning of the mask 314 (or 314a and 314b) and the adjustment of the phase difference between the light beams L.sub.1 and L.sub.2 are precisely accomplished.
The positioning of the patterns P.sub.1 and P.sub.2 on the mask is accomplished by using the alignment mechanism 309. Next, the phase difference is adjusted in a manner to correspond to the surface step (as shown in FIG. 3C) of the specimen 303. This adjustment is performed by controlling (or making a program to control) the piezoelectric control element of the optical path varying mechanism 311 with a computer. Specifically, since the focal point can be shifted to correspond to the phase difference, as shown in FIG. 3D, the specimen can be focused at its upper and lower portions even in case it has a surface step.
After the positioning of the mask 314 and the adjustment of the phase difference between the light beams L.sub.1 and L.sub.2 have thus been accomplished, the original image of the integrated circuit pattern formed on the mask 314 has its size optically reduced to one-fifth and is projected on the wafer 303. The operations thus far described are repeated while moving the wafer 303 sequentially stepwise.
According to the data of FIG. 3D, the transparent shifter layer on the mask is formed and exposed to give phase differences of 150 degrees, 180 degrees and 210 degrees by the on-mask phase shifting method, as shown in FIG. 1K. The experimental conditions are: the minimum pattern size of 0.35 .mu.m; the exposure wavelength of .lambda.=365 nm (i.e., i-line); NA=0.42; the partial coherency of .sigma.=0.3; the resist of "RI17000P of Hitachi Kasei"; and the exposure system of the 5:1 i-line stepper "RA101 of Hitachi, Ltd.".
Incidentally, the principle of the present invention can be realized not only by the pair-mask phase shifting method composing the aforementioned two mask patterns but also the on-mask phase shifting method of exposing one mask to a single luminous flux. In this case, the thickness of the phase shift film 22 of FIG. 1K has to be formed such that the phase difference .phi. may take a desired value ranging from 150 degrees to 210 degrees.
In the "phase shifting method", as has been described in the present embodiment, the method, in which projections are accomplished on a plurality of image planes by setting the shift at other than (2n+1).pi. (wherein n: an integer), will be called the "multi-image plane phase shifting method", and the method of using the two masks will be called the "multi-mask multi-image plane phase shifting method".
Incidentally, the exposure method, in which the simultaneous focusing operations are accomplished on the plural planes having the steps without being accompanied by the phase shift, as will be described in the following embodiments, and the present embodiment will be totally called the "multi-image plane projection exposure method".
(4) Embodiment 4
The present embodiment relates to a modification of the step-and-repeat type 5:1 size reducing projection exposure system (or stepper) which can be applied to Embodiments 1 to 3 and other embodiments, as will be described hereinafter. The present embodiment is effective in case the coherence length is relatively short because an exposing light having a low coherency is used from the requirement of the process or the like.
FIG. 4 is a schematic sectional front elevation showing the exposing optical system of the stepper according to the present embodiment. In the same Figure: reference numeral 402 designates an exposure light source such as the i-line (having a wavelength of 365 nm) of a mercury arc lamp or mercury xenon arc lamp, or an excimer laser (having a wavelength of 249 nm or 308 nm); numeral 403 designates a wafer to be exposed; numeral 404 designates an illuminating optical system including a beam expander, a condenser lens and so on; numeral 405 designates a mirror such as a cold mirror; numeral 406 designates a light dividing half mirror for dividing the light L into substantially equal halves; numerals 407a and 407b designate mirrors for reflecting the divided light beams L.sub.1 and L.sub.2, respectively; numeral 408 designates a corner mirror block for controlling the optical path of the light beam L.sub.1 and the positioning with the mask; numeral 408a designates a front corner mirror; numeral 408b designates a rear corner mirror; numeral 409 designates drive control means for the corner mirror block 408; numeral 410 designates a corner mirror block for controlling the optical path of the light beam L.sub.2 ; numeral 410a designates a front mirror; numeral 410b designates a rear corner mirror; numeral 414a designates a main mask; numeral 414b designates a sub-mask; numerals 412a and 412b front projection lens systems corresponding to the light beams L.sub.1 and L.sub.2, respectively; numeral 411 designates a drive control system for the corner mirror 410; numeral 413 designates a composing half mirror for composing the light beams L.sub.1 and L.sub.2 into the light L'; numeral 415 designates a rear projection lens system for focusing the composed light beam L'; and numeral 416 designates an X-Y stage and wafer sucking table for traversing the wafer 403 in the X and Y directions.
