JPH05175108A - Method and apparatus for exposing with charged particle beam - Google Patents

Abstract

PURPOSE:To correct a proximity effect by providing an actual exposure step of exposing a desired pattern and an inversion exposure step of inverting expo sure information to be used for the actual exposure as it is, accumulating it in a shift register and exposing in an intensity weaker than that by the actual exposure. CONSTITUTION:Pattern data of a pattern to be exposed is supplied to a shift register 1, and sequentially fed from the register 2 to an exposure device for independently controlling a plurality of charged particle beams to expose the pattern. In this case, an output of the register 1 is also supplied to an inverter 2, and fed back to the register 1. After the pattern data is fed from the register 1 to perform an actual exposure for exposing a pattern to be exposed, the inverted pattern data is fed from the register 1, and inversion-exposed. This inversion exposure is executed in an intensity lower than that of the actual exposure in an out-of-focus state. Thus, a proximity effect correction can be efficiently performed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a charged particle beam exposure method, and more particularly to a charged particle beam exposure method in which the proximity effect is corrected.

In recent years, ICs, which are expected to play a role as nuclear technology for technological progress in all industries, will take about 4 years in a few years.
Achieved double high integration. For example, in DRAMs, 1M, 4M, 16M, 64M, 256M, 1G and the like are being integrated.

Such high integration is solely due to the progress of fine processing technology. Optical technology continues to advance to enable fine processing down to 0.5 μm. However, the limit of optical technology is about 0.3 μm. In addition, it is becoming more and more difficult to ensure accuracy in opening a window of a contact hole, aligning it with an underlying pattern, and the like.

In electron beam exposure, fine processing of 0.1 μm or less can be realized with alignment accuracy of 0.05 μm or less. Further, in recent years, by the block exposure and the blanking aperture array type exposure by the present inventors,
1cm ² / 1sec about throughput it can now be expected.

In other words, the electron beam exposure has an outstanding advantage as compared with other lithography techniques regardless of whether the fineness, the alignment accuracy, the quick turnaround, the reliability, or the future due to the improvement of the software are taken.

[0006]

2. Description of the Related Art Electron beam exposure is generally performed using a point beam, a variable rectangular beam, a block pattern beam, or the like. In either case, it is necessary to deflect and position the electron beam to expose the desired area on the wafer.

An electron beam exposure apparatus capable of block exposure will be described with reference to FIG. The exposure apparatus is roughly divided into an exposure unit 10 and a control unit 50. Exposure unit 10
Is a portion for generating an electron beam, shaping it into a spot shape or a pattern shape, and exposing it at a desired position on an exposure object. The control unit 50 is a unit that forms a signal for controlling the exposure unit 10. Under the exposure unit 10, the exposure target W
There is a stage 35 on which is mounted.

First, the exposure section 10 will be described. The electrons generated from the cathode electrode 11 are extracted by the grid electrode 12 and the anode electrode 13. These electrodes 11, 12 and 13 form a charged particle beam generation source 14.

The electron beam generated from the charged particle beam generation source 14 is shaped by, for example, a first slit 15 having a rectangular opening, passes through a first electron lens 16 that focuses the electron beam, and is transmitted onto a transmission mask 20. The beam is incident on the slit deflector 17 for correcting and deflecting the beam irradiation position. The slit deflector 17 receives the corrected deflection signal S
Controlled by 1.

In order to shape the electron beam into a desired pattern, a transmission mask (stencil mask) 20 having a plurality of transmission holes such as a rectangular opening and a block pattern opening having a predetermined pattern is used. The electron beam that has passed through the slit deflector 17 has a second electron lens 1 that is provided to face it.
8, a third electron lens 19, a transmission mask 20 movably mounted between these electron lenses in the horizontal direction, and arranged above and below the transmission mask 20.
1 to P4, the electron beam is deflected according to the transmission mask 2
First to fourth deflectors 21 for selecting one of the transmission holes
Shaped into a desired pattern through an electron beam shaping section including .about.24.

The shaped electron beam is blocked or passed by the blanking electrode 25 to which the blanking signal SB is applied. The electron beam that has passed through the blanking electrode 25 has a fourth electron lens 26, an aperture 27, a refocusing coil 28, and a fifth electron lens 29.
And is incident on the focus coil 30.
The focus coil 30 has a function of focusing the electron beam on the exposure target surface. Further, the stig coil 31 corrects astigmatism.

