JPH06302506A - Method and apparatus for electron beam exposure - Google Patents

Abstract

PURPOSE:To provide a method for electron beam exposure based on blanking aperture array wherein the edges of a pattern are controlled with an accuracy exceeding a specified pitch when exposing the pattern using exposure dots arranged in rows and columns with the specified pitch, and a high throughput is attained. CONSTITUTION:A single electron beam is shaped to obtain N electron beam groups, each composed of a plurality of electron beam elements arranged with the pitch specified. The electron beam groups are formed in a way that, letting the number of electron beam elements in each group be M, M is an arbitrary integer smaller than N and any adjoining groups are shifted from each other by 1/N pitch.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention generally relates to the manufacture of semiconductor devices, and more particularly to an electronic device capable of delicate pattern size setting when drawing a semiconductor pattern on an object such as a semiconductor substrate by a charged particle beam such as an electron beam The present invention relates to a beam exposure method and an exposure apparatus.

Electron beam lithography is an indispensable technique for manufacturing advanced semiconductor integrated circuits with high integration density. By using electron beam lithography,
It is possible to expose a pattern having a width of 05 μm or less with an alignment error of 0.02 μm or less. For this reason, electron beam lithography is central to the manufacture of future semiconductor devices, including DRAMs with large storage capacities exceeding 256 Mbits or 1 Gbits, or high-speed microprocessors with extremely powerful arithmetic functions. It is believed to play a role.

By the way, in manufacturing such a semiconductor device, not only the resolution of the element pattern but also the throughput at the time of manufacturing is of essential importance. In electron beam lithography, since writing is performed using a single focused electron beam, there is no choice but to be disadvantageous in this respect as compared with the conventional optical exposure method in which the entire pattern can be exposed in one exposure.
However, the conventional optical exposure method is reaching the limit of resolution, and therefore, there is a situation in which the electron beam exposure method must be relied upon in the future production of high-speed / high-performance semiconductor devices.

Under these circumstances, various attempts have been made to improve the throughput of electron beam exposure. For example, the applicant of the present invention has previously proposed a so-called block exposure method. In this block exposure method, the element pattern is decomposed into a large number of basic patterns, and the electron beam is shaped according to these basic patterns. The block exposure method is particularly suitable for exposure of a semiconductor memory device in which a relatively small number of basic patterns are often repeated, and a throughput of about 1 cm ² / sec is currently achieved.

On the other hand, in a semiconductor device such as a microprocessor including complicated and various logic circuits, the block exposure method has a problem that the exposure efficiency is lowered. This is because the pattern that can be formed by the block exposure method is limited to a small number of basic pattern combinations. In order to form an irregular exposure pattern by the block exposure method, it is necessary to form a so-called variable rectangular beam by an electrostatic or electromagnetic deflector, but such variable beam shaping is performed by the deflector settling time and control system operation. There is a problem that the exposure efficiency is reduced due to the speed and the like.

[0006]

2. Description of the Related Art In order to perform high-speed exposure of a semiconductor device including such complicated and various logic circuits, a drawing pattern is expressed as a dot pattern consisting of small exposure dots, and an electron beam for forming each dot is used. An electron beam exposure apparatus provided with a so-called blanking aperture array that turns on and off has been proposed.

FIG. 22 shows an outline of a conventional electron beam exposure apparatus using a blanking aperture array. FIG. 22
The electron beam exposure apparatus generally includes an electron optical system 100 that forms an electron beam and focuses the electron beam, and a control system 200 that controls the electron optical system 100. The electron optical system 100 includes an electron gun 104 as an electron beam source.
On the other hand, the electron gun 104 includes a cathode electrode 101, a grid electrode 102, and an anode electrode 103, forms an electron beam, and emits this as a divergent electron beam along a predetermined optical axis.

The electron beam formed by the electron gun 104 is shaped by passing through the beam shaping aperture 105a formed in the beam shaping plate 105. The aperture 105a is formed in alignment with the optical axis O, and shapes the incident electron beam into a rectangular cross section.

The shaped electron beam is transmitted to the electron lens 10
Blanking aperture array (BAA)
Are focused on the formed BAA mask 110, and the lens 107 projects the image of the rectangular aperture onto the BAA mask 110. As will be described later with reference to FIG. 20, BAA
A large number of apertures are formed on the mask 110 corresponding to a large number of exposure dots drawn on the semiconductor substrate, and an electrostatic deflector is formed on each aperture. The electrostatic deflector is controlled by the control signal E and allows the electron beam to pass through as it is in the non-excited state, but deflects the passing electron beam in the excited state, so that the direction of the passing electron beam deviates from the optical axis O. As a result, as described below, an exposure dot pattern corresponding to the non-excited state aperture is formed on the semiconductor substrate.

After passing through the electron lenses 108 and 116, the electron beam passing through the BAA mask 110 is focused at the focal point f ₁ on the optical axis O, and the image of the selected aperture is formed at the focal point f ₁ . To do. The electron beam thus focused passes through the round aperture 117a formed on the round aperture plate 117, and then is held on the movable stage 126 by the electron lenses 119 and 122 forming another reduction optical system. It is focused on the substrate 123. Here, the electron lens 122 acts as an objective lens and includes correction coils 120 and 121 for focus correction and aberration correction, deflectors 124 and 125 for moving the focused electron beam on the substrate surface, and the like.

The lens 108 and the round aperture plate 11
An electrostatic deflector 115 is formed in the middle of 7, and the deflector 115 is driven so that the path of the electron beam is plate 1.
It is deviated from the optical axis O passing through the 17 round apertures 117a. As a result, the electron beam can be turned on / off at high speed on the semiconductor substrate. Further, since the electron beam deflected by the excitation of the electrostatic deflector in the aperture on the BAA mask 110 described above also leaves the round aperture 117a, it does not reach the semiconductor substrate, and as a result, the exposure is performed. The dot pattern can be controlled.

To control the exposure operation, the electron beam exposure apparatus of FIG. 22 uses a control system 200. The control system 200 includes storage devices such as a magnetic tape device 201 and magnetic disk devices 202 and 203 that store data relating to element patterns of a semiconductor device to be drawn. For example, in the illustrated example, the magnetic tape 201 stores various design parameters, the magnetic disk 202 stores drawing patterns, and the magnetic disk 203 stores BAA mask 1.
It may be one used to store 10 aperture arrays.

The data stored in the storage device is stored in the CPU 20.
4, the data compression is released, and then the data is transferred to the interface device 205. At that time, the data designating the pattern on the BAA mask 110 is extracted,
Such data is stored in the data memory 206. The data stored in the data memory 206 is then transferred to the first control circuit 207 which forms the control signal E, which is supplied to the electrostatic deflector corresponding to each aperture on the BAA mask 110. To do. As a result, the excitation / non-excitation of each aperture on the BAA mask described above is controlled, and the exposure dot pattern on the substrate 123 is controlled by such control.

Further, the first control circuit 207 sends a control signal to the blanking control device 210, and the blanking control device 210 forms a blanking signal for interrupting the irradiation of the electron beam in response thereto. The blanking signal is then converted into an analog signal SB by the D / A converter 211, and the analog signal SB controls the deflector 115 to deflect the electron beam away from the optical axis O. As a result, the electron beam leaves the round aperture 117a,
It will not reach the surface of the substrate 123.

The interface device 205 further extracts data for controlling the movement of the electron beam on the surface of the substrate 123 and supplies it to the second control circuit 212. In response to this, the control circuit 212 transmits the electron beam to the substrate 1.
A control signal for deflection is formed on the surface 23, and the formed control signal is supplied to the deflection control circuit 215. The deflection control circuit 215 forms a deflection control signal according to the supplied control signal and supplies it to the D / A converters 216 and 217. D / A converters 216 and 217 drive signals SW1 and SW2 for driving the deflector according to the deflection control signal.
Are formed and are supplied to the deflectors 124 and 125 to deflect the electron beam. The position of the stage 126 is detected by the laser interferometer 214, and the wafer deflection control circuit 215 outputs the output deflection control signal, that is, the drive signal SW1,
SW2 is changed according to the measurement result of the stage position by the laser interferometer. Further, the second control circuit 212 forms a control signal for moving the stage 126 in the horizontal plane, and moves the stage 126 via the stage driving mechanism 213.

FIG. 23 shows an example of the structure of the BAA mask 110 used in the electron beam exposure apparatus of FIG.

Referring to FIG. 23, the mask 110 has a size b of about 1200 μm in the scanning direction and a size c of about 3200 μm in the direction orthogonal to the scanning direction. A plurality of rows of apertures 1A _{1 to} 1A _n , 1 extending in a direction substantially orthogonal to the scanning direction
B _{1 to} 1B _n , 2A _{1 to} 2A _n , ..., 8A _{1 to} 8A
_n , 8B _{1 to} 8B _n are formed. In the figure, the aperture columns includes an aperture columns 1A ₁ to 1A _n 1A, expressed as 1B such an aperture columns containing aperture columns 1B ₁ ~1B _n. In the mask of FIG. 23, the B-series apertures are formed by being shifted by one pitch so that they are staggered with respect to the A-series apertures. In the illustrated example, each aperture array has a side size S of 25 μm and a pitch 2S of 5
BAA including 64 apertures arranged at 0 μm
A total of 16 rows of apertures 1A on the mask 110
8B are formed in parallel with each other with a repeating pitch of 2S in the scanning direction. Corresponding to each aperture, a square dot pattern having a side of 0.08 μm is drawn on the semiconductor substrate 123.

