JP2002162747A - Manufacturing method for three-dimensional structure by multistep exposure - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enlarge a region capable of changing the removing amount of photosensitive materials. SOLUTION: At the first step of the exposure, a resist 12 is exposed with a smaller amount of the light than that of designed in the amount for removing CEL layer and the surface of the resist 12 is removed up to the line 14 at this step (A). At the second step of the exposure, the resist is exposed with the amount of the light as much as deducted at the first step (B). The resist is removed up to the line 14a in total and the resist pattern after the development becomes as shown in 14b (C).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a density distribution mask (also referred to as a gradation mask (GM)) having a two-dimensional transmittance distribution in which the transmittance changes stepwise according to the target surface shape. The present invention relates to a method for producing an article having a three-dimensional structure surface by forming a three-dimensionally structured photosensitive material pattern on a substrate by exposing the photosensitive material pattern to the substrate, and engraving the photosensitive material pattern on the substrate.

[0002]

2. Description of the Related Art A special surface shape represented by a spherical surface or an aspherical surface has been used for a refracting surface or a reflecting surface of an optical element. In recent years, in connection with liquid crystal display elements, liquid crystal projectors, and the like, special surface shapes have been required for microlenses and the like.

In order to form the refraction surface and the reflection surface without using molding or polishing, a layer of a photoresist (a typical example of a photosensitive material) is formed on the surface of an optical substrate. Exposure through an exposure mask having a two-dimensional transmittance distribution, and a convex or concave surface is obtained as the surface shape of the photoresist by developing the photoresist. It is known that anisotropic etching is performed, and the surface shape of a photoresist is engraved on an optical substrate and transferred to obtain a desired three-dimensional structure of a refraction surface or a reflection surface on the surface of the optical substrate ( JP-A-7-230159
JP-A-8-504515).

In this case, a density distribution mask is used as an exposure mask used for obtaining a special surface shape of a three-dimensional structure such as a refraction surface or a reflection surface. Tokiohei 8-5
In the density distribution mask described in Japanese Patent No. 04515 (referred to as prior art 1), in order to form a two-dimensional transmittance distribution pattern, the mask pattern is divided into unit cells called light transmission openings, and The opening size of the cell is set to be a light transmission amount or a light shielding amount according to the height of a corresponding position of a photoresist pattern to be formed.

In the prior art 1, the surface shape of the photosensitive material layer is not smooth unless a process such as defocusing (DF: shifting the focus at the time of exposure from the surface of the photosensitive material) is performed. Unless a photosensitive material having low sensitivity (not available) is used, a three-dimensional film thickness distribution cannot be formed because the photosensitive material is exposed to the bottom of the substrate.

JP-A-7-230159 (prior art)
2), the shape of the dot pattern and the distribution of the dot density are calculated and calculated in accordance with the desired transmittance distribution, and the light of the dot pattern is emitted by the light source device capable of changing the output stepwise or continuously. The output of the output variable writing device that performs writing, writing is performed on the photosensitive material while changing according to the calculated dot pattern shape and the dot density, and the photosensitive material with the written dot pattern is developed, A method for obtaining an exposure mask having a two-dimensional desired transmittance distribution based on the dot pattern shape and the dot density distribution is described.

However, these conventional techniques have the following problems. In the related art, when a desired surface shape is formed using a concentration distribution mask, the light transmittance and the light shielding are reduced because a simple arrangement such as a unit cell square is used to produce a desired target shape. There is a lack of ideas on how to control the quantity globally. Since the density distribution mask is configured as an aggregate of unit cells, it is difficult to consider the adjacent effect of the light transmission amount (difference in the amount of light diffraction). Specifically, Prior Art 1 does not consider the adjacent effect at all. Therefore, if the pattern arrangement in the cell is partially or locally biased, a desired shape cannot be manufactured. This can be corrected empirically with a simple convex shape, but cannot be corrected with a shape having a discontinuous surface such as a Fresnel lens. On the other hand, in the related art 2, the adjacent effect can be leveled. Since the unit cell of the concentration distribution mask is manufactured using digital data (a pattern is formed only by the presence or absence of a light-shielding film), it is possible to form a smooth three-dimensional surface shape with respect to the high sensitivity characteristics of the photosensitive material. Absent. In particular, vertical surfaces such as Fresnel lens shapes cannot be manufactured. In the semiconductor field, it was an object to arrange the presence or absence of a line pattern two-dimensionally. Therefore, it was not intended to control a three-dimensional shape that changes in the height direction with high accuracy. In the case of having a discontinuous surface which rises rapidly like a Fresnel lens, it is difficult to form a vertical portion due to diffraction (wraparound of light) due to an adjacent effect. As another method, there is a method of manufacturing a concentration distribution mask called a HEBS glass method. This is a method in which silver (Ag) ions are scattered in a glass material, and this is blackened by a reduction reaction by electron beam irradiation to control the amount of transmitted light. Blackening occurs in proportion to the irradiation energy. As a method of dispersing silver ions, there are a method of introducing a silver ion component in advance at the time of manufacturing a glass material, and a method of introducing a silver ion component by ion exchange on the surface of the glass material. In the glass material according to the above method, silver is blackened by a chemical reaction inside the glass.
In the glass material according to the method described above, silver is blackened by a chemical reaction on the surface by electron beams. Both methods are the same in that the glass is irradiated with a desired pattern by electron beam irradiation to control the light transmittance. However, in any of these methods, in the case of the i-line stepper, the light transmittance T =
There is a problem that the number of gradations is small and the range is only 6.3 to 39%. For this reason, there is a problem that the three-dimensional structure that can be manufactured is limited.