The electron beam is further transmitted to the sixth electron lens 3
2. The exposure target W according to the exposure position determination signals S2 and S3
The position is controlled by the main deflector 33 that performs the upper positioning and the sub deflector 34 that is an electrostatic deflector, and the desired position on the exposure target W is irradiated.

The object W to be exposed is placed and moved on a stage 35 which is movable in the XY directions. The exposure unit 10 is further provided with first to fourth alignment coils 36, 37, 38, 39.

The control unit 50 has a memory 51 and a CPU 52. Design data of the integrated circuit device is stored in the memory 51, read by the CPU 52, and processed.
The CPU 52 controls the whole other charged particle beam exposure apparatus.

The interface 53 receives drawing information fetched by the CPU 52, for example, drawing position information on the wafer W where a pattern is to be drawn, and the transparent mask 20.
Various types of information such as the mask information of are transferred. The data memory 54 stores and holds the drawing pattern information and the mask information transferred from the interface 53.

The pattern controller 55 receives the drawing pattern information and the mask information from the data memory 54, designates one of the transmission holes of the transmission mask according to the information, and indicates the position of the designated transmission hole on the transmission mask. Designating means, holding means, computing means, and output means for generating various signals P1 to P4 and performing various processing including processing for computing a correction value H according to the shape difference between the pattern shape to be drawn and the designated transmission hole shape. Have.

An amplifier unit 56 having a digital / analog converter function and an amplifier function receives the correction value H and generates a corrected deflection signal S1. The mask moving mechanism 57 moves the transparent mask 20 as needed according to a signal from the pattern controller 55.

The blanking control circuit 58 controls an amplifier section 59 having a digital / analog converter function and an amplifier function in response to a signal from the pattern controller 55 to generate a blanking signal SB.

The sequence controller 60 receives the drawing position information from the interface 53 and controls the drawing processing sequence. The stage moving mechanism 61 moves the stage 35 as needed in response to a signal from the sequence controller 60.

The movement of the stage 35 is detected by the laser interferometer 62 and supplied to the deflection control circuit 63. The deflection control circuit 63 calculates the exposure position on the wafer W and generates the exposure position determination signals S2 and S3.
The signals are supplied to the signals 4 and 65, and also to the sequence controller 60. The amplifier units 64 and 6
5 has a digital / analog converter function and an amplifier function, respectively.

In the ordinary electron beam exposure, the main deflector 33, which is an electromagnetic deflector, is used for 2 to 10 mm.
A deflection field of □ is beam-deflected, and a sub-field of about 100 μm □ is deflected by the sub deflector 34 which is an electrostatic deflector.

The pattern data is read from the memory 51 by the CPU 52, transferred to the data memory 54, and stored therein. The pattern controller 55 decomposes the pattern for each shot according to the pattern data read from the data memory 54.

The pattern data decomposed into each shot is the data for the main deflector 33, the data for the sub deflector 34, the data for the slit deflector 17,
It is separated into a blanking signal SB and the like to control the deflection of the electron beam.

By the way, charged particles are injected into a resist layer formed on an object such as a semiconductor wafer, and the resist is exposed. The charged particles incident on the resist undergo multiple scattering as they progress.

When a resist layer is formed on a silicon substrate and charged particle beam exposure is carried out, the charged particles incident on the resist layer undergo forward scattering while advancing in the resist layer,
Proceeds from the resist layer into the silicon substrate. The charged particles returning from the substrate by multiple scattering and re-incident on the resist layer are further scattered in the resist layer to form backscattering.

Due to such forward scattering and back scattering, an incidental pattern due to scattering is formed around the desired exposure pattern. When the patterns to be exposed are dense, incidental patterns due to scattering from the respective patterns overlap with each other, and the exposure intensity exceeds the development threshold value.

As a result, a phenomenon occurs in which a pattern larger than the desired pattern size is drawn. This phenomenon is more prominent when the patterns are dense and is called a proximity effect.

FIG. 6 is a diagram for explaining the proximity effect. FIG. 6A is a graph showing electron tracks when a resist layer 72 made of PMMA is placed on a silicon substrate 71 and electrons are irradiated from above. The diagram on the left side of FIG. 6 (A) shows a track when electrons are irradiated with acceleration energy of 10 kV, and the diagram on the right side of FIG. 6 (A) shows a track of electrons when the acceleration energy is 20 kV.

The tracks of 100 electrons are obtained by simulation by the Monte Carlo method. The horizontal axis of the graph represents the distance from the electron irradiation position of the resist layer 2 in microns, and the vertical axis represents the depth from the surface of the resist layer 2 in microns.