Therefore, in the conventional electron beam exposure apparatus using such a blanking aperture array, a group of individual electron beams aligned in a direction substantially orthogonal to the scanning direction of the electron beam by the individual aperture rows is used. Become,
An electron beam group having a flat cross-sectional shape is formed, and the writing pattern is exposed as an aggregate of exposure dot patterns on the semiconductor substrate 123 by deflecting the flat electron beam group in the scanning direction. At that time, by arranging the respective apertures alternately, it is possible to avoid that the intervals of the electron beams irradiated at the same time are too close,
It becomes possible to avoid the problem of Coulomb interaction between electron beams which occurs due to the electron beams being too close to each other. Further, by performing multiple exposure using a plurality of aperture rows formed, it becomes possible to arbitrarily adjust the exposure amount of the exposure pattern formed on the substrate. As a result, it is possible to reduce the problem of the exposure pattern deformation depending on the pattern size, such as the proximity effect due to backscattered electrons, by changing the exposure amount so that the deformation of the exposure pattern is compensated. Become.

FIG. 24 shows the aperture formed on the mask 110 and the electrostatic deflector associated therewith.

Referring to FIG. 24, each electrostatic deflector is
Separate the shaded ground electrode GND from the aperture
It is composed of the drive side electrode ACT facing this,
The BAA control circuit is provided with each array arranged in the scanning direction in FIG.
Perch row COL₁, COL ₂, COL₃, ... CO
L₁₂₈Is provided with a plurality of drive circuits. rear
As explained in more detail in the same column, eg aperture
Row COL₂The apertures that make up the
The same drive signal from the path sequentially corresponds to the beam scanning speed.
It is supplied with a delay of
The electron beam passing through each aperture in the row causes
Square dot pattern with sides of 0.08 μm repeatedly exposed
Be illuminated. In order to perform such exposure, the aperture of each row
Driven through separate lines from the corresponding drive circuit
Signal is supplied. FIG. 24 shows the aperture array COL.₂To
Such lines are collectively indicated by the symbol L.
Lines corresponding to other aperture rows are not shown for simplicity.
Omitted.

FIG. 25 shows an example of an exposure pattern formed on the semiconductor substrate 123 by the mask 110 shown in FIG. Referring to FIG. 25, the exposure pattern is configured by a two-dimensional array of A-series and B-series exposure dots, and each exposure dot is formed as a rectangular exposure section having a side size of 0.08 μm. Therefore, the exposure dots are
By controlling the mask 110 to be turned on / off, the exposure pattern can be deformed with a rectangular area having a size of 0.08 × 0.08 μm as one unit.

[0022]

Problems to be Solved by the Invention FIGS.
(E) shows changes in the exposure pattern on the semiconductor substrate due to such multiple exposure, in particular, changes in the pattern size.
In the figure, the horizontal axis represents the coordinate axis set on the surface of the substrate 123 so as to be orthogonal to the scanning direction, and the vertical axis represents the exposure amount. The symbol TH on the vertical axis represents the exposure threshold of the electron beam resist formed on the semiconductor substrate 123. The hatched area indicates that the BAA mask 110 is used for exposure. In the figure, each rectangular area has a size of 0.08 μm on a side, as described above.

In the example shown in FIGS. 26A to 26E, exposure is performed.
Sometimes the row of apertures 1A on the BAA mask 110₁~ 1A
_n, 1B₁~ 1B_n, 2A₁~ 2A_n, 2B₁~ 2
B_n, 3A₁~ 3A_n, 3B₁~ 3B_n, 4A₁~ 4A
_n, 4B₁~ 4B_n1st group, aperture row 5A₁~
5A_n, 5B₁~ 5B_n, 6A₁~ 6A_n, 6B₁~ 6
B _n, 7A₁~ 7A_n, 7B₁~ 7B_n, 8A₁~ 8A
_n, 8B₁~ 8B_nAs a second group
Exposure, first group aperture row and second group aperture row
I went twice with. That is, the first group aperture
After exposing using columns, use the second group of aperture columns to
Double exposure is performed. In the figure, such exposure exposes the substrate.
The exposure amount profile realized by
It

FIG. 26A shows the case where the first group and the second group of aperture rows are used to expose the same pattern having an edge formed at position P ₀ . In the figure, the proximity effect associated with exposure is omitted to simplify the description. As described above, when the same pattern is exposed in the first exposure and the second exposure, the exposure amount profile shown by the solid line in the figure becomes very steep at the edge portion and is very sharp. The boundaries of the exposure pattern are obtained. Similarly, FIG. 26 (E) shows that the aperture array of the first group and the second group exposes a pattern having an edge at a position P ₁ offset by 0.08 μm from the position P ₀ of FIG. 26 (A). Here is an example that is used. Also in this case, the exposure amount profile becomes extremely steep at the edge portion, and a sharp boundary is obtained. In other words, FIG.
In the exposure process shown in 6 (A) and 26 (E),
It is possible to change the size of the exposure pattern in steps of 0.08 μm.

On the other hand, in the exposure method using the blanking aperture array, it may be desirable to change the size of the exposure pattern at intervals smaller than the interval of 0.08 μm. For example, when correcting the proximity effect, it is necessary to selectively change the exposure amount at the edge portion of the exposure pattern. FIG. 26
In the double exposure method shown in (A) to 26 (E), for example, as shown in FIG. 26 (B), at the time of the second exposure, the exposure of the exposure dot inside by one dot from the pattern edge is selected. It is possible to move the edge of the exposure pattern inward from the position P ₀ by 0.02 ± 0.02 μm by suppressing it. Further, in the example shown in FIG. 26C, at the time of the second exposure, the exposure of the dots forming the edge portion at the outermost portion of the pattern is suppressed, and the pattern edge is inward by 0.04 ± 0.01 μm from the position P _0. It is possible to move to. Further, in the example shown in FIG. 26 (D), during the second exposure, by suppressing the exposure of the two dots from the pattern edge, 0 for the position P _0.
The pattern edge is displaced by 06 ± 0.02 μm.

However, in the examples of FIGS. 26B to 26D, particularly the examples of FIGS. 26C and 26D, the actual exposure amount profile shown by the solid line is shown in FIG. ) Or 26 (E), the surface becomes flatter, resulting in the problem that the edge of the exposure pattern formed on the resist becomes uncertain. For example, if the irradiation amount of the electron beam or the exposure amount threshold value of the resist on the semiconductor substrate fluctuates even a slight amount, the edge position of the exposure pattern drawn on the semiconductor substrate changes greatly. Therefore, as shown in FIG.
In the example of 26 (D), the error of the pattern edge is large. On the other hand, as described above, in the example of FIGS. 26 (A) and 26 (E), the edge of the exposure pattern is determined to be very sharp. Therefore, in the double exposure method of FIGS. 26 (A) to 26 (E) using the BAA mask 110,
There is a problem in that the sharpness of the image formed on the electron beam resist on the substrate 123 changes depending on the size of the exposure pattern.

In order to solve this problem, exposure dots are first formed on the substrate using a BAA mask as shown in FIG. 23, and then X is formed on the exposure pattern composed of the exposure dots thus formed. An exposure method has been proposed in which another exposure pattern composed of exposure dots shifted by a half pitch in the Y direction and the Y direction is superimposed and exposed. According to such a multiple exposure method, the edge position of the exposure pattern can be finely displaced without lowering the exposure accuracy, as will be described later. However, in such multiple exposure, it is necessary to repeat the exposure once or more on the already exposed substrate, and there is a problem that the throughput of the exposure is significantly reduced.

The BAA mask 1 shown in FIGS.
In the conventional electron beam exposure apparatus using 10, the exposure timing of each aperture is shifted in FIG.
As shown in FIG. 7, there is a problem that the edge of the exposure pattern is wavy, and the desired exposure pattern cannot be obtained. Moreover, in general, when performing exposure on a wafer using the exposure apparatus of FIG. 22, the exposure is performed by dividing the wafer surface into a plurality of bands or cell stripes. If the deviation occurs, the exposure patterns will be different at the band boundaries.

Generally, the factors that cause the exposure timing of the aperture to be deviated are systematic changes in the line lengths of the individual lines L in the mask 110 as shown in FIG. 24, the drive circuit in the BAA control circuit 209 to the mask 11.
A change in the length of each signal line reaching 0, a clock skew in the drive circuit, and the like can be considered. These factors cause the displacement of the exposure dots in the scanning direction within each row, for example, the aperture row COL ₂ . Of these, a change in the line length of the line L in the mask and a clock skew only cause a signal delay of about several tens of picoseconds at most, and are not a serious problem. On the other hand, the BAA control circuit 20
The change in the line length from 9 to the mask 110 may not be negligible depending on the mounting state of the BAA control circuit 209 in the control system 200. Further, as described above, the apertures are divided into a plurality of groups by each aperture row COL ₁ , COL ₂ , COL _3, ...
Even when driven by the control circuit 209 to finely adjust the pattern, the exposure pattern may deviate depending on the mounting state of the BAA control circuit.