[0008]

SUMMARY OF THE INVENTION An object of the present invention is to improve the above-mentioned problems of the prior art and to provide a novel method for forming a surface shape. The present invention solves the following problems. To demonstrate the concept of flexible process setting that can respond to the sensitivity characteristics of various photosensitive materials (different for each photosensitive material). Propose a process with high reproducibility, thereby reducing costs and improving yield. Specifically, to produce a three-dimensional structure having a smooth curved surface and a perpendicular surface, such as a Fresnel lens shape. In particular, to propose a process for making the inclination of the vertical plane 90 degrees as designed.

Here, the "sensitivity curve" of the photosensitive material is given as schematically shown in FIG. As a matter of course, the sensitivity curve differs depending on the photosensitive material. A is the coating thickness of the photosensitive material, B is the amount of the photosensitive material removed at the beginning of the exposure (this amount may be considerably small depending on the photosensitive material),
C is the irradiation light amount at which the photosensitive material starts to be exposed, the initial irradiation light amount (Eth) by the CEL (contrast enhanced lithography) effect, and D is the irradiation light amount required for removing the entire thickness of the photosensitive material, that is, the photosensitive material Exposure amount (Eop) when light is exposed to the bottom. Ordinary photosensitive materials for semiconductors are designed to be exposed at a certain exposure amount or more. Eth and Eop are characteristic values of the resist used by the photosensitive material manufacturer.

[0010] Since the light transmission coefficient differs depending on the molecular structure contained in the photosensitive material, the CEL amount and the sensitivity curve differ depending on the photosensitive material. When light travels through the photosensitive material, the light energy (light amount) depends on the depth
Decrease. That is, the thickness (depth) of the photosensitive material and the amount of irradiation light energy are in inverse proportion to each other and decrease in an exponential function. Therefore, when the “light transmittance” and the remaining film amount of the photosensitive material are obtained from the experimental data, it is possible to form a resist film thickness distribution having a distribution in the thickness direction of the photosensitive material.

The present invention does not form a two-dimensional line pattern of a photosensitive material having a certain height as in a semiconductor process. Instead, the present invention discloses a "structure having a three-dimensional shape, that is, a pattern having a controlled height. Object ". In the semiconductor field, there must be a certain amount of Eth (CEL effect), and the ratio of the light amount (Eop) for completely exposing the entire film thickness: Eth / Eop should be appropriately small (high sensitivity). It is required that the inclination (AB / CD inclination) is large. Further, it is also required that the removal amount (B) of the photosensitive material in Eth is large. The above are the target values of photosensitive material manufacturers in the semiconductor field and development themes.

However, in the method using the concentration distribution mask as in the present invention (referred to as the present method), it is desirable that Eth is small (the CEL effect is small). This is because the method of the present invention utilizes the photosensitivity of the photosensitive material between AB (or CD).
In addition, the resist of the photosensitive material removal amount B in the initial stage of the exposure is not only wasted (unusable area) by the present method, but also the photosensitive material is suddenly removed by the B amount when exposed to the light amount C. Occurs. A region where the photosensitive material removal amount can be changed by solving this point (between AB in FIG. 1)
It is an object of the present invention to attempt to broaden this.

[0013]

SUMMARY OF THE INVENTION The present invention proposes a method for solving the above-mentioned problems by using a density distribution mask method. According to the present invention, a photosensitive material is applied on a substrate, and the photosensitive material is coated using a density distribution mask having a two-dimensional transmittance distribution in which the transmittance changes stepwise according to a target surface shape. A method for manufacturing an article having a three-dimensional structure surface shape by exposing and developing a material to form a three-dimensional structure photosensitive material pattern on the substrate, and engraving the photosensitive material pattern on the substrate, The exposure is a multi-step exposure including a plurality of exposure steps with different masks, and one of the exposure steps includes a step of irradiating the photosensitive material layer with a light amount for removing the CEL layer of the photosensitive material. It is characterized by having.

The density distribution mask used for exposure is formed by forming a light-shielding pattern having a two-dimensional light intensity distribution on a transparent substrate, and is divided without gaps by unit cells having an appropriate shape and size. The light-shielding pattern in each unit cell is configured such that the light-transmitting amount or the light-shielding amount is set according to the height of the corresponding position of the photosensitive material pattern. A photosensitive material layer is formed thereon, and exposure is performed using the concentration distribution mask.

In the present invention, the above-mentioned "unit cells having an appropriate shape and size are divided without gaps, and the light-shielding pattern in each unit cell has a light transmission amount and a light-shielding amount corresponding to a desired shape. In the three-dimensional structure forming technology manufactured using the "density distribution mask set in", a plurality of pre-designed density distribution masks are prepared and a desired structure is obtained by sequentially exposing with this mask. Things.