As is apparent from the figure, the electrons irradiated on the resist layer 2 spread to a range of about 2 μm when the acceleration energy is 10 kV due to forward scattering and back scattering.
When the acceleration energy is 20 kV, it spreads to about 4 μm or more. The proximity effect actually observed is 3 to 5 μm.
It has a range of width.

FIG. 6B is a graph schematically showing the distribution of exposure intensity generated by such electron beam exposure. In the portions of the exposure patterns P1, P2, P3,
The exposure intensity is high, and tail portions T1, T2, and T3 due to forward scattering and back scattering are formed in the periphery thereof. The strength of the tail portion T depends on the pattern area,
A wide pattern is high, and a narrow pattern is low.

If the acceleration energy of the charged particle beam is constant, the spread of the tail pattern T due to scattering is almost constant. The tail T spreading from each pattern P to the periphery becomes weaker in intensity as the distance from the pattern increases.

By the way, the tail portions T1 and T2 between the patterns P1 and P2 shown in the figure overlap each other, and the influence on the resist is the sum thereof. Therefore, in the portion where the patterns are dense, the tail portions from the respective portions may overlap with each other and the development threshold value may be exceeded.

FIG. 6C is a diagram for schematically explaining the proximity effect caused by the overlapping of tail portions due to forward scattering and back scattering. As shown on the left side of FIG. 6C, when the parallel rectangular patterns P5 and P6 are exposed, a design pattern is created according to the desired pattern, and the charged particle beam exposure is performed according to the design pattern. The subsequent patterns become patterns P5a and P6a shown on the right side.

That is, in the central portion of the gap between the patterns P5 and P6, the tail portions from the respective pattern portions are strongly overlapped with each other, and the pattern width becomes wider than the designed value. In this way, when the proximity effect occurs, a pattern having a desired shape cannot be obtained.

A method for correcting the proximity effect in advance to obtain a pattern having a shape as designed is called proximity effect correction. FIG. 7 is a diagram for explaining a proximity effect correction method according to a conventional technique.

FIG. 7A schematically shows a case where the proximity effect is not corrected and the proximity effect occurs between the figures. When trying to expose the rectangular patterns P5 and P6,
As shown on the right side of FIG. 7A, the pattern spreads in the central portion, and the expanded patterns P5a and P6a are connected at the central portion P7. A method for correcting such a proximity effect will be described below.

FIG. 7B shows a method of correcting the proximity effect by changing the irradiation intensity of the charged particle beam. When the patterns P5 and P6 are exposed, if the entire area in the pattern is exposed with the same intensity, a proximity effect occurs as shown in FIG. 7 (A).

Therefore, the proximity effect is corrected by reducing the irradiation intensity of the charged particle beam in the portion close to another pattern. For example, as shown in the figure, a sample point X is set at the center of the opposite sides of the patterns P5 and P6, and the charged particle beam irradiation dose including scattering caused by the surrounding pattern at these sample points is calculated,
The doses of the adjacent portions P8 and P9 are adjusted so that the desired dose is obtained.

By reducing the irradiation amount of the charged particle beam in the proximity portions P8 and P9, the scattering amount of the charged particle beam in the region P7 between the patterns is reduced and the proximity effect can be corrected.

FIG. 7C is a diagram for explaining a figure deletion method which is another proximity effect correction method. Pattern P5
When P6 and P6 are exposed as designed, each pattern becomes larger than the designed value, and the proximity effect as shown in FIG. 7A occurs. Therefore, the pattern is reduced to a design value or less in advance, and the pattern after exposure is adjusted to have a desired value including scattering.

For example, a sample point X is set at the center of the opposite sides of the patterns P5 and P6, the charged particle beam irradiation dose including the scattering at these points is calculated, and the opposite sides are adjusted so that the desired pattern size is obtained. Support the proximity areas P10 and P11. As a result, the pattern thickened by the proximity effect forms patterns P5 and P6 having a desired width.

However, since the correction is performed for every existing pattern in both the figure deleting method and the irradiation intensity changing method, the time required for the correction process increases rapidly as the pattern increases. Further, in both correction methods, the representative points are set for each pattern and the correction is performed based on the exposure intensity value at those points. Therefore, if the number of the representative points is small, the distortion of the pattern due to the indirect effect cannot be completely corrected. Increasing the number of representative points leads to an increase in processing time.