Further, as shown in FIG. 24, different aperture arrays COL ₁ , COL ₂ , COL ₃ ... Are driven by different drive circuits, so that variations in characteristics of individual drive circuits and BAA control circuit. 209 control system 2
There is a possibility that the exposure dots may be displaced due to the mounting state within 00. The deviation of the exposure dot is as shown in FIG.
7 causes an error in the edge of the exposure pattern. In the past, since the accuracy of the exposed pattern was not so high, such a shift of the exposure pattern did not cause much problem, and therefore no attempt was made to correct the shift of the exposure dot. But,
In an advanced electron beam exposure apparatus in which demands for exposure accuracy become strict and subtle exposure pattern correction such as proximity effect correction is a problem, it is necessary to correct the exposure dot shift during BAA exposure by some method. There is.

Therefore, it is a general object of the present invention to provide a new and useful electron beam exposure method and exposure apparatus that solve the above problems.

A more specific object of the present invention is an exposure method for exposing a pattern as an aggregate of exposure dots on an object by an electron beam exposure apparatus having a blanking aperture array, in which the size of the exposure pattern is blanked. The size of each exposure dot formed on the object by the aperture array can be changed at intervals smaller than that, and the accuracy of the exposure pattern does not substantially change when the size of the exposure pattern is changed, and An object of the present invention is to provide an electron beam exposure method capable of efficiently performing exposure, and an electron beam exposure apparatus capable of carrying out such an electron beam exposure method.

Another object of the present invention is to provide an electron beam exposure method and apparatus which eliminates the positional deviation of the exposure dots caused by the deviation of the exposure timing when forming the exposure pattern as a set of exposure dots. Especially.

[0034]

The present invention solves the above-mentioned problems.
A beam shaping step of shaping a single electron beam to form an electron beam pattern composed of a plurality of electron beam elements;
An electron beam exposure method comprising: an exposure step of drawing a predetermined exposure pattern on an object as a set of exposure dots arranged at a predetermined pitch by irradiating a plurality of electron beam elements forming the electron beam pattern. In the shaping step, the plurality of electron beam elements are repeatedly formed in a second direction which is aligned in the first direction and is orthogonal to the first direction by shaping the single electron beam. As a plurality of electron beam groups consisting of a set of mutually parallel electron beam elements, and forming each of the plurality of electron beam elements simultaneously so that different electron beam groups are formed. In between, the electron beam element has at least the first direction and the second direction, where N is an integer representing the number of electron beam groups and M is an arbitrary integer smaller than N. M mutually in either direction countercurrent
By an electron beam exposure method characterized by being shifted by / N pitch, or electron beam source means for forming an electron beam and emitting the electron beam along a predetermined optical axis; formed on the predetermined optical axis, Focusing means for focusing the electron beam on an object; electron beam shaping means formed on the predetermined optical axis for shaping the electron beam to form a plurality of electron beam elements; controlling the electron beam shaping means And shaping means for forming the plurality of electron beam elements according to a predetermined exposure pattern; and deflecting means for deflecting the shaped electron beam elements, the exposure pattern being formed on the object by the electron beam elements. In an electron beam exposure apparatus for drawing: the electron beam shaping means is formed with a plurality of openings, and a single electron beam is formed into a plurality of electrons corresponding to the plurality of openings. A mask plate for shaping into a beam element, and a deflecting device formed in each of the plurality of openings for deflecting an electron beam element passing through the opening; the shaping control means causes the deflecting device to expose an exposure pattern. A plurality of apertures formed on the mask plate in a first direction and a second direction different from the first direction. To form a plurality of opening groups arranged in a predetermined pitch, each extending in the first direction and repeated N times in the second direction at the predetermined pitch; in each opening group, Each opening is repeated at the predetermined pitch in the first direction, and N is the number of opening groups and M is an integer less than N with respect to the corresponding openings in the adjacent opening groups. As the above Switch M / N in the first and second directions, and the shaping control means is provided on each of the deflecting devices formed in the plurality of opening groups. A beam shaping step of forming a plurality of electron beam elements by an electron beam exposure apparatus or shaping a single electron beam to form an electron beam pattern composed of a plurality of electron beam elements, characterized by supplying a plurality of independent drive signals. An electron beam exposure method, comprising: an exposure step of drawing a predetermined exposure pattern as a set of exposure dots arranged at a predetermined pitch by irradiating a plurality of electron beam elements forming the electron beam pattern: the beam shaping A step of applying the single electron beam,
Passing through a mask having a plurality of openings turned on and off according to the electron beam pattern; the beam shaping step further includes a timing detection step of detecting a driving timing at which the openings are turned on and off; And an electron beam exposure method, which includes a timing adjusting step of adjusting the timing to a predetermined timing.

[0035]

According to the present invention, when the exposure pattern expressed as a set of dots is formed by the multiple exposure process, the exposure pattern is shifted by M / N pitch for each exposure, and the exposure is performed for each exposure. By changing the pattern, the edge position of the exposure pattern can be set very finely. Moreover, by forming a plurality of electron beam elements from a single electron beam and simultaneously exposing a plurality of exposure dots shifted by M / N pitch, it is possible to greatly improve the throughput in the exposure process.

More specifically, by shifting the exposure pattern by M / N pitch for each exposure, the exposure amount at the edge portion of the exposure pattern changes abruptly in any of the exposure patterns. Even when the edge position is finely set, the position of the edge is unlikely to become uncertain. In other words, according to the present invention, it is possible to change the exposure pattern very finely without sacrificing the exposure accuracy, and particularly, the deformation of the exposure pattern due to the correction of the proximity effect can be accurately and easily performed. It can be executed with high throughput.

Hereinafter, the principle of the present invention will be described in detail with reference to FIG. FIG. 1 corresponds to a case where double exposure is performed using a BAA mask formed by shifting the row of openings by 1/2 pitch in the X and Y directions.

Referring to FIG. 1, the first exposure pattern
X₁, X for the second exposure pattern₂Expressed as
Light pattern X₁, X₂Are each 0.08 × 0.0
A rectangle having a size of 8 μm and shown by a solid line in the figure
Each exposure dot is formed by a two-dimensional array of exposure dots
Each exposure pattern X by turning on / off₁, X ₂
Is exposed. In this case, each exposure pattern X₁, X₂To
The rectangular exposure dot is 0.08 in the X and Y directions.
It is repeatedly formed with a pitch of μm. Furthermore, the first
Second exposure pattern X₁And the second exposure pattern X₂
And, the position of the exposure dot is half the pitch, that is, 0.04
It is formed so as to be displaced by .mu.m in the X and Y directions.
Furthermore, pattern X₁, X₂The semiconductor substrate on which the exposure of
Appropriate photosensitivity threshold of electron beam resist coated on the plate
By setting to 1, the exposure of the resist is
The exposure pattern X shown by₁And the exposure pattern
X₂To occur only in the overlapping part of
You can

In the double exposure process, when the exposure pattern X ₁ and the exposure pattern X ₂ are the same, it is obvious that the exposure pattern changes by 0.08 μm. On the other hand, in the present invention, the edge of the exposure area shown by hatching in FIG. 1 is changed by the 0.08 μm step by changing the _first exposure pattern X ₁ and the second exposure pattern X ₂ separately. It becomes possible to change the distance substantially smaller than that. At that time, in the present invention, the exposure pattern becomes uncertain at a specific pattern size as shown in FIGS. 25 (C) and 25 (D) as described in FIGS. 25 (A) to 25 (E). Points can be avoided, and it is possible to ensure substantially uniform exposure accuracy in any case.

2 (A) -2 (C) and FIG. 3 (D)-
3 (E) is a diagram showing the size of the exposure pattern formed as a result of the double exposure on the electron beam resist by independently changing the exposure patterns X ₁ and X ₂ in the double exposure process shown in FIG. is there. In the figure, as in the case of FIGS. 25A to 25E, the vertical axis represents the exposure amount and the horizontal axis represents the coordinate axis on the semiconductor substrate. The exposure dose threshold of the electron beam resist is indicated by TH. In the figure, a ₁ , a ₂ , ... Are the exposure dots in the _first exposure pattern X ₁ ,
₁ , b ₂ , ... Represent exposure dots in the _second exposure pattern X ₂ . In addition, in each of FIGS. 2A to 2C and FIGS. 3D to 3E, the drawing on the right side is an exposure formed on the substrate corresponding to the double exposure. Represents a pattern. In the figure, each exposure dot is shown as being separated from each other, but this is merely for the convenience of illustration, and in reality, each exposure dot is formed overlapping with the adjacent exposure dot. There is.