More specifically, a photosensitive material layer is formed by applying a photosensitive material to a predetermined thickness on the surface of a substrate material on which a desired special surface shape is to be formed, and the photosensitive material layer is formed according to the desired surface shape. Then, a mask pattern is exposed on the photosensitive material layer by using a concentration distribution mask in which the transmittance changes stepwise, which is designed in advance by a separate method. That is, a photosensitive material is applied on a target substrate material, the photosensitive material is exposed, and the photosensitive material is exposed using a concentration distribution mask used for patterning into a final target three-dimensional structure.

The present invention includes the following steps. (Step 1) As a preparatory step, a photosensitive material to be used is irradiated with light of various intensities, and a sensitivity curve (characteristic) of the material is measured in advance. (Step 2) From this result, an initial light removal amount (CEL) specific to the photosensitive material is obtained.

(Step 3) A plurality of exposure masks are prepared and the photosensitive material is exposed in multiple stages. Here, the amount of exposure to be irradiated is represented by the following relationship: exposure amount required for patterning a target structure = first-stage exposure amount + second-stage exposure amount +... + N-th-stage exposure amount. The exposure amount at each stage is determined so as to satisfy this equation. Here, the “exposure amount required for patterning the target structure” means that the applied photosensitive material is securely exposed to the substrate surface, and after development and rinsing, an independent pattern not connected by the photosensitive material is formed. The amount of light irradiation required for

Here, for example, in the case of two-stage exposure, the exposure required for patterning a target structure = the first-stage exposure + the second-stage exposure. In other words, (first-stage exposure amount) = (exposure amount required for patterning a target structure)-(second-stage exposure amount). Here, assuming that the second stage exposure amount = the light amount for CEL removal, (first stage exposure amount) = (the exposure amount required for patterning a target structure) − (the light amount for CEL removal). That is, the first-stage exposure amount is a light amount obtained by subtracting the “light amount for CEL removal” from the “exposure amount required for patterning”.

Here, since the "light amount for CEL removal" is a value unique to the photosensitive material, the first-stage exposure amount can be easily obtained by calculation even if the resist changes, and is necessarily determined. .

In the case where three or more exposure steps are performed, the first-step exposure is divided and exposed as necessary. In the present invention, the “exposure for removing the CEL layer” step is provided independently, but there is no reason to limit this step to one exposure step. That is, in this step, the light amount required for CEL removal can be divided into two or more times for exposure.

The concentration distribution mask method has the following advantages over the resist thermal deformation method. (1) In a micro-lens array (MLA) having a very small pitch in which the distance between adjacent pixels is as close to zero as possible, the heights of adjacent portions are different in adjacent tangential cross sections (shown in FIG. 9D). In FIG. 9, (A) is a partial plan view of a micro-lens array having a fine pitch, (B) is a cross-sectional view taken along the line AA, and (C) is a cross-sectional view taken along the line BB. Sectional view,
(D) is a cross-sectional view taken along the line C-C, which is called an adjacent tangential cross-section. However, it is attempted to fabricate such a micro-pitch micro-lens array with a conventional resist thermal deformation method. However, the MLA structure cannot be manufactured as designed. In the conventional resist thermal deformation method, the interval is reduced by applying the resist multiple times,
Alternatively, the height of the adjacent cross section can be made close to the target value, but it is practically impossible to make the interval between adjacent MLAs zero to form an independent (isolated) resist block. (2) The Fresnel lens cannot be formed by the conventional resist thermal deformation method. (3) In the conventional resist thermal deformation method, the diameter is 500 μm.
The extent was the limit of large aperture lenses. (4) The density distribution mask method can easily produce an aspherical shape. (5) A deformed lens such as a toroid can be easily manufactured. (6) A monotonically increasing convex structure represented by a prism, a pyramid or the like can be easily manufactured. (7) Complex structures such as micromachining can be easily manufactured.

[0023]

DESCRIPTION OF THE PREFERRED EMBODIMENTS In the first stage of exposure, the exposure is smaller than the exposure amount at the time of design (the exposure amount required for patterning a target structure). It is preferable to include a step of additionally exposing the exposed light amount.

The amount of light to be exposed in the first stage is the amount of light for removing the CEL layer from the amount of exposure at the time of design (claim 2). In this case, the mask used for the exposure for CEL removal includes a uniform GM pattern No. (the GM pattern is a light shielding film pattern of a unit cell, and a GM pattern No. A batch exposure can be performed with a uniform density mask in which the transmittance corresponds to the transmittance of the unit cell.

In the present invention, if the technical idea of "irradiating the light amount for CEL removal" is adhered to, the number of irradiations and the combination of the irradiation amounts are within the range of selection by a technician in process setting. is there. Therefore, even if the first-stage exposure process is performed with a light exposure smaller than the designed exposure amount, there is no reason to define the first-stage exposure as one exposure. That is, this exposure step may be divided into two or more exposure steps for exposure.

A preferred GM pattern arranging method for realizing the present invention is a method in which the amount of light to be exposed for removing the CEL layer is taken into consideration. In other words, since it is necessary to irradiate a small amount of light for exposure to remove the CEL, the GM mask pattern arrangement for manufacturing a desired structure has a smaller number than the GM pattern No. obtained from the sensitivity curve (ie, , Light transmission amount is small). For the same reason, the GM pattern No. corresponding to the uppermost part of the desired structure is set to the zeroth (all black: zero light transmittance) (claim 5). By setting the number to zero, the resist remains without being removed at all under the zero number pattern (C
All the EL layers remain. ). “CEL layer removal exposure” is performed to remove this and form a desired structure.