Further, regarding the block exposure method for collectively transferring the complicated repeating pattern formed on the mask,
In practice, it is extremely difficult to carry out the method of deleting a figure and the method of changing the irradiation intensity for correcting each pattern.

FIG. 7D shows a ghost exposure method which is another proximity effect correction method. The ghost exposure method uses a main pattern for forming an exposure pattern and an auxiliary exposure pattern that is a black-and-white inverted pattern of the main pattern. The auxiliary exposure is usually performed with an intensity of about 30 to 50% of the actual exposure for exposing the main pattern.

After exposing the main pattern with a desired exposure intensity, the auxiliary exposure pattern is exposed with an intensity corresponding to the intensity of the scattering pattern additionally formed by the exposure of the main pattern. When such overlapping exposure is performed, almost uniform exposure is performed on the entire surface except for the main pattern, and only the main pattern can be developed by adjusting the development level.

The present inventors have proposed a blanking aperture array exposure method in which exposure is performed while controlling a large number of openings independently. FIG. 8 schematically shows the blanking aperture array. FIG. 8A shows a schematic plan view of the blanking aperture array. Light-shielding substrate 80
A large number of openings 81 are formed therein.

For example, 64 openings 81 are arranged in the row LA1 shown at the top of the figure, and 64 openings 81 are arranged in a staggered pattern in the next row LB1.
Further, when the two opening rows LA1 and LB1 are combined into one set, eight opening rows are arranged in the vertical direction in the drawing.

The pattern on the mask is reduced to 1/500 on the sample surface. On the mask, each opening 81 has a size of 25 μm □ and has a width of 50 μm.
They are arranged at a pitch of μm, and the vertical pitch is set to 100 μm. However, with respect to the vertical direction, another one row of opening columns arranged in a zigzag pattern is arranged in the middle of the 100 μm pitch.

The blanking aperture array as a whole is approximately 3200 μm in the horizontal direction in the figure and approximately 80 in the vertical direction in the figure.
The dimension is 0 μm. The opening of 25 μm □ has a size of 0.05 μm □ on the object.

By irradiating the charged particle beam through the opening 81 while moving such a blanking aperture array in the vertical direction in the figure, the entire area on the object can be exposed by the pair of opening rows LA1 and LB1. With eight sets of aperture rows, the entire surface can be subjected to multiple exposure in eight layers.

In each opening 81, electrodes 82a and 82b are formed along two opposite sides, and these electrodes 82 are formed.
The electron beam passing through the opening 81 can be deflected to a region outside the object by the voltage applied to a and 82b.

That is, the electrodes 82a and 82b have openings 81.
It can serve as a shutter for the electron beam passing through the. One blanking aperture BA is formed by the opening 81 and the electrodes 82a and 82b formed on both sides of the opening 81.

In the above structure, 128 × 8 blanking apertures BA are formed. When viewed in the vertical direction, the eight blanking apertures BA are arranged at the same horizontal position, so that the same position on the object can be exposed eight times.

FIG. 8B is a graph for explaining the exposure method in the case where the blanking aperture array as shown in FIG. 8A is used to perform exposure on the sample surface. Figure 8
In (B), the horizontal axis represents time in nsec, and the vertical axis represents distance on the sample surface in μm.

For the blanking aperture array,
It is assumed that the sample is moving at a constant velocity of 0.5 μm / 50 nsec. The curve indicated by LA1 is one blanking aperture BA belonging to the column LA1 shown in FIG.
Shows the exposure target of.

In the time up to 5 nsec, the position on the sample surface from the reference position to 0.05 μm is exposed to light for 5 nsec.
To 10nsec, 0.05μ on the sample surface
The area between m and 0.1 μm is exposed.

In this way, by exposing the area exposed under the blanking aperture sequentially every 5 nsec, one row of exposure stripes can be formed on the sample surface.

By the way, in FIG. 8A, when the sample is moving from the upper side to the lower side, the same vertical position on the sample surface is below the other blanking apertures which are staggered after 10 nsec. Appear in.

For example, the area that can be exposed by one blanking aperture array LA1 is a region having a width of 0.05 μm and a pitch of 0.1 μm, and what is formed after exposure is a line-and-space having an equal pitch.

In order to fill the entire surface, it is necessary to perform exposure with the blanking aperture array LB1 in the next stage, which is arranged complementarily to the blanking aperture array LA1. A region appearing below the blanking aperture array LB1 in the second row is indicated by a broken line LB1.