The operations of the present system are substantially identical to those of the aforementioned individual systems, and their repeated descriptions will be omitted here.
(5) Embodiment 5
The present embodiment relates to the step-and-repeat type 5:1 size-reducing projection exposure system which is characterized in that the optical distance from the main mask to the wafer and the optical distance from the sub-mask to the wafer are substantially equal, and in that the optical distance from the main mask to the light source and the optical distance from the sub-mask to the light source are substantially equal. However, it is natural that these characteristics are not essential to the present invention.
FIGS. 5A and 5B are a section showing the i-line exposure system of the present embodiment and an additional explanatory diagram showing a representative optical beam, respectively.
In these Figures, reference numeral 502 designates a light source which is constructed to include: an ultraviolet lamp such as a high-pressure Hg arc lamp or a Xenon Mg lamp; and a filter group or mirror for extracting only a substantially monochromatic i-line (having a wavelength of 365 nm) from the emission spectrum. Numeral 504 designates a condenser lens or a lens system composed of a group of plural lenses (made of synthetic quarts) to form a kohler illumination for the mask. Reference numeral 551 designates a first prism (made of synthetic quartz) applied to the half mirror face for adjusting the optical path, and numeral 506 designates a half mirror plane for dividing the exposing luminous flux L into the main exposing luminous flux L.sub.1 and an auxiliary exposing luminous flux L.sub.2. This half mirror is desired to have a substantially equal reflectivity and transmittance in an identical polarization mode. Reference numerals 507a and 508a designate the mirror surfaces for deflecting the main luminous flux L.sub.1 at an angle of 90 degrees; numeral 507b designates a mirror surface for deflecting the auxiliary luminous flux L.sub.2 at an angle of 90 degrees; and numerals 552a and 552b designate polarizing prisms (made of synthetic quartz) having individual evaporated mirror surfaces. Numerals 514a and 514b designate a main mask (or rectile) and an auxiliary mask (or rectile) having an exposed or transferred pattern, and numerals 561a and 561b designate a mask holder and fine drive means in the Z-axis (i.e., the direction of the optical axis) and in the X-Y directions. Numerals 540a and 540b designate phase difference setting means for setting the phase difference .phi. between the two luminous fluxes L.sub.1 and L.sub.2 by adjusting the optical length inbetween, and numeral 541 designates a communication pipe. Numerals 562a and 562b designate pre-projection lens groups; numeral 554 designates a second prism (made of synthetic quartz) for adjusting the optical path; numeral 553b designates a deflection prism (made of synthetic quartz) for deflecting the luminous flux L.sub.2 at an angle of 90 degrees; numerals 549a, 549b and 508b designate deflecting mirror surfaces; and numeral 513 designates a composing half mirror surface for composing the luminous fluxes L.sub.1 and L.sub.2 into the (composed) luminous flux L'. The half mirror 513 has characteristics similar to those of the aforementioned dividing half mirror 506. Numeral 515 designate an exposing post-projecting lens group; numeral 565 designates a referring post-projecting lens group; numeral 566 designates optical detection means disposed in the image plane of the referring projection lens group; numeral 503 designates an exposed wafer; numeral 576 designates a wafer chuck for retaining the flatness of the wafer by sucking the wafer by vacuum and a state for .theta. rotations (i.e., rotations on the vertical axis extending through the center of the wafer); numeral 577 designates a stage for moving in the Z-axis (i.e., vertical axis); numeral 578 designates horizontality adjusting means including three Z-axis drive means; numeral 579 designates an X-stage; and numeral 580 designates a Y-stage.