2 (A) to 2 (C) and 3 (D) to
3 (E), FIG. 2 (A) corresponds to the state shown in FIG. 1, and the same patterns are mutually in the X direction and the Y direction.
The exposure pattern X ₁ and the exposure pattern X ₂ are formed by being shifted by half a pitch in the direction. In the figure, the solid line represents the exposure amount profile associated with such double exposure. However, in FIGS. 2A to 2C and FIGS. 3D to 3E, changes in the exposure amount due to the proximity effect are omitted for simplicity of description. Further, the exposure order of the exposure pattern X ₁ and the exposure pattern X ₂ is not important, and the exposure pattern X ₁
The same result can be obtained by first exposing and then exposing the exposure pattern X _{2 or} by reversing the order. Therefore, in FIGS. 2 (A) to 2 (C) and FIGS. 3 (D) to 3 (E), the vertical relationship of the exposure patterns X ₁ and X ₂ may be reversed in some figures, but This has no effect on the exposure result.

In the case of FIG. 2A, corresponding to the threshold value TH,
It can be seen that the exposure pattern formed on the semiconductor substrate has an edge at position P ₀ . However, the position P ₀ corresponds to the point where the exposure profile crosses the threshold value TH. On the other hand, in the example of FIG. 2B, the exposure of the exposure dot a ₂ among the exposure dots a ₁ , a ₂ , ... Which constitute the _first exposure pattern X ₁ is suppressed, and as a result, the solid line The exposure profile indicated by is formed at a position slightly displaced from the case of FIG. 2A by about 0.02 μm. At that time, the region on the semiconductor substrate corresponding to the exposure dots a ₂ exposure is suppressed, the exposure dots b _1, since each of the b ₂ are overlapped exposure by a half pitch, also than particular exposure dots a ₂ edge P Exposure dot b ₁ close to ₀ by half pitch
Since the exposure is continued, the shape of the exposure profile shown by the solid line, particularly the inclination thereof, does not change much compared with the case of FIG. 2 (A). In other words, the exposure profile in FIG. 2B maintains a steep shape at the edge portion of the exposure pattern, so that the exposure dose threshold T with respect to the electron beam exposure dose.
Even if H changes a little, the edge of the exposure pattern formed on the semiconductor substrate does not change much from the predetermined position. In this case, the edge of the exposure pattern is determined with an accuracy of ± 0.01 μm.

FIG. 2C shows the first exposure pattern X ₁
The exposure profile corresponding to the case where the exposure of the outermost exposure dot a ₁ is suppressed, and the exposure pattern formed on the semiconductor substrate accordingly are shown. In the case of FIG. 2C, the exposure dot whose exposure is suppressed is the outermost exposure dot a ₁
Therefore, the displacement amount of the exposure profile is larger than that in the case of FIG. 2B and is about 0.04 μ with reference to the position P _0.
However, since the exposure of the exposure dot b ₁ immediately inside, which overlaps the exposure dot a ₁ by a half pitch, is continued, the shape of the exposure profile is different from the case of FIG. 2A except for the effect of parallel movement. And the steep exposure profile shape is maintained at the edge portion of the exposure pattern. That is, in this case, the exposure accuracy equivalent to that in FIG.

FIG. 3D shows the first exposure pattern X ₁
The exposure profile corresponding to the case where the exposure of the exposure dot a ₁ therein and the exposure dot b _{2 in} the second exposure pattern X ₂ is suppressed, and the exposure pattern formed on the semiconductor substrate accordingly are shown. In this case, the exposure profile is translated by about 0.02 μm from the case of FIG. 2C, and as a result, the edge of the exposure pattern is displaced 0.06 μm inward from the position P ₀ . Even in this case, the shape of the exposure profile does not change significantly, and the shape is ± 0.01 μm.
Exposure accuracy can be obtained.

FIG. 3E shows the first exposure pattern X ₁
The case where the exposure of the exposure dot a ₁ therein and the exposure dot b _{1 in} the second exposure pattern X ₂ is suppressed is shown. Figure 3
When the exposure profile of (E) is moved in parallel, FIG.
2 is completely the same as the exposure profile of FIG. 2 and the corresponding exposure pattern on the semiconductor substrate is displaced 0.08 μm inward as compared with the case of FIG.

FIG. 4 shows FIGS. 2 (A) -2 (C) and FIG.
4D shows changes in the edge position on the semiconductor substrate of the exposure pattern achieved by the exposure processes shown in FIGS. In FIG. 4, position A on the horizontal axis corresponds to FIG. 2 (A), and B corresponds to FIG. 2 (B). Similarly, C is shown in FIG.
3D corresponds to FIG. 3D, and E corresponds to FIG.
Corresponds to (E). Of course, the accuracy of the exposure pattern is maximized in the cases of FIGS. 2 (A), 2 (C) and 3 (E), but according to the gist of the present invention, the exposure pattern is multiplexed while being displaced in the X and Y directions. By performing the exposure, it is possible to perform high-precision exposure without sacrificing the accuracy of the exposure pattern even in an intermediate exposure pattern as shown in FIGS. 2B to 3D.

In the present invention, a plurality of electron beam elements corresponding to the exposure pattern shown in FIG. 1 and having M / N pitch shifts are simultaneously formed to expose a plurality of points on the substrate at the same time. By sequentially scanning the plurality of electron beam elements at each of the points, it is possible to repeatedly expose the exposure dots that are shifted by M / N pitch from each other by one deflection scan of the electron beam, and thus it is possible to perform highly accurate exposure. Can be executed with high throughput.

Further, according to the present invention, since the timing deviation of the drive signal for turning on / off the opening formed on the beam shaping mask is compensated for, the length of the wiring for transmitting the drive signal depends on the opening. Even if it changes
It is possible to minimize the error in the exposure pattern by aligning the timing of exposure of each exposure dot.

[0049]

Next, an embodiment of the present invention will be described with reference to the BAA mask 220 shown in FIG. This BAA mask 22
0 is used to form an exposure dot pattern instead of the BAA mask 110 in the electron beam exposure apparatus of FIG. Alternatively, the mask 220 is shown in FIG.
It can also be applied to the electron beam exposure apparatus described in relation to No. 4.

Referring to FIG. 5, the mask 220 has a row of openings 1A, 1B, 2A, 2B, ... Composed of openings arranged alternately with a pitch 2S similar to the mask 110.
.., 8A, 8B, of which opening row 1A-4B
And opening column 5A~8B form a group EX _1, EX ₂ of the first and second openings columns respectively. However, in the mask 220 of FIG. 5, the openings 1A ₁ , 1A ₂ , ..., 4B _n−1 , 4 forming the first group EX ₁ are formed.
B _n and the opening 5 forming the second group EX _2.
A ₁ , 5A ₂ , ..., 8B _n-1 , 8B _n are half pitches in the X direction and the Y direction, respectively, at the boundary between the groups EX ₁ and EX ₂ shown by the broken lines in FIG. S only and staggered in, X with respect to the electron beam group electron beams formed at the opening group of the group EX ₁ along with this is formed by the opening group of the group EX ₂
And a distance S in the Y direction. Each opening is of course provided with a corresponding deflector,
_The double exposure shown in FIGS. 2 (A) to 2 (C) and FIGS. 3 (D) to 3 (E) is performed by controlling the deflectors of the opening group of _{1 and} the opening group of EX ₂ independently. It becomes possible to execute the process.

Typically, in each group EX ₁ and EX ₂ of the BAA mask shown in FIG. 5, each opening group is driven by a common drive signal. More specifically, for example, when an exposure pattern is formed on the substrate 123 shown in FIG. 22, all the openings shown in FIG. 5 are simultaneously driven by respective drive signals, and the substrate 123 is shown on the substrate 123 at each moment. , An exposure dot pattern corresponding to the turned-on opening is formed. For example, all openings on the mask 220 in FIG. 5 are on (that is, a state in which an electron beam is allowed to pass).
In this case, the exposure dot pattern corresponding to the opening on the mask 220 is reduced and formed on the substrate 123. Further, the electron beam element is sequentially deflected by one dot in the Y direction and sent, and in a state of being deflected by 6 dots, the electron beam element is exposed on the exposure dot on the substrate previously exposed by the electron beam element from the opening 1A ₁ and then, Exposes the exposure dots in duplicate by the electron beam elements from the corresponding openings in the opening row 2A. In this way, at each point on the substrate 123, after the exposure dots are sequentially and multiple-exposed by the electron beam elements from the opening rows 1A to 4A or the opening rows 1B to 4B, the opening rows 5A to 8A or 5B-8B
Exposure dots corresponding to are sequentially formed with a half pitch shift in the X and Y directions. At that time, at each point, the drive signal is changed between the opening group EX ₁ and the opening group EX ₂ , so that FIG. 2 (A) to 2 (C), FIG.
As described above, the precise position determination of the pattern edge becomes possible.

In the mask of FIG. 5, each opening row 1A-8 is formed.
Since the electron beam is simultaneously formed in each of the openings in A and 1B to 8B, the above-described overlapping exposure is actually performed on the substrate 123.
2 (A) to 2 (C) and FIG. 3 () by repeatedly performing overlapping exposures in the Y direction, because a plurality of points (16 × n points in the example of FIG. 5) on the surface are simultaneously executed. Even when the effects shown in D) and 3 (E) are obtained, it is possible to obtain a very excellent throughput.