As described above, in order to produce a desired structure as a mask used in the first-stage exposure, the light transmission amount is set to be smaller than the GM pattern No. obtained from the sensitivity curve of the photosensitive material. By using the mask constituted by the GM pattern No., it is possible to perform the exposure while keeping the light source light amount of the exposing machine constant between the exposure in the first stage and the exposure in the second and subsequent stages.

The conventional method has the following problems. The Fresnel lens shape has an inflection point (the height suddenly increases in a portion where the number of lens zones changes). When this is performed by the conventional method, the rising angle does not become 90 ° due to the adjacent effect. Due to the adjacent effect, the height of the lens is reduced in the peripheral portion of the lens (the portion after the two orbicular zones). In the case of a simple convex shape, the shape may be narrowed by side etching of the vertex of the lens. When the above method is used, the rising angle is close to 90 °, the lens height is equal, and the side etching is prevented or reduced. As a result, a three-dimensional microstructure and a Fresnel lens structure can be easily manufactured. When a structure composed of a photosensitive material is transferred to a substrate material using a dry etching method, a three-dimensional structure having the same material as the substrate material can be obtained. If the substrate material is a glass material, it can exhibit optical performance,
If a material such as Si is used, a micro machine or the like can be manufactured. When a three-dimensional structure is transferred by a dry etching method, a dry etching component slightly exists in a direction parallel to the substrate surface. This etching component breaks the shape as the photosensitive material is exposed to the plasma for a longer time, but reduces the effect of side etching by leaving the photosensitive material slightly at the top of the desired shape. It becomes possible to do.

The present invention will be described more specifically with reference to the drawings. In the following, a specific resist will be described. Naturally, the CEL and the sensitivity curve differ depending on the resist, so the following description is an example. FIG. 2 shows measured values of a sensitivity curve when exposure is performed in two stages. The horizontal axis is the GM pattern No. corresponding to the light transmittance of the unit cell, and the light transmittance increases linearly as the numerical value increases. The vertical axis represents the resist removal amount after exposure and development, and X represents the removal amount when the entire resist film thickness is removed.

Assuming that the exposure required for patterning the target structure is W = 1000 mJ and the second-stage exposure is W = 240 mJ for the CEL removal amount, the first-stage exposure is W = 240 mJ.
= 760 mJ. In the first stage exposure, the relationship between each GM pattern and the amount of resist removal is represented by EFGH lines. Since no development is performed at this stage, the resist is not actually removed, but the corresponding resist removal amount is as shown in the figure. In the unit cell having the transmittance equal to or higher than the GM pattern 82, the light amount is suppressed, and the unit cell is saturated at the resist removal amount Y.

Next, in the second stage of exposure, the entire surface of the resist is exposed with an equal amount of light using a mask in which GM pattern No. 42 is uniformly arranged on the entire surface. The relationship between each GM pattern and the resist removal amount at the end of the second stage exposure is IJKLM.
It is represented by the N line. If you develop at this stage,
A three-dimensional resist pattern from which the resist has been removed in the relationship represented by the IJKLMN line is obtained.

Referring to the results of FIG. 2, the exposure is divided into two stages, the amount of resist removed by the second stage exposure increases, the graph moves upward, and the resist film thickness that can be patterned is Y. To X. Further, the number of gradations that can be patterned by the first-stage exposure is 36 to FG.
The gradation (82-46) is increased to 49 gradations (91-42) between J and M by the exposure at the second stage.

FIG. 3 conceptually shows a change in the resist pattern in each of the two exposure steps. (A) shows the state after the first-stage exposure. Reference numeral 10 denotes a substrate, 12 denotes a resist layer, and 14 denotes a region exposed by the first-stage exposure. If development is performed at this stage, the resist 12 is removed only from the surface side to the portion indicated by the line 14. The region below the GM pattern No. 46 is not removed by the CEL effect, and even in the region above the GM pattern No. 82, the resist removal range does not reach the substrate because the amount of light is suppressed.

(B) shows a state after the second stage exposure. By irradiating the entire surface with the light amount corresponding to the GM pattern No. 42, the resist removal region as a whole is advanced to the position indicated by the line 14a. (C) shows the resist pattern 14b after development.

[0035]

DESCRIPTION OF THE PREFERRED EMBODIMENTS (Design of Density Distribution Mask) An example of a fine pitch MLA in which the distance between adjacent microlenses is made as close as possible to zero is shown. 0.9 "in MLA for liquid crystal projector
The pixel size for XGA is 18 μm × 18 μm. In this MLA, if there is a lens non-forming portion of 1 μm on each side of the lens, 17 μm × 17 μm
And the ML occupying the entire area
A area is 17 × 17/18 × 18 = 289/324 =
It is 0.89, and even if all the light can be effectively collected by the MLA, the light collection efficiency is only 89%. In other words, it is important to reduce the area of the non-formed portion of the MLA in order to improve the light use efficiency. In this embodiment, the non-formed portion of the MLA is made close to zero.