FIG. 5C is a section showing an essential portion of the phase difference setting means 540a of the aforementioned stepper. In the same Figure: reference numerals 542a and 543a designate synthetic quartz glass plates; numeral 541a designates means for adjusting their gap; numeral 544a designates a metal bellows; numeral 547a designates a pressure reservoir; numeral 546a designates a communication pipe made of austenite stainless steel pipe; and numeral 545a designates an optical path control chamber, in which the single or mixed gases having a refractive index different from that of the atmospheric gases in a chamber arranged with the stepper or the major atmospheric gases in the exposing luminous flux passage are held under a constant pressure. Incidentially, this optical path control chamber 545a or 547a can be evacuated by a vacuum pump. In case of this evacuation, it is unnecessary to consider the temperature rise of the gases in the optical path control chamber.
FIG. 5D is a top plan view showing the wafer stage portion of the aforementioned stepper. In the same Figure, numeral 503 designates a wafer to be exposed; numeral 576 designates a wafer chuck and .theta. stage; numeral 577 designates a Z state; numerals 578a to 578c designate individual Z-axis direction drive elements composing the horizontality adjusting means 578; numeral 579 designates an X-table; and numeral 580 designates a Y-table.
Next, the exposing operations of the present stepper will be described in the following. First of all, the optical path between the point on each mask corresponding to the exposure region and the light source is equalized as much as possible by adjusting the gradients of the main mask 514a and the auxiliary mask 514b. Moreover, the optical path between each mask and the corresponding point on the wafer 503 is equalized as much as possible (in terms of the gradient of the wafer). Next, as has been described in connection with Embodiment 3, the positioning mark M is used to accomplish the focusing, the mask registration in the X-Y plane and the adjustment of the phase difference to the phase difference .phi.=.pi. (after this, the phase difference (which may be a relative one as long as the interference is concerned) .phi. is readjusted to correspond to the step so that .phi. falls within the range from ##EQU1## After this, the exposure at the same site is executed.
The adjustment of the phase difference is executed by changing the thickness of the optical path control chamber 540a or 540b. Specifically, the distance between the quartz plate 542a and 542b is moved in parallel with one quartz plate.
Moreover, the gradient adjustment of each mask or wafer is executed by the movement in the Z-axis direction by three gradient adjusting means 578a to 578c (in case of the wafer, but by a similar mechanism in case of the mask), as shown in FIG.
5D.
The post-projection lens group 515 (as shown in FIG. 5A) itself has its two sides constructed of the "telecentric" such that the main luminous flux advances in parallel with the optical axis at the two sides of the same lens group. As a result, as in the infinite cylinder length correcting system of a telescope, it is possible to minimize the change of the focusing characteristics as a whole in case a variety of optical elements are interposed between the preprojection lens group 562a or 562b and the postprojection lens group 515. Since, moreover, the preprojection lens groups 562a and 562b are disposed separately of the post-projection lens group 515 in the vicinity of the masks 514a 514b, it is easy to retain the optimum object side numerical aperture.
(6) Embodiment 6
The present embodiment to be described is directed mainly to a mask pattern to be used in the invention, in which a main mask and an auxiliary mask are separately exposed to luminous fluxes and in which these luminous fluxes are composed to have a phase difference of (2n+1).pi. so that the wafer may be exposed to the composed light. In the following description, the main and auxiliary patterns corresponding to identical patterns (on the wafer) on the submask and the main mask are conveniently projected on a common plane, as shown. Moreover, the sizes attached to the identical patterns are converted to those on the wafer in case of the 5:1 size-reducing projection. For the auxiliary pattern, broken lines indicate the boundary between the shielding regions and the opening regions. The opening regions of the auxiliary patterns have their corresponding portions indicated by scattered points.
FIG. 6A shows the patterns of main and auxiliary masks in case isolated Al lines (or metal wiring lines, insulating film strips, strip-shaped openings, poly-Si wiring or gate lines, poly-cide wiring lines or gate lines, all of which are represented by the isolated Al lines) are to be exposed by the negative process. (In case linear openings are to be formed, it is naturally necessary to use the positive type resist process in the present mask patterns.) In the same Figure: reference numeral 601a designates that opening on the main mask, which corresponds to the Al line; numerals 604d and 605d designate shielding portions of a chromium film of the same main mask; and numerals 602b and 603b designate the auxiliary patterns (or shift patterns or compensating patterns, which are phase-inverted or merely inverted patterns or inverted slits) on the auxiliary mask. The size A is 0.3 to 0.4 .mu.m; the size B is about 0.2 .mu.m; and the size E is about 0.1 .mu.m.