FIGS. 6A to 6D show an example in which the present invention is applied to a main exposure pattern. In the figure, the exposure pattern is formed by the exposure dots A and the exposure dots B shifted by a half pitch from the exposure dots A. Of these, in the example of FIG. 6A, a total of 25 exposure dots A arranged in 5 columns and 5 rows. Superimposed on 4
A total of 16 exposure dots B arranged in 4 rows are exposed. On the other hand, in the example of FIG. 6B, a total of 9 exposure dots A arranged in 3 rows and 3 columns are overlapped with a total of 16 exposure dots B arranged in 4 rows and 4 columns. As a result, an exposure pattern having a boundary line shifted inward by one pitch from the case of FIG. 6A is formed. In the case of FIG. 6C, the boundary line is shifted further inward by one pitch.

On the other hand, in the example of FIG. 6D, 4 dots of the exposure dots B of 4 rows and 4 columns which are exposed by being superposed on the exposure dots A of 5 rows and 5 columns are selectively exposed. In this example, the exposure line B is prevented from being exposed, so that the boundary line of the exposure pattern can be changed at a finer pitch than one pitch.

FIG. 7 shows an exposure mask 220 'according to a second embodiment of the present invention.

Referring to FIG. 7, the exposure mask 220 'is
In each of the opening groups, a plurality of rows of opening groups A to D are included.
Is two rows of openings where each opening is offset by one pitch, eg
For example, the opening row A₁And A₂, B₁And B₂...
Formed as. Furthermore, an opening row of the opening group A,
A corresponding opening row of the opening group B adjacent to this, for example,
Opening row A₁And the opening row B forming the₁To form
The opening row does not have a 1/4 pitch in the X and Y directions.
Is being used. A similar relationship is for the opening row A₂And B ₂, B₁
And C₁, B₂And C₂And so on. Sanawa
In general, in an exposure mask having an opening group of N rows
Is the opening in one opening group and the opening group adjacent to it.
Corresponding openings in the X and Y direction, M from N
It is shifted by M / N pitch as an arbitrary small integer.

8A to 8E show the mask 22 of FIG.
An example of exposure using 0'is shown. In this way, by selectively suppressing the exposure of the exposure dots,
Fine adjustment of the exposure edge position as described in (A) to 3 (E) is possible. For example, in the example of FIG. 8B, the edge position of the exposure pattern is displaced inward (leftward in the figure) by 0.01 ± 0.05 microns with respect to that of FIG. 8A. On the other hand, in the example of FIG. 8C, the edge of the exposure pattern is offset by 0.02 micron to the left side from that of FIG. 8A. In the case of FIG. 8C, the edge position is determined with the same precision as in the case of FIG. In the example of FIG. 8D, the pattern edge is displaced 0.07 ± 0.005 micron to the left side as compared with the case of FIG. 8A, and in the example of FIG. It is displaced by 0.08 micron to the left with respect to the case of 8 (A), and the edge position accuracy at that time is the same as in the cases of FIGS. 8 (A) and 8 (C), and is the highest. Further, in the mask of FIG. 7, the mask of FIG.
It is possible to combine the exposure patterns other than those shown in FIGS.

Next, the structure of the BAA mask according to the third embodiment of the present invention will be described with reference to FIG.

Referring to FIG. 9, BAA according to the present embodiment.
The mask includes a plurality of opening groups A, B, C, and D, each of which is composed of two rows of openings extending in the X direction and offset from each other by one pitch in the X and Y directions. In the opening group B, the openings are shifted by a half pitch in the X direction. More specifically, the openings composing the opening group A are composed of the openings A ₁ and A ₂ which are offset from each other by one pitch in the X and Y directions, and the openings A ₁ and A ₂ are Two rows of openings are formed, each extending in the X direction. Similarly, the opening group B is composed of an opening B ₁ and an opening B ₂ that form two opening rows extending in the X direction, and the opening B ₁ is relative to the opening A ₁ . 1/2 in X direction
The pitch is shifted. Further, the opening B _{2 is} also displaced from the corresponding opening A ₂ by ½ pitch in the X direction.

Similarly, the opening group C is composed of openings C ₁ and C ₂ forming two rows of openings aligned in the X direction,
The opening group D is composed of openings D ₁ and D ₂ forming two rows of openings aligned in the X direction, each opening C ₁ being in the opening group A and corresponding. It is displaced by 1/2 pitch in the Y direction with respect to the opening A ₁ . Similarly, each opening C ₂ has a corresponding opening A in opening group A.
It is shifted a half pitch in the Y direction with respect to _2. Further, the openings D ₁ in the opening group D are displaced by 1/2 pitch in the X and Y directions with respect to the corresponding openings A ₁ .
The opening D _{2 is} also displaced from the corresponding opening A ₂ by ½ pitch in the X and Y directions.

FIGS. 10A and 10B show exposure dot patterns drawn on the substrate using the mask of FIG. In the figure, exposure dots A to D correspond to the opening groups A to D on the BAA mask, respectively. As shown in detail in FIG. 10 (B), the exposure dots A to D in FIG. 10 (A) are overlapped with each other by a half pitch, and the exposure pattern is changed in each opening group. This makes it possible to finely adjust the pattern edge. In particular, when the mask of FIG. 9 is used, as shown in FIGS. 10A and 10B, when the exposure dots are overlapped and exposed on the substrate using the mask of FIG. Problems such as diagonally arranging downward can be solved, and highly accurate exposure can be performed regardless of the direction in which the exposure pattern extends.

11 (A) to 11 (E) use FIGS. 2 (A) to 2 (C) and FIG. 3 using the mask shown in FIG.
An example in the case where the same multiple exposure as in (A) and 3 (B) is performed is shown.

FIG. 11A shows a reference pattern, in which exposure 1 to exposure 4 are overlapped and executed on the substrate corresponding to the opening groups A to D. In the reference pattern of FIG. 11A, the opening groups A to D are driven by the same pattern, and as a result, FIG.
As shown on the right of (A), a sharp pattern edge is formed on the substrate at the reference position P ₀ . On the other hand, FIG.
In the example of 1 (B), by changing the exposure pattern of the opening group A, the formation of the exposure dot closest to the edge in exposure 1 is suppressed, and as a result, the exposure edge is shifted by 0.01 μm to the left in the figure. Further, in the example of FIG. 11C, only the exposure pattern of the opening group B is changed, and as a result, the edge of the exposure pattern is further 0.01 μm to the left in the figure.
It just shifts. In the example of FIG. 11D, the exposure patterns of the opening group A and the opening group D are changed, and as a result, the edge of the exposure pattern is further shifted by 0.01 μm to the left in the drawing. Further, in FIG. 11E, the exposure pattern is changed in all of the opening groups A to D, and the edge of the exposure pattern is further shifted to the left by 0.01 μm.

As described above, according to this embodiment, as shown in FIG.
It is possible to adjust the edge position of the exposure pattern in finer steps than in the case described with reference to (A) to 2 (C) and FIGS. 3 (D) to 3 (E). At this time, there is no problem that the edge position becomes uncertain even in an intermediate pattern as shown in FIGS. 11 (B) to 11 (D).

FIGS. 12A to 12C show an example in which the fine adjustment of the exposure pattern according to the present invention is applied to the proximity effect correction. In the figure, FIG. 12 (B) shows the exposure of the main exposure pattern, and FIG. 12 (A) shows the auxiliary exposure that is executed to compensate for the decrease in the exposure amount that occurs in the peripheral portion of the minute pattern or the large area pattern due to the proximity effect. Is shown in FIG.
(C) shows the exposure actually performed on the semiconductor substrate. In such auxiliary exposure, FIG.
As shown in (1), it is necessary to precisely overlap the minute pattern or to selectively expose the edge portion of the exposure pattern, so that precise control of the exposure pattern is required.
By applying the present invention to such auxiliary exposure, it is possible to effectively correct the proximity effect.

FIGS. 13A to 13C show the BA of FIG.
An example is shown in which the throughput of electron beam exposure is improved by performing proximity effect correction by one electron beam scan using the A mask 220.

Referring to FIGS. 13A to 13C,
FIG. 13A shows the result of main exposure performed using the mask 220. In this exposure, each exposure dot is exposed according to the main exposure pattern, and FIG. 13A shows the total exposure amount of each exposure dot. When such exposure is performed, the background exposure amount increases in proportion to the pattern density due to the well-known proximity effect.
In the isolated pattern on the left of the figure (minimum pattern density), the background exposure amount is almost zero, whereas in the continuous pattern on the right of the figure (maximum pattern density), the background exposure amount is almost the same as the actual exposure amount. ing. Moreover, in the middle intermediate pattern, intermediate background exposure occurs.