After 20 nsec, the region exposed under the blanking aperture array LA1 of the first column at time 0 is exposed under the blanking aperture array LA2 of the third column. Similarly, 40nsec, 60nsec, ...
The same position is exposed under the blanking aperture BA arranged at the succeeding position. By performing the exposure at each time point, the same position is multiply exposed.

According to such multiple exposure, the exposure of the same target area is divided into, for example, 8 steps. Current on /
The unit of OFF becomes smaller, and the electron beam as a whole gradually increases and gradually decreases. That is, a rapid change in the exposure current is prevented, and refocusing for compensating the Coulomb interaction is facilitated.

FIG. 9 is a schematic diagram for explaining exposure by the blanking aperture array. For simplicity, a case where the target pattern is exposed by only one exposure will be described. In the case of multiple exposure, similar exposure may be repeated.

As shown in FIG. 9A, it is assumed that blanking apertures BA1 to BA5 are arranged in two rows in a staggered pattern. The pattern to be exposed is assumed to be a figure as shown in FIG. That is, it is considered that an area as shown in FIG. 9B is exposed on the object. Here, it is assumed that the object moves downward from above the blanking aperture shown in FIG.

First, as shown in FIG. 9C, the lower end of the area to be exposed reaches under the blanking apertures BA1 and BA2. Where the blanking aperture BA
1 and BA2 are turned on, and the corresponding area is exposed.

At the next timing, as shown in FIG. 9D, the upper three blanking apertures BA
Areas to be exposed appear below 1, BA2 and BA3. Therefore, these areas are exposed.

At the next timing, as shown in FIG. 9 (E), the upper blanking apertures BA1 and BA are shown.
The area to be exposed appears below 2. Therefore, these areas are exposed.

At this time, the blanking aperture B
Areas to be exposed are arranged in the areas adjacent to A1 and BA2, but since there is no blanking aperture in the areas corresponding to these areas, exposure is not performed.

In this manner, the region exposed under the blanking aperture is selectively exposed as the target sample is moved downwardly under the blanking aperture BA.

After the three regions are exposed in FIG. 9F,
Proceeding to FIG. 9G, the two areas that were not exposed in FIG. 9E are arranged below the blanking apertures BA4 and BA5. Therefore, at this stage, these areas are exposed by the blanking apertures BA4 and BA5.

That is, among the patterns in the five columns, 1, 3,
The fifth row is exposed using the blanking apertures BA1, BA2, BA3 shown in the upper part of FIG. Exposed.

In this way, FIG. 9 (H), (I),
As the object moves in sequence with (J), the area that can be exposed is exposed, and the pattern shown in FIG. 9B is formed.
Thus, according to the blanking aperture array,
Any figure can be exposed.

The blanking aperture array is compatible with the block exposure, not in opposition to the block exposure. That is, a block pattern and a blanking aperture array may be provided together on a stencil mask, a pattern with a high appearance frequency may be exposed by block exposure, and a pattern with a low appearance frequency may be exposed by a blanking aperture array.

[0075]

As described above,
According to the blanking aperture array, an arbitrary pattern can be exposed by controlling each blanking aperture while irradiating a predetermined region on the sample with the charged particle beam.

However, in the charged particle beam exposure, auxiliary exposure for correcting the proximity effect is necessary. By the way, using a blanking aperture array,
The method of executing the auxiliary exposure has not been established yet.

An object of the present invention is to correct the proximity effect in an exposure technique for exposing a pattern on a target surface by independently controlling ON / OFF of a plurality of charged particle beams such as a blanking aperture array system. It is to provide the technology that can do it.

[0078]

According to a charged particle beam exposure method of the present invention, a plurality of charged particle beams are independently on / off controlled according to exposure information to irradiate a target surface to expose a desired pattern. It includes an actual exposure step to be performed and an inverse exposure step in which the exposure information used for the actual exposure is inverted as it is and accumulated in a shift register to perform exposure having a weaker intensity than the actual exposure.

[0079]

The information of the reversal pattern can be obtained by directly reversing the exposure information used when exposing a desired pattern on the target surface by independently controlling the on / off of a plurality of charged particle beams.

By accumulating the inversion pattern information in the shift register and performing the inversion exposure having a weaker intensity than the actual exposure for exposing the desired pattern, the auxiliary exposure for correcting the proximity effect can be performed.