FIG. 6B shows the pattern of the main mask and auxiliary mask of the present Embodiment B. This embodiment corresponds to contact holes, through holes or other isolated holes and uses the positive type resist process (although the negative type resist process is used in case of the isolated film pattern). In the same Figure: reference numeral 611a designates an opening corresponding to the hole (or opening) on the main mask; numeral 612d designates the shielding portion on the same main mask; and numerals 613b, 614b, 615b and 616b designate a group of inverted slits on the auxiliary mask. The sizes are substantially identical for the common symbols to those of the foregoing embodiment.
FIG. 6C shows a mask pattern corresponding the isolated openings or the like of the main and auxiliary masks of Embodiment 6C or a modification of the foregoing Embodiment 6B. In the same Figure, reference numerals 613C, 614C, 615C and 616C designate auxiliary opening patterns (or corner enhancement patterns or enhancers) for preventing the openings from being rounded, and the remaining numerals are wholly identical to those of the foregoing Embodiment 6B. The enhancers a square of about
0.1 .mu.m. This method is effective according to the foregoing Embodiment 6B for preventing the rounding of the corners from being extraordinarily enlarged.
FIG. 6D shows the main and auxiliary mask pattern in case the "L"-shaped opening pattern having its width corresponding to the minimum line width in said exposure process like the foregoing embodiments is to be treated by the positive resist process. In the same Figure: reference numeral 621a designates an opening on the main mask; numeral 622d designates the shielding portion (i.e., a portion of the shielding portion as the auxiliary mask like before. Namely, the portion other than the inverted shifter portion, as indicated by broken lines, is wholly the screening or shielding portion) on the main mask; and numerals 623b, 624b, 625b, 626b, 627b and 628b designate individual shifter region openings on the auxiliary mask. The sizes are designated at the same symbols as those of FIG. 6B. (These symbols designate the equal sizes, unless otherwise specified.) Incidentally, the present pattern is an isolated film pattern such as the "L"-shaped pattern if the negative type resist process is used.
FIG. 6E shows a modification 6E of the aforementioned Embodiment 6D. In the same Figure, reference numeral 621a' designates that opening pattern on the main mask, which corresponds to that 621a of FIG. 6D, and numeral 621d designates an auxiliary screening pattern (or a corner reduction pattern or a reducer) for preventing an excessive expansion inside of the corners of the "L"-shaped type opening on the same main mask. The sizes are equal to those of the enhancers. Numerals 623c,
624c, 625c, 626c and 627c designate opening patterns corresponding to the enhancers and formed on the main mask so as to prevent the excessive reduction of the corners; numeral 622d designates the shielding portion on the main mask; and numerals 623b,
624b, 625b, 626b, 627b and 628b designate individual shifter patterns (or inverted openings).
FIG. 6F designates main and auxiliary patterns corresponding to the negative type resist process of the isolated and bent Al wiring pattern of Embodiment 6F. In the same Figure: reference numeral 631a designates that opening on the main mask, which corresponds to the Al wiring line; numerals 638d and 639d designate the shielding portions on the main mask; and numerals 633b, 634b, 635b and 636b designate shifters running along the Al wiring line. The individual sizes are equal to the others on principle. This pattern can be applied to the formation of band-shaped openings if applied to the positive type resist process.
FIG. 6G shows the main and auxiliary mask pattern (corresponding to the negative process of the isolated Al bent pattern or the like) of Embodiment 6G. The present embodiment corresponds to a modification of the foregoing Embodiment 6F. In the same Figure, reference numeral 631c designates an opening pattern acting as the enhancers, and numeral 631d designates a shielding pattern acting as the reducer. Both of these patterns are formed on the main mask and have sizes equal to those of similar patterns of FIG. 6E. The remaining portions are absolutely identical to those of the foregoing Embodiment 6F.