In order to compensate for such a change in the exposure amount due to the pattern density, conventionally, a correction exposure pattern which is substantially the same as the main exposure pattern is superposed, and the total exposure amount is substantially constant for each pattern. An exposure method has been proposed which is performed so that FIG. 13B shows such correction exposure, in which the same isolated pattern is overlapped and exposed to the left isolated pattern with the same exposure amount as in the case of main exposure.
Further, in the continuous pattern on the right side of the figure, the exposure amount is set to substantially zero, and the correction exposure is not substantially performed. Further, in the intermediate pattern, correction exposure is performed with an intermediate exposure amount according to the pattern density, and as a result, the total exposure amount shown in FIG. 13C is almost the same for all exposure patterns. Such an intermediate exposure amount can be obtained by roughening the density of the exposure dots when performing the correction exposure. In such a proximity effect correction method, when the development is carried out at the exposure threshold value shown by the broken line in FIG. 13 (C), a highly accurate exposure pattern with less spread is obtained as compared with the case where correction exposure is uniformly applied. Can be done.

In the present invention, when the auxiliary exposure pattern is superposed on the main exposure pattern, the BAA mask 220 shown in FIG. 5 is used.
High-precision auxiliary exposure is possible using. Further, conventionally, the exposure of FIG. 9 (A) had to be performed again after the exposure of FIG. 9 (A), whereas in the present invention, by using the mask 220 of FIG. FIG.
It becomes possible to execute the pattern of FIG. 13B by using the opening group EX ₂ immediately after the pattern of (A) is executed by the opening group EX ₁ , and the main exposure and the subsequent auxiliary exposure can be performed. It becomes possible to execute it efficiently.

FIG. 14 shows the structure of an electron beam exposure apparatus for carrying out the present invention.

Referring to FIG. 14, the electron beam exposure apparatus uses the BAA mask 220 or 2 as the electron optical system 100.
An electrostatic deflector having an electron optical system having the same configuration as that of the conventional electron beam exposure apparatus shown in FIG. 22 except that 20 'is used and formed corresponding to each opening in the BAA mask. def is driven by the drive unit 300 including the corresponding drive circuits 301 _{1 to} 301 _N. For example, as shown in FIG. 5, each opening row 1A, 1B, ... Includes 128 openings, and each opening row is formed by two opening groups that are offset from each other by a half pitch. When such a BAA mask is used, the driving unit 300 includes 128 × 2 driving circuits, and each driving circuit drives the electrostatic deflector of the corresponding opening. On the other hand, when the BAA mask shown in FIG. 6 is used, 128 × 4 driving circuits are used as the driving unit 300. In general, if each opening row includes N groups of openings that are offset from each other by M / N pitch, the driving unit 300 has 128 units.
It is formed by × N driving circuits. Each drive circuit 301
_{1 to} 301 _N includes a D / A converter and an analog amplifier.

[0072] Further, BA of the BAA data generating circuit 302 _{1-302 128} corresponding to the respective driving circuits are N columns formed
A data storage and output portion 302 is provided in each BAA data generating circuit 302 _{1-302 128} is supplied exposure data from the data expansion unit 303 via a bus. The data expansion unit 303 includes N sets of data processing units 303 _{1 to} 3 corresponding to the N groups of opening portions that are shifted from each other by M / N pitch.
Is composed of 03 _N, the data processing unit, for example unit 303 ₁ via the CPU 304, or directly to a buffer memory 303a to be supplied to the exposure data from the external storage device 201 shown for example in FIG. 22, the buffer memory The data expansion circuit 303b is supplied with exposure data from 303a and expands it into an exposure pattern for each opening row, and a canvas memory 303c for storing the expanded exposure pattern. The configuration of the data expansion unit 303 will be described in detail later. Further, the data expansion unit 303 is provided with a refocus control circuit 303d for correcting the shift of the focal length due to the Coulomb repulsive force generated when the electron beam is focused. Since the structure and operation of the refocus control circuit 303d are not related to the gist of the present invention, the description thereof will be omitted.

The CPU 304 controls the exposure control device 305, while the exposure control device 305 controls the main deflection drive circuit 306 and the sub deflection circuit 307 corresponding to the deflection circuit 215 shown in FIG. Vessel 124,1
25 is driven to deflect the electron beam on the substrate 123. The main deflection drive circuit 306 includes a distortion correction circuit 306a, and the correction circuit 306a corrects aberrations and the like via a focus / stig correction circuit 310. Furthermore, CPU3
Reference numeral 04 controls the stage moving mechanism 308 and the autoloader 309 corresponding to the stage moving mechanism of FIG.

FIG. 15 shows a buffer memory of the data expanding unit 303 from the CPU 304, for example, the buffer memory 303.
The example of the exposure pattern supplied to a is shown. Referring to FIG. 15, white circles represent exposure dots, and the exposure pattern P is represented by a set of white circles. In the figure, the exposure pattern P is divided into a plurality of cell stripes extending in the vertical direction, while each cell stripe is formed by a bit line formed by a set of 128 exposure dots aligned in the horizontal direction. Typically, the cell stripe is approximately 100 μm in the longitudinal direction.
m, and has a size of about 10 μm in the lateral direction.

FIG. 16 is a diagram for explaining the configuration of the data expansion unit by taking the data processing unit 303 ₁ as an example.

Referring to FIG. 16, the exposure data supplied from the external storage device is shape data OP representing the pattern shape, starting point data X ₀ and Y ₀ representing the starting points in the X direction and the Y direction, and in the X direction. It is composed of data X ₁ representing the size of the extending pattern and data Y ₁ representing the size of the pattern extending in the Y direction. Of these, the shape data OP is directly connected to the control device 317 in the developing unit. The block pattern library 319 and the start point data Y ₀ are supplied to the address counter 303d in the data expansion section. On the other hand, the starting point data X ₀ is supplied to the bitmap data shifter 312, and further the length data X ₁ is supplied to the block pattern library 319.

The length data X ₁ read from the block pattern library 319 is held in the register 311 and then supplied to the bitmap data shifter 312, and the origin is moved based on the start point data X ₀ . The bit map data for one line thus formed is held in the output register 313.

The bit matrix held in the output register 313
The adder 314 and the subtractor 315.
Is already supplied to the canvas memory 303c.
Correspond out of the exposure pattern bitmap data
Addition to the bitmap data of the bit line
Rui is subtracted. As a result, the exposure dots have already accumulated
Is added to the exposure pattern bitmap
It will be deleted. Of the adder 314 and the subtractor 315
The control of which to use is controlled by the control unit in the data expansion unit 303.
The position 317 is used based on the shape data OP.
It Also, depending on the shape data OP, addition and subtraction
In some cases, none of the above is done. Processed in this way
The bit map data for one bit line is OR times
After being processed in path 316, the register for memory writing is
After passing through the star 318, the ad in the canvas memory 303c is
Stored at the address specified by the response counter 303d
It This address is the start point data Y described above.₀From
One of the addresses to be started (Y ₀+ ΔY, ΔY <Y₁) To
Correspond. Further, the control device 317 is the exposure control device of FIG.
It is controlled by the position 305 and is based on the shape data OP.
Canvas memory 303c, writing register 318, addition
It controls the calculators and subtractors 314 and 315.

In the present invention, as shown in FIG.
Data developing device 303 of the structure _{1 ~303 N} is, BAA
N openings are provided corresponding to the N openings formed on the mask 220 or 220 ′ by M / N pitch shift with respect to each other, and are provided in FIGS. 2, 3A to 7E, or FIG. A)
(E), or the exposure shown in FIGS. 10A and 10B is possible.

Next, a fourth embodiment of the present invention will be described with reference to FIG. The purpose of this embodiment is to correct the timing deviation of the exposure dots in the BAA exposure performed previously with reference to FIG.

[0081] Referring to FIG. 17, BAA data generating circuit 302 _{1-302 128} is driven by the clock supplied from the oscillator 302x, also delay circuit 302d are provided corresponding to the BAA data generation circuit for each. The delay circuit 302d is controlled by the output of the control circuit 302c, while the control circuit 302c cooperates with the memory 302m,
The output of the corresponding BAA data generation circuit is stored in the memory 302.
Delayed by the amount of delay stored in m corresponding to each aperture on the BAA mask 220. As described in the previous embodiment, the masks 220 are actually 1 / m on each other.
Aperture groups that are offset by two pitches are formed.
In FIG. 7, only one aperture group is shown for simplicity. Furthermore, the response to these two apertures group BAA data generating circuit 302 _{1-302 128} also provided two groups.
Further, when the aperture group N group is formed as in the embodiment of FIG. 7 or FIG. 9, the delay circuit 302d corresponding to the BAA data generation circuit 302 _{1-302 128} and also provided N group.