Specifically, when a 1/5 (reduced) stepper is used, the actually manufactured density distribution mask reticle pattern dimension is 90 μm × 90 μm. This one
MLA are divided into 3.0 μm unit cells, and the height × width = 3
It is divided into 0 × 30 (pieces) = 900 (pieces) unit cells.

Next, in the central 2 × 2 unit cell (6 μm × 6 μm on the density distribution mask reticle, 1.2 μm × 1.2 μm in the actual pattern), cell No. 1 (all chromium remains) Deploy. Cell No. 80 (no chrome remaining portion) is arranged at the four corners of the lens. The “opening area” corresponding to each “gradation” is made to correspond to the cells of No. 1 to No. 80 during this time. This relationship is obtained from the exposure process and the resist sensitivity curve. Of course, it is necessary to grasp the sensitivity curve each time the resist material or process is different. Thus, CAD data of the MLA density distribution mask reticle is created.

(Production of Concentration Distribution Mask Reticle) The CAD data created as described above is irradiated with a laser beam using a laser beam irradiation apparatus (a product of Ricoh Optical Co., Ltd.) shown in FIG. Done. In this laser beam irradiation, an optimal beam shape can be determined according to a desired shape, and a polygonal shape, a circular shape, or the like can be shaped by an aperture.

The laser beam irradiation device shown in FIG. 4 includes a laser beam oscillator 1, a beam splitter 2 for dividing the laser beam from the laser beam oscillator 1 into a plurality of laser beams, a mirror 3 for bending the optical path of the laser beam, and a mirror. 3
The optical modulator 4 modulates the laser light bent by the optical modulator 4. The optical modulator 4 controls the optical modulator 4 by a signal from the data bus to control ON / OFF of each laser light. A light deflector 6 for deflecting the laser light, an objective lens 7 for condensing the laser light on the resist material layer, and a mask blank placed thereon in the X direction and Y direction.
It is composed of main components such as an XY stage 8 that moves in the direction, and a control device 9 that controls the operation of the optical deflector 6 and the operation of the XY stage 8.

This laser beam irradiation apparatus controls the operation of the XY stage 8 according to the design data, and controls the ON / OFF and deflection of each laser beam so that a desired mask is formed on the resist material layer of the mask blank. Draw a pattern. That is, the resist material layer is irradiated with laser light by the laser light irradiation device so that the light transmitting region or the light shielding region is set to have a desired transmittance distribution for each unit cell.
Pattern formation is performed in a two-dimensional manner. At this time, the irradiation of the laser beam is controlled in accordance with the transmission amount distribution of each unit cell calculated according to the desired special surface shape, and the light transmitting area or the light shielding area in each unit cell is increased or decreased. The arrangement of the dots is controlled. An appropriate unit cell shape and dot shape may be selected depending on the target product.

The CAD data created as described above is installed in the laser beam irradiation device shown in FIG.
The density distribution mask blanks were exposed by a predetermined method while controlling the ON / OFF of the XY stage and the laser beam, and the beam shape by the aperture. Then, development and rinsing were performed by a predetermined method to pattern the resist material layer. Thereafter, the Cr film was patterned by dry etching.

By using a laser beam drawing method and making the drawing beam shape rectangular and increasing the drawing area spirally, higher reproducibility than the electron beam drawing method can be obtained. When the drawing area is circular, the laser beam drawing method can obtain extremely high reproducibility when the diameter of the drawing area is 0.2 μm or more. When the diameter of the drawing area is smaller than 0.2 μm, the reproducibility deteriorates. However, in the electron beam drawing method, when the dimension of the drawing area is smaller than 0.5 μm, the reproducibility deteriorates. Is excellent.

The prediction of the "adjacent effect" depends on the shape of the unit cell. When the unit cell shape is a square or a rectangle, it can be accurately drawn by a rectangular dot, so that the adjacent effect can be predicted by calculation.

(Specific Example of Fabricating Concentration Distribution Mask Reticle) Fabrication of MLA for Sparse Liquid Crystal: In producing concentration distribution mask reticle, TGMR-950BE of a positive resist material is used as a resist material which is a photosensitive material.
(A product of Tokyo Ohka Co., Ltd.).

The density distribution mask is composed of unit cells divided into squares, and the amount of light transmission or light shielding in each unit cell is controlled. Of course, an optimum unit cell may be determined according to a desired shape and manufactured with optimum dots. Here, for the sake of simplicity, the description will be made using a square. As a method of controlling the amount of light transmission, there is a method of combining the control of the Cr opening area and the control of the Cr film thickness. here,
Method was adopted.

In order to manufacture a concentration distribution mask reticle, a Cr film having a thickness of, for example, 200 nm is formed on a transparent glass substrate, and the above-mentioned resist material is applied thereon. The resist material was irradiated with laser light using the laser light irradiation apparatus shown in FIG. 4 to perform drawing. Thereafter, a mask pattern was formed on the resist material layer through development and rinsing, and the Cr film was dry-etched using the resist pattern as an etching mask, thereby patterning the Cr film to produce a concentration distribution mask.