FIG. 6H shows the main and auxiliary mask patterns for the line-and-space pattern of Embodiment 6H. In this case, the negative resist process is adopted. In the same Figure: reference numerals 641a, 642a and 643a designate those band-shaped opening pattern portions on the main mask, which correspond to the Al line patterns; numerals 641b, 642b and 643b designate those band-shaped shifter opening pattern portions (or complementary line patterns) on the auxiliary mask, which correspond to the Al line pattern portions; and numerals 645d, 646d, 647d and 648d designate shielding portions on the main mask. The sizes are 0.3 .mu.m for both the lines and spaces (as converted on the wafer). Incidentally, in the positive case, it is necessary, as shown, to replace the shielding portions between the opening on the main mask and the opening on the auxiliary mask by their adjacent openings. In other words, the opening of the main or auxiliary mask has to be located in the portion corresponding to the space. This location is identical to that of the case in which the periodic band-shaped openings are to be formed.
The mask patterns of Embodiments 6A to 6H can be applied not only to the aforementioned multi-mask systems (of Embodiments 1 to 5) but also to the on-mask phase shift (i.e., the phase shifting exposure method using one mask which is formed with both a shifter pattern having an inverted transparent film of a relative phase difference of .phi.=.pi. and a main pattern of .phi.=0 on one mask). In this case, the masks may be prepared while leaving the patterns of FIGS. 6A to 6H as they are on the masks.
(7) Embodiment 7
Here will be described the wafer processing and exposure process to be used in the embodiment of the present invention.
FIG. 7A is a top plan view of the wafer showing the exposure flow of the 5:1 size-reducing step-and-repeat projection exposure. In the same Figure: numeral 703 designates a wafer (e.g., wafer of 8 inches and single crystal Si) to be exposed; numeral 702 designates an orientation flat of the wafer; numerals 731 and 732 designate the exposed regions (i.e., the unit exposure regions which are optically irradiated by a single exposure action); and numerals 733 to 736 designate the individual unit exposure regions to be exposed. These regions cover the substantially whole area of the upper face of the aforementioned wafer 703.
FIG. 7B is a top plan view showing the relations among the unit exposure regions 733, individual chip regions 721 and 722, and an inter-chip region 723 in case of a memory IC.
FIGS. 7C to 7E and FIGS. 7F to 7H are schematic sections for explaining the exposure process using the positive and negative resists of the present invention and the flow of the wafer process. FIGS. 7C and 7F show optical diagrams and examples of the on-mask phase shift (i.e., the phase shifting method using one mask, which shows the main pattern only but omits the shifter) as to the mask. In the multi-mask case, the presentation is absolutely identical to this because the optical path is divided midway into two halves which are composed into one on the wafer surface.
In FIGS. 7C to 7E: reference numeral 714 designates a positive type mask; numeral 745 designates the opening of the mask 714; numeral 714 designates a size-reducing projection lens system which is shown in another embodiment; numeral 703
designates a wafer to be processed, which is sucked by the vacuum onto the wafer stage of the stepper; numeral 741 designates a first oxide film on the principal plane of the semiconductor wafer; numeral 742 designates an Al wiring pattern formed thereover; numeral 743 designates a second oxide film formed all over the former; and numeral 744 designates a positive type resist (which should be referred to Embodiment 16) film applied (to a thickness of 0.6 .mu.m) to the whole surface of the former by the spinner.
In FIG. 7D, reference numeral 746 designates an opening formed in a predetermined portion in the resist film 744.
In FIG. 7E, reference numeral 747 designates a through hole of the second oxide film, which is formed as the mask of the resist film 744.
In FIGS. 7F to 7H: reference numeral 714 designates a negative type mask; numeral 744 designates an opening or transparent pattern of the mask 714; numeral 715 designates the same size-reducing projection lens system as the foregoing one; numeral
703 designates the semiconductor wafer which is sucked onto the wafer stage of the stepper, as before; numeral 741 designates an oxide film formed over the principal plane of the wafer; numeral 742 designates an Al film applied to cover the whole surface of the oxide film by the sputtering; and numeral 754 designates a negative type photo resist film formed on (or applied to) the Al film and having a thickness of about 0.6 .mu.m.