[0081]

1 shows a charged particle beam exposure method according to a basic embodiment of the present invention. As shown in FIG. 1 (A), pattern data of a pattern to be exposed is supplied to a shift register 1 and is sequentially sent from the shift register 1 to an exposure device for exposing a pattern by independently controlling a plurality of charged particle beams. Be done. At this time, the output of the shift register 1 is also supplied to the inverting circuit 2, and the pattern data inverted by the inverting circuit 2 is fed back to the shift register 1.

After the pattern data is sent out from the shift register 1 and the actual exposure for exposing the desired pattern is performed, the reverse pattern data is sent out from the shift register 1 and the reverse exposure is performed.

The exposure on the object is carried out as shown in FIG. The exposure target area 3 on the object is divided into a plurality of unit strip areas 4a, 4b, 4c, ... And pattern data for exposing each unit strip area is transferred from the shift register 1 shown in FIG. Supplied. According to this pattern data, the actual exposure for exposing the unit strip area 4a is performed as indicated by the arrow 5a indicated by the solid line.

Subsequently, the inverted pattern data inverted by the inversion circuit 2 is supplied from the shift register 1 to the exposure device, and the auxiliary exposure of the unit area 4a is performed as shown by the arrow 6a indicated by the broken line. It should be noted that the reversal exposure is performed with an exposure intensity lower than that of the actual exposure and with a defocused state.

FIG. 1C is a graph showing changes in focus when performing actual exposure and reverse blur exposure. When the actual exposure 5a is performed with the charged particle beam focused, the reverse blur exposure 6a is performed with the charged particle beam defocused. Subsequently, the actual exposure 5b is executed in a focused state, the focus is defocused again at the end of the actual exposure, and the reverse blur exposure 6b is executed. In this way, the actual exposure and the reverse blur exposure are performed while changing the focus state.

FIG. 2 is a block diagram showing the structure of an electron beam exposure apparatus according to a more specific embodiment of the present invention.
An exposure system 10 of the electron beam exposure apparatus includes a blanking aperture array 8 and pattern data is supplied to each blanking aperture of the blanking aperture array 8 via a delay circuit D. That is, the same data is supplied to the openings of one row of the blanking aperture array at different timings.

In order to supply the pattern data to the delay circuit D at high speed, the delay circuit D has two sets of shift registers 1A,
1B is connected. While using one shift register, the other shift register can be ready for data. For example, 16 sets of scrollers S1, S2, ..., S16 are connected to these two sets of shift registers 1A and 1B, and one set of data can be alternately supplied to the shift registers 1A and 1B.

Each scroller S is basically a circuit having the same structure as the shift register 1. Scroller S
In order to supply the pattern data to each of the scrollers, a combination of four sets of pattern data developing devices PD and memories SM such as SRAMs is connected to each scroller. That is, the pattern data expansion device PD and the static RAM
4 × 16 = 64 combinations are provided in total.

Each pattern data developing device PD receives the pattern data from the data memory, creates one unit of pattern data to be exposed, and stores the pattern data in the memory SM formed by SRAM. 4 sets of memory S
M1 to SM4 supply pattern data to the corresponding scrollers S alternately.

Inverters 2a and 2b are connected to the shift registers 1A and 1B, respectively, and the output signals of the shift registers can be inverted and fed back to the shift register inputs.

Therefore, when the pattern data is supplied from the data expansion device PD to the shift register 1 through the scroller S and is supplied from the shift register 1 to the delay circuit D, the inversion data of the pattern data is automatically converted. 2 and is fed back to the shift register 1 to output inverted pattern data following the pattern data.

Incidentally, the exposure system 10 can be constituted by an exposure system as shown in FIG. 5 provided with a blanking aperture array. If the blanking aperture array has a 128 × 8 blanking aperture as shown in FIG. 8, the pattern data supplied to the blanking aperture array 8 is in units of 128 × n. Here, n is a number defining the length of the unit strip area 4 shown in FIG.

FIG. 3 shows an example of the structure of a scroller capable of transmitting such 128 × 2000 pattern data at high speed. FIG. 3A is a block diagram showing the circuit configuration of the scroller. 128 pattern data IN
A scroller capable of receiving 0 to IN127 in parallel and accommodating 2000 in series is a 128 × 2000 shift register sr connected as shown in the figure.

That is, 2000 shift registers s
128 sets of r series connections are arranged in parallel. 128
128 input signals IN0 to IN127 in parallel,
The individual output signals EX0 to EX127 can be supplied in parallel.