FIG. 5 shows an example of a typical arrangement of density distribution masks for a microlens of 20 μm × 20 μm. The unit cell has a square shape like a grid. The unit cell does not necessarily have to be a square, but is desirably another polygonal shape according to a desired shape. The hatched portion is the portion where the Cr film remains.

(Specific Example of Manufacturing Micro-Dimensional MLA for Liquid Crystal) Using the density distribution mask reticle (shown in FIG. 5) of the specific example of manufacturing the density distribution mask reticle as a mask, a reduction projection exposure apparatus (1/1) shown in FIG. An example of an MLA for a liquid crystal projector manufactured by exposing using a 5 stepper) to form a resist pattern and transferring it to a material for an optical device will be described.

First, the reduced projection exposure apparatus will be described. Light from the light source lamp 30 is condensed by the condenser lens 31 and irradiates the exposure mask 32. The light transmitted through the mask 32 is incident on an imaging lens 33 having a reduction magnification, and a reduced image of the mask 32, that is, a reduced image of the transmittance distribution, is formed on the surface of the optical device material 37 placed on the stage 34. Is imaged. The stage 34 on which the optical device material 37 is placed is moved by the action of the step motors 35 and 36.
In a plane perpendicular to the optical axis of the imaging lens 33, the optical device can be displaced in two directions perpendicular to each other, and the position of the optical device material 37 can be aligned with the optical axis of the imaging lens 33. I have.

The reduced image of the mask 32 by the imaging lens 33 is formed on the surface of the photoresist layer of the optical device material 37. This exposure is performed densely over the entire surface of the optical device material 37.

A density distribution mask reticle was manufactured. Here, the mask was designed using GM pattern Nos. 42 to 91 as an example as shown in FIG. As shown in FIG. 2, in the exposure in the first stage, the sensitivity is only up to 82, but by performing the exposure in the second stage, the range of sensitivity is widened. This is the same as that the sensitivity curve changes when the exposure amount is changed.
At the same time, the thickness of the resist that can be exposed increases.

The exposure was performed with a 1/5 stepper shown in FIG. The following exposure conditions were continuously performed by a stepper device. First, an example is shown in which exposure is performed using a GM mask pattern for producing a target shape. The first stage is and the second stage is. GM pattern mask Defocus: +5 μm, irradiation amount: 390 mW × 0.8
0 second GM pattern mask Defocus: +1 μm, irradiation amount: 390 mW × 0.2
8 seconds # 42 mask on the entire surface Defocus: +5 μm, irradiation dose: 390 mW × 1.0
A plus sign in the display of the 2 second defocus amount means that the focal point is above the resist surface. Under these conditions, the first stage exposure amount is 390 mW × 1.08 seconds (illuminance: 421 mJ), and the second stage exposure amount is 390 mW × 1.02 seconds (illuminance: 3
98 mJ). The total exposure is about 820 mJ.
It is. Of course, there is also a method of first exposing with an entire mask. In the case of this embodiment, when the exposure was performed in the order of →→, the exposure amount at the tip of the MLA was increased, and the curvature of the tip was increased. In other words, the height has been reduced. )

After exposure under these conditions, the photosensitive material PEB
(Post exposure bake) was performed at 110 ° C. for 480 seconds. Next, the photosensitive material was developed and rinsed. Thereafter, the resist was hardened by evacuating while irradiating ultraviolet rays for 180 seconds with an ultraviolet curing device. The ultraviolet curing device irradiates a wavelength that can cure the resist at a wavelength shorter than the wavelength used for exposing the resist. By this operation, the plasma resistance of the resist is improved, and the resist can be processed in the next step. The resist height at this time is 7.0 μ
m. At this time, the initial coating resist thickness is
It was 8.5 μm. Therefore, the thickness of the CEL layer was approximately 1.5 μm.

Due to the defocusing effect, the shape could be manufactured without any particular step. Also,
The number of mask gradations could be increased, and the CEL layer was also removed from the entire black pattern portion disposed on the mask, and the desired good curvature was obtained.

Thereafter, the substrate was set in a TCP (inductively coupled plasma) dry etching apparatus, and the degree of vacuum was set as follows:
1.5 × 10 ⁻³ Torr, CHF ₃ : 5.0 sccm, C
F ₄ : 50.0 sccm, O ₂ : 15.0 sccm, substrate bias power: 600 W, upper electrode power: 1.25 kW,
Dry etching was performed at a substrate cooling temperature of -20 ° C. At this time, the etching was performed while changing the substrate bias power and the upper electrode power over time, and changing the selectivity to decrease with time. The average etching rate of the substrate was 0.67 μm / min,
The actual etching time required was 10.5 minutes. The lens height after the etching was 5.3 μm.

(Manufacture of Fresnel Lens) The following products can be mentioned as products in which the density distribution mask method is most effective. (1) With the resist thermal deformation method, it was impossible to form a Fresnel lens having a very small size. → The density distribution mask technology makes this possible. (2) In the resist thermal deformation method, the lens size that can be manufactured is limited, and the diameter of about 500 μm is the limit of the large diameter lens. → The large-diameter lens can be manufactured by the density distribution mask method. (3) A lot of data accumulation and know-how were required to produce an aspherical shape. → The density distribution mask method can produce various shapes.