In FIG. 7G, reference numeral 754x designates a patterned resist film.
In FIG. 7H, reference numeral 742x designates an Al wiring pattern which is patterned by using the resist film 754x as the mask.
FIGS. 7J to 7P are sections showing the process flow for fabricating a CMOS static RAM (i.e., SRAM) according to the twin well method, and FIG. 7Q is a layout showing the chip. The structure will be sequentially described in the following.
FIG. 7J shows a process for forming n- and p-wells according to the twin well process. In the same Figure: reference numeral 703 designates an n--type Si single crystal wafer (or substrate); numeral 760n designates the n-type well region; and numeral 760p designates the p-type well region.
FIG. 7K shows a subsequent gate forming process and a process for forming the source and drain of each FET by the ion implantation in a self-alignment manner by using the formed gate as the mask. In the same Figure: reference numerals 761a to
761c designate LOCOS oxide films; numerals 762p and 762n designate gate oxide films; numerals 763p and 763n designate polysilicon gate electrodes (of polycide); and numerals 764p and 764n designate p- and n-type highly doped source and drain regions, respectively.
FIG. 7L shows an inter-layer PSG film forming process and a second-layer poly-Si wiring line and high resistor forming process. In the same Figure: reference numeral 765 designates an inter-layer PSG film; numeral 766 designates a second-layer poly-Si wiring line; and numeral 766r designates a poly-Si high resistor acting as a load resistor of the SRAM memory cell.
FIG. 7M shows a flattening process by an SOG and a contact hole or through hole forming process. In the same Figure: reference numeral 767 designates an SOG film; numerals 768a, 768b, 768d and 768e designate contact holes with the Si substrate; and numeral 768c designates a through hole in the second-layer poly-Si wiring line and the upper layer.
FIG. 7N shows a first-layer Al wiring line forming process. In the same Figure, reference numerals 769a to 769e designate the first-layer Al wiring line.
FIG. 70 shows a process for forming an inter-layer insulating film over the first-layer Al wiring line and a second-layer Al wiring line forming process. In the same Figure, reference numeral 770 designates an inter-layer insulating film over the first-layer Al wiring line, and numerals 771a and 771b designate second-layer Al wiring line connected with the underlying Al wiring line through through holes.
FIG. 7P shows a process for forming a final passivation film over the second-layer Al wiring line. In the same Figure, reference numeral 772 designates a final passivation film.
FIG. 7Q is a top plan view showing the layout of the aforementioned SRAM at the unit of a chip. In the same Figure: numeral 721 designates the chip; numeral 722 designates a memory cell mat; and numeral 723 designates a peripheral circuit including an address decoder and a rear/write circuit.
FIG. 7I is a flow chart showing a process for photolithography in the aforementioned SRAM fabricating process, i.e., an exposure process in the flow form. In the same Figure, an n-well photo step 7P1 is one for forming a resist pattern over a Si.sub.3 N.sub.4 film (or substrate) so as to cover those other than the portion to be formed into the n-well, and a field photo step 7P2 is one for covering and patterning a photo resist film so that the Si.sub.3 N.sub.4 may be patterned to cover the P-channel and N-channel active regions. A p-well photo step 7P3 is one for patterning the resist film to cover the n-well so as to form the p-well channel stopper region, and a gate photo step 7P4 is one for patterning the resist film over the poly-Si or polycide layer covering the whole surface so as to pattern the gate electrodes 763p and 763n. The detail of the process till this step has been briefly described because it will be described in more detail with reference to FIGS. 8A to 8E. An n-channel photo step 7P5 is one for patterning the resist film at the p-channel side so as to implant the n-channel side with ions of n-type impurity by using the gate 763n as the mask; a poly-Si photo step 7P7 is one for forming a resist pattern over the poly-Si film covering all over the surface, so as to pattern the second-layer poly-Si film to be formed into the second-layer wiring line