FIG. 3B is a circuit diagram showing a configuration example of each shift register sr. The output signal of the transistor Tr1 is applied to the gate electrode of the second-stage transistor Tr2 via the gate formed of the transistor TrA.

For example, if the gate voltage of the transistor Tr1 is high, the output of the transistor Tr1 becomes low, and when the gate TrA opens, the transistor Tr2 receiving this signal is turned off and its output becomes high.
Thus, the two transistors Tr1 and Tr2 form a shift register of one stage. The output signals of the transistors Tr1 and Tr2 are the gate Tr.
It is selectively transmitted to the output side by A and TrB.

Here, an operation example for performing exposure by the blanking aperture array at a practical speed will be briefly described. Consider the case where exposure is performed at 1 cm ² / sec, assuming that the size of one electron beam is 0.05 μm □. An area of 1 cm ² includes 4 × 10 ¹⁰ beams of 0.05 μm □. 4 × 10 per second
^In order to perform ¹⁰ exposures, the number of beams is 128, for example.
× 8 is used.

The stage on which the sample is placed is about 5 in the Y direction.
Assuming that the electron beam travels at 0 mm / sec, the electron beam as a whole is 2 mm / 2 nsec in the X direction due to the main deflector.
You can move with.

The sub deflector moves in the Y direction by 100 μm five times every 5 μsec. At this time, the scanning speed is 100 μm / 5 μsec, and the time required for scanning per 1 μm is 50 nsec. This value is 2. with respect to the length per beam of 0.05 μm.
The length is 5 nsec, and one shot is 2.5 nsec.
Need to be processed.

In the case of the above-described structure in which the blanking aperture array is arranged in eight layers in the beam traveling direction, eight shots are performed to expose one point. It takes 8 shots to complete the exposure at the same position, ie 2
It takes 0 nsec. Therefore, when refocusing the electron beam intensity, the time required for refocusing is 20 nsec instead of 2.5 nsec.

When the area to be exposed is 0.5 cm ² / sec, the stage moving speed can be halved, and the sub-deflector takes 5 μs to expose the same area as described above.
It becomes possible to use 10 μsec which is twice the ec. In this case, the time required for one shot is doubled to 5
μsec. According to the configurations shown in FIGS. 2 and 3, it is possible to perform such high-speed exposure.

FIG. 4 schematically shows an exposure method in which an actual exposure for exposing a desired pattern and a ghost (auxiliary) exposure for correcting the proximity effect are performed by using the configuration as shown in FIGS. Show. FIG. 4A shows the actual exposure, and FIG.
Indicates ghost exposure.

In the case of actual exposure, as shown in FIG. 4A, exposure is performed by using all blanking apertures in which 128 blanking apertures BA are arranged in parallel in eight columns in the beam traveling direction.

In the case of ghost exposure, as shown in FIG. 4B, among the rows L1 to L8 of the blanking apertures BA arranged in eight rows, only three rows L1, L2 and L3 are used and the remaining five rows are used. The columns L4 to L8 are always off. The exposure data is, for example, a blanking aperture BA11,
Triple exposure is performed by BA21 and BA31.

In the actual exposure, the pattern data is exposed eight times by the blanking apertures, for example, BA11 to BA81. In the ghost exposure, the pattern data is exposed three times. The exposure intensity of ghost exposure becomes 3/8 of the actual exposure.

The proximity effect can be compensated by exposing the inversion pattern with a predetermined weak intensity while slightly blurring the beam. In addition, the actual exposure is performed by double exposure,
Although the case where the ghost exposure is performed by the triple exposure has been described, the ratio of the ghost exposure to the actual exposure can be changed according to the conditions.

The method of reducing the intensity of ghost exposure with respect to the actual exposure is possible in addition to the method shown in FIG. For example, the effective exposure intensity can be reduced by activating all the blanking apertures BA even in the ghost exposure and making the scanning speed faster than the actual exposure.

For example, if the clock signal is set to 50 MHz in the actual exposure and the clock signal is set to 100 MHz in the ghost exposure, the exposure intensity of the ghost exposure becomes about ½ of the actual exposure.

Further, when creating the inversion data, the intensity of the ghost exposure can be reduced also by creating the data for turning off the ghost exposure for a predetermined position.

For example, in FIG. 4B, if all the blanking apertures in the even columns are turned off,
The exposure intensity of ghost exposure becomes 1/2. in this way,
The method of thinning out the ghost exposure pattern data is not limited to every other row, and can be arbitrarily selected such as one in three rows and two in five rows.