The density distribution mask method has the features described above. An example of a Fresnel lens is given as one of the shapes that best express these features. In this embodiment, an eight-zone Fresnel lens having a focal length: F = 6.25 (mm) and a numerical aperture: NA = 0.4 (10λ gap) was manufactured as the Fresnel lens.

FIG. 7 shows a part of the Fresnel lens 42 on the right side from the optical axis O, where t is the thickness of the substrate and h (n) (n = 1, 2, 3,...). ) Is the surface height from the optical axis O to each annular zone, ΔZ (n) is the sag amount representing the depth from the surface of each annular zone to the valley, and θ is the inclination angle.

[0059] Here, aspheric expression representing the curved shape of each annular zone of the Fresnel lens: Z (n) is, Z (n) = ch ( n) 2 / {1+ (1- (K + 1) c 2 h
(n) ² ) ^1/2 } + Ah (n) ⁴ + Bh (n) ⁶ + Ch (n)
⁸ + Dh (n) ¹⁰ + Eh (n) ¹² + Fh (n) ¹⁴ + Gh
(n) ¹⁶ and the surface height: h (n) is h (n) = (X ² + Y ² ) ^1/2 , and the inclination angle θ is θ = arctan (dz / dh). There, slope dz / dh is, dz / dh = ch (n ) / {(1- (K + 1) c 2 h (n)
² ) ^1/2 } + 4Ah (n) ³ + 6Bh (n) ⁵ + 8Ch
(n) ⁷ + 10Dh (n) ⁹ + 12Eh (n) ¹¹ + 14Fh
(n) ¹³ +16 Gh (n) ¹⁵ . Design wavelength λ = 980 (nm), substrate thickness t = 1 ± 0.05 (mm) (parallelism: 0.1 μm or less at φ5 mm), refractive index of substrate n = 1.511118, back focus BF = 5.588268. Also, spherical equation = ch (n) ² / {1+ (1−c ² h (n) ² )
^1/2 }. FIG. 8 shows the result of calculating the sag amount of the ring zones 1 to 8. Based on the above, the lens was designed and manufactured so that the height of each of the zones of the Fresnel lens of the eight zones was the same as 21.0 μm.

(Specific Example of Fresnel Lens Manufacturing) Based on the above lens design, the arrangement of the density distribution mask pattern No. is determined from the sensitivity curve and the calculation formula, and the CAD data of the CAD data is determined using the laser beam irradiation apparatus shown in FIG. I drew a pattern.
After that, a mask pattern is formed on the resist material layer through development and rinsing, and the Cr film is dry-etched using the resist pattern as an etching mask.
The Cr film was patterned to produce a concentration distribution mask. Here, one density distribution mask reticle was manufactured. Here, the mask was designed using GM pattern Nos. 42 to 91 as an example as shown in FIG.

Next, an example of a Fresnel lens manufactured using the density distribution mask reticle will be described. An SF-6 glass substrate was prepared and TGMR-95 described on this substrate was prepared.
An OBE resist is applied to a thickness of about 8.5 μm. Next, prebaking was performed at 100 ° C. for 180 seconds using a hot plate.

The substrate was exposed with a 1 / 2.5 stepper using a density distribution mask reticle. Exposure was performed under the following exposure conditions. The first stage is
Then, the second stage is. GM pattern mask Defocus: +5 μm, irradiation amount: 390 mW × 0.4
5 seconds GM pattern mask Defocus: +3 μm, irradiation amount: 390 mW × 0.3
5 seconds GM pattern mask Defocus: +1 μm, irradiation amount: 390 mW × 0.2
8 seconds # 42 mask on the entire surface Defocus: +5 μm, irradiation dose: 390 mW × 1.0
2 seconds

Under these conditions, the first-step exposure amount was 390 mW × 1.08 seconds (illuminance: 421 mJ), and the second-step exposure amount was 390 mW × 1.02 seconds (illuminance: 3
98 mJ). The total exposure is about 820 mJ.
It is. Of course, there is also a method of first exposing with an entire mask. In the case of this embodiment, when exposure is performed in the order of →→→, the exposure amount at the Fresnel lens tip becomes large,
The tip became rounded.

After exposure under these conditions, PEB was performed at 110 ° C. for 25 minutes. Next, the photosensitive material was developed and rinsed. Thereafter, the resist was hardened by evacuating while irradiating ultraviolet rays for 480 seconds with an ultraviolet curing device. The resist height at this time is 6.
It was 7 μm.

At this time, the initial applied resist thickness is:
It was 8.2 μm. Therefore, the thickness of the CEL layer was approximately 1.5 μm. In the above MLA embodiment, about 1.
5 μm, but from other experiments 1.5 to 1.7 μm
It seems to be m-th place. Of course, it varies slightly depending on the process conditions. Due to the defocus effect, the shape could be manufactured without any particular step. In addition, the number of mask gradations could be increased, and the CEL layer was also removed from the entire black pattern portion disposed on the mask, so that a desired desired Fresnel shape was obtained.