Here, the memory 302m is the BAA mask 2
It is possible to store all the delay amounts corresponding to the respective apertures on the 20. However, as described above, the delay of the exposure timing is mainly caused by the wiring length of the wiring associated with each BAA data generation circuit. Therefore, the aperture on the mask 220 is divided into an aperture group including a plurality of apertures driven by the same BAA data generation circuit, and the delay amount of the drive signal is reduced for each aperture group in the memory 302m. Is preferably stored. As explained previously, the circuit 302 ₁ to 302 may be ignored by the clock skew order of several tens of picoseconds at _128, also several picoseconds also shift the exposure timing by the line length of the wiring in BAA mask 220 Therefore, it can be ignored. In contrast, the line length of the signal path from the BAA data generating circuit 302 _{1-302 128} BAA mask 220 circuit 302 ₁
Depending on the mounting state of .about.302 ₁₂₈ , it may change within a range of about 1 m, and in such a case, a delay of about 10 ns may occur in the exposure timing. Generally, the delay due to the line length of the signal path is about 10 nsec / m. See also FIG.
BAA data generating circuit 302 _{1-302 128} at 4
The situation is the same even when the D / A converter 301 is provided between the BAA mask 220 and the BAA mask 220.

When operating the exposure apparatus according to the present embodiment, first, the control circuit 302c controls each delay circuit 302d so that the corresponding BAA data generation circuits 302 ₁ to 302 _1-3.
The drive signal output from the output circuit 02 _{12 8}
To each aperture on the BAA mask 220 without including the delay due to Further, the delay detection circuit 302f detects the timing of the drive signal in each of the BAA aperture groups driven by the same BAA data generation circuit, and the memory 3 stores data representing the obtained timing.
At 02 m, it is stored corresponding to each of the BAA aperture groups. Next, the control circuit 302c reads the timing of the drive signal in each aperture group, and based on the timing in the aperture group having the largest delay amount,
Calculate the delay amount of each aperture group and store it again in memory 3
It is stored in 02f.

Next, when the actual exposure is performed, the control circuit 302c causes each of the delay circuits 302d to match the timing of the drive signal in each aperture group with the timing of the aperture group having the largest delay amount. Control to do.

FIG. 18 shows a flowchart corresponding to the above operation.

Referring to FIG. 18, parameter I is initialized to 1 in step S1, and then BAA data generation circuit 3
02i (302 _{1 in} this case) is driven. Further, in step S3, the timing Ti of the drive signal in the aperture group which is simultaneously driven by the BAA data generation drive circuit 302i is detected by the delay detection circuit 302f, and this is stored in the memory 302m. further,
By the steps S4 and S5, the above process is executed for all BAs.
It is executed for the A data generating circuit 302 _{1-302 128.}

Further, in step S6, the memory 30
Of the timings Ti stored in 2 m, the one having the maximum delay amount T _MAX is searched.

Next, in step S7, the parameter I
Are initialized to 1 again, and the delay amount Δ
Ti is calculated based on the delay amount T _MAX . Further, the delay amount ΔTi is stored in the memory 302m in step S9, and by executing steps S10 and S11,
Delay ΔTi is for all BAA data generating circuit 302 _{1-302 128} stored in the memory 302m. Further, on the basis of the delay amount ΔTi stored in the memory 302m in step S11, the BAA data generation circuit 302 _{1 is} formed on the BAA mask 220.
The delay amount of each delay circuit 302d is set so that the timings of the drive signals from ˜302 ₁₂₈ are aligned.

FIG. 19 shows an example of the delay circuit 302d used in the circuit of FIG.

[0090] Referring to FIG. 19, the delay circuit delay element 302d gives each different delay 302d ₁ to 30
2d _n and a switch SW for selecting one of the delay elements
And the switch SW is the control circuit 3
One delay element is selected according to the output control signal of 02c. Such delay element 302d ₁ ~302d _n is preferably one which provides a delay of about several tens of nm seconds at maximum.

FIG. 20 shows another example of the delay circuit 302d used in the circuit of FIG.

Referring to FIG. 20, the delay circuit 302d includes delay elements 302d ₁ ′ to 30 d each having a predetermined delay amount.
2d _n 'and a switch SW' for connecting a selected one of the delay elements in series, the switch SW '
Supplies a desired amount of delay to the input signal by connecting a predetermined number of delay elements in series according to the output control signal from the control circuit 302c. The configuration and operation of the delay circuit 302d are apparent from FIG. 20, and further description will be omitted.

Next, a fifth embodiment of the present invention will be described with reference to FIG.

In the present embodiment, the delay detection times shown in FIG.
A marker provided on the stage 126 instead of the path 302f.
Use a high-speed electron detector 126a such as a multi-channel plate.
BAA data generation circuit 302₁~ 302₁₂₈Out of
The BAA aperture on the mask 220 is driven by the applied drive signal.
As a result of turning on / off the char
Detect the child beam. The electron detector 126a is an electron beam
Pulse signal is generated according to the arrival of
Supply 2m. At that time, the BAA data generation circuit 302
₁~ 302₁₂₈Are sequentially driven, and the memory 302m is electronic.
The arrival timing of the beam is determined by each BAA data generation circuit
02₁~ 302₁₂₈Memorize corresponding to. Next, the system
The control circuit 302c is used for each BAA data generation circuit 302.₁
~ 302 ₁₂₈Drive detected by electronic detector 126
The timing of the signal is read out and B with the largest delay amount is read.
Each B based on the timing in the AA data generation circuit
Calculate the delay amount of the AA data generation circuit and write it down again.
It is stored in the memory 302f. When performing the actual exposure,
The control circuit 302c includes the delay circuits 302d.
The drive signal of each BAA data generation circuit.
The imming is the tie of the aperture group with the largest delay amount.
Control to match the ming. However, in this example
The delay circuit 302d is not the drive signal itself, but B
AA data generation circuit 302₁~ 302₁₂₈Operation of Taimi
It is designed to delay the clock
It The above operation is substantially the same as the flowchart described in FIG.
The detailed description will be omitted.

According to this embodiment, it is possible to actually measure the on / off timing of the electron beam by the BAA mask 220, and therefore, it is possible to perform the control with higher precision and reliability. Although illustration is omitted, in the configuration of FIG. 17 or 21, the memory 302m is omitted, the timing of each detection signal is compared with a predetermined reference timing in the control circuit 302c, and the difference is obtained, and the delay circuit is accordingly obtained. The delay amount of 302d may be set.

Further, the timing adjustment of the BAA exposure according to this embodiment is performed by 1 / N shown in FIG. 5, FIG. 7 or FIG.
The present invention is not limited to the BAA mask having the aperture group whose pitch is shifted, but can be applied to the conventional BAA mask 110 shown in FIG.

Furthermore, the present invention is not limited to the above embodiments, and various modifications within the scope of the present invention,
It can be changed.

[0098]

According to the present invention, N groups of opening groups are formed on the BAA mask by being shifted from each other by M / N pitches, and the opening groups are independently controlled so that they are mutually formed on the semiconductor substrate. M / N
By forming an exposure pattern composed of N groups of exposure dots displaced by a pitch by multiple exposure, it becomes possible to change the edge position of the exposure pattern with high accuracy,
At that time, there is no problem that the exposure accuracy is lowered in a specific multiple exposure pattern. Further, by providing a delay circuit corresponding to each drive circuit for driving the aperture on the BAA mask 220, it becomes possible to align the drive timing of the BAA aperture and expose a delicate and fine pattern with high accuracy. It will be possible.

[Brief description of drawings]

FIG. 1 is a diagram illustrating the principle of the present invention.

2A to 2C are views (No. 1) for explaining the operation of the present invention.

3 (D) and 3 (E) are views (No. 2) for explaining the operation of the present invention.

FIG. 4 is a diagram showing the displacement of an exposure edge with high accuracy achieved by multiple exposure according to the present invention.

FIG. 5 is a view showing a BAA exposure mask according to the first embodiment of the present invention.

6A to 6D are diagrams showing an example in which the present invention is applied to precision exposure of a rectangular pattern.

FIG. 7 is a diagram showing a BAA exposure mask according to a second embodiment of the present invention.

8A to 8E are diagrams showing an example of exposure using the exposure mask of FIG.

FIG. 9 is a diagram showing a structure of a BAA mask according to a third embodiment of the present invention.

10A and 10B are views showing exposure patterns exposed on a substrate using the mask of FIG.

11A to 11E are diagrams showing fine adjustment of the edge position of the exposure pattern on the substrate using the mask of FIG.

12A to 12C are diagrams showing an example in which the present invention is applied to correction of a proximity effect.

FIG. 13 is a diagram showing another example in which the present invention is applied to the correction of the proximity effect.

FIG. 14 is a diagram showing a configuration of an electron beam exposure apparatus used in the present invention.

15 is a diagram showing an example of an exposure pattern exposed by the exposure apparatus of FIG.

16 is a diagram showing a configuration of a data expansion unit in the exposure apparatus of FIG.

FIG. 17 is a block diagram showing a configuration of a fourth exemplary embodiment of the present invention.

18 is a flowchart showing an exposure operation with the configuration of FIG.

19 is a diagram showing a configuration of a delay circuit used in the circuit of FIG.

20 is a diagram showing another example of the configuration of the delay circuit used in the circuit of FIG.

FIG. 21 is a block diagram showing a configuration of a fifth exemplary embodiment of the present invention.

FIG. 22 is a diagram showing a configuration of a conventional electron beam exposure apparatus.

23 is a diagram showing a configuration of a BAA mask used in the electron beam exposure apparatus of FIG.