As described above, by inverting the pattern data used in the actual exposure, the inverted pattern data can be easily obtained. By using this reverse pattern data as it is and performing reverse blur exposure after actual exposure, the proximity effect can be efficiently corrected.

The present invention has been described above with reference to the embodiments.
The present invention is not limited to these. For example,
It will be apparent to those skilled in the art that various modifications, improvements, combinations and the like can be made.

[0113]

As described above, according to the present invention,
Proximity effect correction can be efficiently performed even in a system in which exposure is performed by independently controlling ON / OFF of a plurality of charged particle beams according to exposure information.

By subdividing the pattern and using a plurality of beams, an arbitrary pattern can be exposed at high speed, and by performing proximity effect correction, highly accurate exposure can be performed.

[Brief description of drawings]

FIG. 1 is a block diagram, a schematic plan view and a graph showing a basic embodiment of the present invention.

FIG. 2 is a block diagram of an electron beam exposure apparatus according to an embodiment of the present invention.

3A and 3B are a schematic plan view and a circuit diagram showing a configuration of a scroller used in the configuration of FIG.

FIG. 4 is a schematic plan view illustrating an exposure method according to an exemplary embodiment of the present invention.

FIG. 5 is a block diagram showing a typical configuration of an electron beam exposure apparatus.

FIG. 6 is a conceptual diagram for explaining a proximity effect.

FIG. 7 is a conceptual diagram for explaining correction of a proximity effect.

FIG. 8 is a schematic plan view and a graph for explaining a blanking aperture array.

FIG. 9 is a schematic plan view for explaining exposure by a blanking aperture array.

[Explanation of symbols]

 1 shift register 2 inversion circuit 3 exposure target area 4 unit strip area 5 actual exposure 6 inversion blur exposure 8 blanking aperture array 10 exposure system D delay circuit S scroller SM SRAM PD pattern data expansion device BA blanking aperture

─────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] October 28, 1992

[Procedure Amendment 1]

[Document name to be corrected] Drawing

[Name of item to be corrected] Figure 5

[Correction method] Change

[Correction content]

[Figure 5]

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁵ Identification number Internal reference number FI Technical display location G03F 7/20 521 7818-2H 7352-4M H01L 21/30 321 8831-4M 341 B

Claims (7)

[Claims] 1. An actual exposure step (5) of performing on / off control of a plurality of charged particle beams independently according to exposure information to irradiate a target surface to expose a desired pattern, and the actual exposure step (5) A charged particle beam exposure method, which comprises a reversal exposure step (6) of reversing the exposure information as it is, accumulating it in a shift register, and performing exposure with a weaker intensity than the actual exposure. 2. Each of the plurality of openings has an on / off controllable opening, and when the plurality of openings are moved in a predetermined direction, an arbitrary point on the target surface can be irradiated with a charged particle beam a plurality of times. In the actual exposure process, a blanking aperture array with openings is used to delay the exposure information to obtain signals for exposing the same pattern on the object at different times, and also to reverse the exposure information. The charged particle beam exposure method according to claim 1, wherein information is formed, and in the reversal exposure step, signals for exposing the same pattern on the object at different times are obtained by delaying the reversed information. 3. The charged particle beam exposure method according to claim 2, wherein the number of times the same pattern is exposed in the reverse exposure is smaller than the number of times the same pattern is exposed in the actual exposure. 4. The charged particle beam exposure method according to claim 2, wherein the speed of delaying the reverse information in the reverse exposure is higher than the speed of delaying the exposure information in the actual exposure. 5. The charged particle beam exposure method according to claim 1, wherein the reversal exposure is performed in a state where the focus is deviated from that in the actual exposure. 6. A plurality of openings each of which can be controlled to be turned on and off, and when moved in a predetermined direction, the plurality of openings can be irradiated with a charged particle beam at an arbitrary point on a target surface a plurality of times. A blanking aperture array (8) in which openings are arranged; a shift register (1) for sequentially storing and sequentially supplying exposure information to be sequentially supplied to a pair of openings of the blanking aperture; Inversion circuit that inverts the output (2)
And a feedback path for returning the output of the inverting circuit to the input of the shift register. 7. A delay circuit which receives the output of the shift register and supplies the output of each stage while giving a delay of a plurality of stages, and a control circuit which switches the number of stages of the delay circuit according to an exposure condition. Charged particle beam exposure apparatus as described.