The substrate was set in a TCP dry etching apparatus, and the degree of vacuum was 1.5 × 10 ⁻³ Torr, CHF ₃ :
20.0 sccm, BCl ₃ : 1.0 sccm, CF ₄ : 1
Dry etching was performed under the following conditions: 5.0 sccm, O ₂ : 0.9 sccm, substrate bias power: 1.2 kW, upper electrode power: 1.25 kW, and substrate cooling temperature: −20 ° C. At this time, the etching was performed while changing the substrate bias power and the upper electrode power with time, and changing the selection ratio to increase with time. The average etching rate of the substrate was 0.84 μm / min. The selectivity is 3.15 and the lens height after etching is 2
It was 1.1 μm. The actual etching time is 26.
It took 5 minutes. As a result, the curvature of each wheel has a shape according to the design value, and the rising angle of each wheel is also 8 on average.
A steep slope of 8.6 ° could be obtained. Also, the gap distance between the rising portions between the ring bodies was extremely small, at most about 1 μm, and the optical performance as designed could be obtained.

[0067]

According to the present invention, the exposure for forming a photosensitive material pattern having a three-dimensional structure to be transferred onto a substrate is a multi-step exposure including a plurality of exposure steps using different masks. Exposure is performed with a light amount smaller than the exposure amount at the time of design, and in the second and subsequent stages of the exposure, the light amount that was less exposed in the first stage is additionally exposed, so that the sensitivity characteristics of various photosensitive materials (photosensitive Flexible process that can be used for different materials).
In addition, the reproducibility of the process is improved, whereby the cost can be reduced and the yield can be improved. When the three-dimensional structure is transferred by the dry etching method, a slight dry etching component in the direction parallel to the substrate surface exists, but the photosensitive material is slightly applied to the apex portion of the desired shape as in claims 5 and 6. By leaving it, the influence of side etching can be reduced, and deformation of the shape of the transferred three-dimensional structure can be suppressed.

[Brief description of the drawings]

FIG. 1 is a graph schematically showing a “sensitivity curve” of a photosensitive material.

FIG. 2 is a graph showing how a sensitivity curve of a photosensitive material changes by two-step exposure according to the present invention.

FIG. 3 conceptually shows a change in a resist pattern for each exposure step in the two-step exposure.

FIG. 4 is a schematic configuration diagram showing an example of a laser beam irradiation device used for manufacturing a concentration distribution mask reticle.

FIG. 5 is a diagram showing an example of a light-shielding pattern of a density distribution mask for microlenses.

FIG. 6 is a schematic configuration diagram illustrating an example of a reduction projection exposure apparatus.

FIG. 7 is a sectional view showing a part of a Fresnel lens.

FIG. 8 is a diagram showing design data of a Fresnel lens.

9A is a partial plan view of a micro-lens array having a minute pitch, FIG. 9B is a cross-sectional view taken along line AA, and FIG. 9C is a cross-sectional view taken along line BB. , (D) are cross-sectional views taken along the line CC.

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 laser light oscillator 2 beam splitter 4 light modulator 5 light modulation controller 6 light deflector 7 objective lens 8 XY stage 10 substrate 12 resist 14, 14 a resist removal area 14 b resist pattern

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G03F 7/095 501 G03F 7/095 501 F Term (Reference) 2H025 AA03 AB14 AB20 AC01 AC04 AC08 AD01 AD03 DA40 FA04 2H095 BB01 BB31 2H097 AA12 BB01 JA02 LA17

Claims (6)

[Claims] 1. A photosensitive material is applied on a substrate, and the photosensitive material is applied using a density distribution mask having a two-dimensional transmittance distribution in which the transmittance changes stepwise according to a target surface shape. Forming a three-dimensionally structured photosensitive material pattern on the substrate by exposing and developing a photosensitive material, and engraving the photosensitive material pattern onto the substrate to produce an article having a three-dimensionally structured surface shape. The exposure is a multi-step exposure including a plurality of exposure steps using different masks, and as one of the exposure steps, the photosensitive material is exposed to light in order to remove a CEL (contrast enhanced lithography) layer of the photosensitive material. A method for manufacturing a three-dimensional structure, comprising a step of irradiating a layer. 2. In the first stage of the exposure, the exposure is performed with a light amount smaller than the exposure amount at the time of design by an amount of light for removing the CEL layer, and in the second and subsequent stages of the exposure, the exposure is reduced in the first stage. The method for manufacturing a three-dimensional structure according to claim 1, further comprising exposing a light amount. 3. The three-dimensional structure manufacturing according to claim 2, wherein the mask used in the step of irradiating the photosensitive material layer with the light amount for removing the CEL layer is a uniform concentration mask for uniformly exposing the entire surface. Method. 4. The method for manufacturing a three-dimensional structure according to claim 2, wherein the first-stage exposure is divided into two or more exposure steps. 5. A GM pattern No. corresponding to an uppermost part of a desired structure is used as a mask used in the first stage exposure.
Is set to zero (all black: light transmittance is zero), and the GM pattern No. of the other part is more C than the GM pattern No. obtained from the sensitivity curve of the photosensitive material in order to manufacture a desired structure.
5. The method for manufacturing a three-dimensional structure according to claim 1, wherein the mask comprises a GM pattern No. set so that the light transmission amount is reduced by an amount of light for removing the EL layer. 6. The method for manufacturing a three-dimensional structure according to claim 1, wherein the photosensitive material layer is left at a vertex of the three-dimensional structure.