FIG. 24 is a diagram showing a detailed configuration of the mask of FIG. 23.

FIG. 25 is a diagram showing an exposure pattern exposed on the semiconductor substrate by the mask of FIG. 23.

FIGS. 26A to 26E are views for explaining problems that occur in exposure using the mask of FIG.

27 is a diagram showing a shift in exposure timing that occurs when exposure is performed using the masks shown in FIGS.

[Explanation of symbols]

X ₁ 1st exposure pattern X ₂ 2nd exposure pattern a _{1 to} a ₆ , a to d exposure dot b _{1 to} b ₅ exposure dot P _{0 to} P ₄ exposure edge position TH development threshold value EX ₁ , EX ₂ , A , B, C, D Opening group 1A, 1B, 2A, 2B, ..., 8A, 8B, A ₁ ,
A ₂ , ..., D ₁ , D ₂ aperture array 1A _{1 to} 1A _n , 1B _{1 to} 1B _n , ... aperture 100 electron optical system 101 cathode 102 grid 103 anode 104 electron gun 105 aperture plate 105a beam Shaping aperture 107, 108, 116, 119 Electron lens 110, 220, 220 'BAA mask 122 Objective lens 123 Substrate 124, 125 Deflector 126 Stage 126a Electron detector 200 Exposure control system 201-203 Storage device 204 CPU 205 Interface 206 memory 207,209,210,212,215 control circuit 214 laser interferometer 215 stage moving mechanism 211,216,217 driving circuit 301 driving circuits 302 ₁ to 302 ₁₂₈ BAA data output unit 302c control circuit 302d Extending circuit 302f timing detection circuit 302m memory 302x clock generator 303 the data expansion unit 304 CPU 305 exposure control unit 307 sub-scanning control device 308 main scanning control unit 308 stage controller 309 autoloader 310 main deflection control

Claims (18)

[Claims] 1. A beam shaping step of shaping a single electron beam to form an electron beam pattern composed of a plurality of electron beam elements, and a predetermined exposure pattern on an object, and a plurality of electrons constituting the electron beam pattern. In an electron beam exposure method including an exposure step of drawing as a set of exposure dots arranged at a predetermined pitch by irradiation of a beam element, the beam shaping step, the plurality of electron beam elements,
Each of the first electron beams is shaped by shaping the single electron beam.
A second line aligned in the direction of and perpendicular to the first direction.
A step of forming as a plurality of electron beam groups consisting of a set of mutually parallel electron beam elements repeatedly formed in the direction of, and forming each of the plurality of electron beam elements simultaneously. Between different electron beam groups, the electron beam elements have at least one of the first direction and the second direction, where N is an integer representing the number of electron beam groups and M is an arbitrary integer smaller than N. An electron beam exposure method characterized in that they are offset from each other by an M / N pitch. 2. The beam shaping step is performed so that electron beam elements forming the electron beam group are displaced by M / N pitch in the first direction between one electron beam group and another electron beam group. The electron beam exposure method according to claim 1, wherein: 3. The beam shaping step is performed so that electron beam elements forming the electron beam group are displaced from each other by M / N pitch in the second direction between one electron beam group and another electron beam group. The electron beam exposure method according to claim 1, wherein: 4. In the beam shaping step, the electron beam elements forming the electron beam group are M in one electron beam group and another electron beam group in the first direction and the second direction.
The electron beam exposure method according to claim 1, wherein the electron beam exposure is performed so as to be shifted by / N pitch. 5. The beam shaping step is repeated at a pitch of 2 pitches in the first direction, and a first exposure dot row composed of exposure dots repeated at a pitch of 2 pitches in the first direction, and And a step of forming exposure dots such that a second exposure dot row composed of exposure dots that are offset by one pitch in the first direction with respect to the first exposure dot row is alternately repeated. The electron beam exposure method according to claim 1. 6. The exposing step comprises scanning the plurality of electron beam elements aligned in the first direction on the object in the second direction in all of the plurality of electron beam groups, 2. The electron beam exposure according to claim 1, further comprising the step of exposing a row of exposure dots formed of the exposure dots on the object, corresponding to each of the plurality of electron beam groups, a plurality of times of overlapping exposure. Method. 7. In at least one of the steps of overlappingly exposing the exposure dot row a plurality of times corresponding to the plurality of electron beam groups, the exposure dot row to be exposed is the exposure dot row previously exposed. 5. The electron beam exposure method according to claim 4, wherein the two have different exposure patterns. 8. The electron beam exposure method according to claim 7, wherein the exposure dot rows having different exposure patterns have exposure dot densities different from those of other exposure dot rows. 9. An electron beam source means for forming an electron beam and emitting the electron beam along a predetermined optical axis; a focusing means formed on the predetermined optical axis for focusing the electron beam on an object; An electron beam shaping means formed on a predetermined optical axis for shaping an electron beam to form a plurality of electron beam elements; and controlling the electron beam shaping means to expose the plurality of electron beam elements in a predetermined exposure pattern. In the electron beam exposure apparatus, which comprises a shaping control means for forming an exposure pattern on the object by the electron beam elements, the shaping control means for forming the plurality of shaped electron beam elements according to the above; , A mask plate formed with a plurality of openings to shape a single electron beam into a plurality of electron beam elements corresponding to the plurality of openings, and a shape for each of the plurality of openings. The shaping control means supplies a driving signal to the deflecting device according to an exposure pattern so as to expose the plurality of electron beam elements to the predetermined exposure. A plurality of openings formed on the mask plate are arranged at a predetermined pitch in a first direction and a second direction different from the first direction, and each of the first openings is formed in the first direction.
Forming a plurality of opening groups extending in the first direction and repeating N times at the predetermined pitch in the second direction; in each of the opening groups, each opening is arranged in the first direction in the predetermined direction. Repeated by a pitch, and for corresponding openings in adjacent openings, N is the number of the openings and M is an arbitrary integer smaller than N, and the distance is M / N of the predetermined pitch. The shaping control means supplies a plurality of drive signals independent of each other to each of the deflecting devices formed in the plurality of opening groups. An electron beam exposure apparatus characterized by the above. 10. The plurality of opening rows, each of which is composed of a plurality of openings aligned in the first direction, extends in the first direction, and is repeated in the second direction. The electron beam exposure apparatus according to claim 9, further comprising: 11. The first opening row having a plurality of openings arranged in the first direction at a pitch of 2 pitches, and the second opening row having a second pitch in the first direction. Are arranged at intervals, and yet the first row of openings is provided with the first
11. The electron beam exposure apparatus according to claim 10, further comprising a second opening row formed of a plurality of openings formed by being shifted by one pitch in the direction of. 12. The electron beam exposure apparatus according to claim 9, wherein the shaping control unit includes a delay unit that aligns the timings of the plurality of drive signals with a predetermined timing. 13. A beam shaping step of shaping a single electron beam to form an electron beam pattern composed of a plurality of electron beam elements, and a predetermined exposure pattern on an object, a plurality of electrons constituting the electron beam pattern. An electron beam exposure method comprising: an irradiation step of writing as a set of exposure dots arranged at a predetermined pitch by irradiation of a beam element: the beam shaping step turns on or off the single electron beam according to the electron beam pattern. A step of passing a mask having a plurality of openings formed therein; the beam shaping step further includes a timing detection step of detecting a drive timing at which the openings are turned on and off; and the drive timing at a predetermined timing. An electron beam exposure method, which comprises a timing adjusting step of aligning. 14. The beam shaping step aligns each of the plurality of electron beam elements in a first direction and further comprises:
Forming a plurality of electron beam groups consisting of a set of mutually parallel electron beam elements repeatedly formed in a second direction perpendicular to the direction of The electron beam elements are formed so that N is an integer representing the number of electron beam groups, M is an arbitrary integer smaller than N, and the electron beam elements are displaced from each other by at least M / N pitches in the first direction. 14. The electron beam exposure method according to claim 13. 15. The timing detection step includes a step of obtaining the drive timing for each of a plurality of opening groups including a plurality of openings and being turned on and off in synchronization with each other, and the timing adjustment step includes the timing detection. A step of storing the drive timing obtained in the step in a memory; a step of searching for an opening having the largest delay amount based on the drive timing stored in the memory; 14. The electron beam exposure method according to claim 13, further comprising a delay step of delaying the drive timing of the opening group. 16. The timing detection step includes a step of obtaining the drive timing for each of a plurality of opening groups including a plurality of openings and turned on and off in synchronization with each other, and the timing adjustment step includes the timing detection. A step of comparing the drive timing obtained in the step with the predetermined timing to obtain a delay amount; and a delay step of delaying the drive timing of the plurality of opening groups based on the delay amount so as to match the predetermined timing. 14. The electron beam exposure method according to claim 13, comprising: 17. The timing detecting step includes a step of detecting a timing of a drive signal which is supplied to the opening and turns on and off the opening.
The electron beam exposure method described. 18. The electron beam exposure method according to claim 13, wherein the timing detection step includes a step of detecting the timing of the electron beam with which the object is irradiated.