GB2349737A - Electron beam exposure apparatus - Google Patents

Abstract

An electron beam exposure apparatus, that does not require a lead time and that can achieve a high effective throughput, comprises a beam source 1 for generating an electron beam, shaping means 3-6,9 for shaping the electron beam, deflecting means 12,13,36 for changing the position of the electron beam irradiated on a sample, and projection means 10-1 to 10-9 for focusing the shaped electron beam onto the sample, wherein the shaping means 3-6,9 comprises a shaping aperture array 3 for generating a plurality of sub-beams by splitting the electron beam, rectangular shaping means e.g. arrays 4-6 for respectively shaping each of the plurality of sub-beams into a desired rectangular shape, and sub-beam deflection means 9 for respectively moving the irradiation position of each of the plurality of sub-beams. The electron beam exposure apparatus thus splits the electron beam generated by a single electron beam source into a plurality of sub-beams, applies shaping to each sub-beam using a variable size rectangle method, and deflects each shaped beam. If the axial distance between the sub-beams is small, the effects of the displacement between the sub-beams can be made small and the problem of displacements at joins does not occur. Moreover, all the sub-beams can be deflected over a wide range using prior known deflection means. As a result, the beam can be split into a large number of sub-beams, and the throughput improved greatly.

Description

ELECTRON BEAM EXPOSURE APPARATUS BACKGROUND OF THE INVENTION The present invention relates to an electron beam exposure apparatus and, more particularly, to a technique for improving throughput in an electron beam exposure apparatus.

In recent years, semiconductor technology has been advancing at a rapid pace, achieving higher integration levels and higher functionality of semiconductor integrated circuits (ICs), and is now expected to play a major role as a core technology in the advancement of technologies spanning a wide variety of industrial fields including computers and communication equipment control.

The integration level of ICs has been increasing by a factor of 4 every two to three years and, in the case of dynamic random access memory (DRAM), its storage capacity has quadrupled from 1M to 4M, to 16M, to 256M, and then to 1G. These high integration levels of ICs have been made possible in large part by advances in miniaturization techniques in semiconductor technology.

At present, the limits of the miniaturization techniques are determined by pattern exposure techniques (lithography techniques). The pattern exposure technique commonly practiced today uses an optical exposure (optical lithography) apparatus called a stepper. In the optical lithography apparatus, the minimum linewidth of a pattern that can be formed is limited by the wavelength of the exposure light source used because of diffraction phenomena. Currently, light sources that emit ultraviolet rays are used, but it is difficult to use light of shorter wavelengths, and various new exposure methods other than optical lithography are being studied to achieve finer-featured processing. Among them, electron beam exposure is capable of processing much smaller features and patterns than optical lithography is. The development of electron beam exposure techniques has been progressing faster than other methods in that practical machines are already available, and much attention has been directed to electron beam exposure as a technique that can replace optical lithography.

However, it has been thought that electron beam exposure techniques cannot be used for mass production of LSIs because of low throughput compared with steppers. This thinking has been based, for example, on examples of electron beam exposure of a single-stroke writing type in which a single electron beam is continuously scanned for exposure, and has not been the conclusion arrived at after seriously analyzing and studying the causes by focusing attention on physical and technical obstacles to the improvement of throughput. In other words, the judgement that electron beam exposure cannot be used for mass production of LSIs because of low throughput has only been made in view of the productivity of the electron beam exposure that uses the prior known single beam exposure method.

Various methods for improving the throughput of electron beam exposure have been proposed in recent years. In electron beam exposure using a single beam, a pattern is written by repeatedly scanning the beam across the pattern portion to fill the pattern image. To accurately write fine corners of the pattern, the beam must be focused into a smaller spot, which correspondingly increases the time required for image filling. In view of this, there has been proposed a blanking aperture array (BAA) method in which a plurality of electron beams, each capable of being turned on and off independently of the others by means of a device called a blanking aperture array (BAA) having a plurality of arrayed apertures, are generated and scanned simultaneously across the pattern. The BAA method, like the single beam method, does not require the use of a mask that is needed in optical lithography. In practice, the plurality of electron beams are arranged twodimensionally, increasing the amount of exposure while reducing the rate of change of the overall current amount at pattern edges, etc. The efficiency of image filling according to the BAA method is greatly improved compared with the single beam method in the case of patterns whose width as measured perpendicularly to the scan direction is large, but does not improve much when, for example, there is a fine pattern extending in a direction parallel to the scan direction. In any case, the BAA method needs to scan the entire exposure range and, when the pattern to be exposed is small, the exposure time increases, contrary to its intended purpose, and sufficient throughput cannot be obtained at the present state of technology. Furthermore, the BAA has a large number of apertures which need to be made smaller than 70 to 50% of the pattern rule, and all of which are required to operate properly. Therefore, the BAA must be managed strictly, this increasing the overhead time and leading to the problem of decreased throughput.

Other methods proposed for the improvement of throughput include a variable size rectangle method. In the variable size rectangle method, two substrates each having a rectangular aperture are arranged with the apertures facing opposite each other, and a beam rectangularly shaped by being passed through the aperture in the first substrate is deflected so as to be irradiated on the aperture in the second substrate, then the beam passed through it is deflected so as to be brought back into its original orientation. The shape of the beam passed through the aperture in the second substrate is determined by the amount of deflection, that is, the degree of overlapping between that aperture and the beam irradiated on the second substrate; thus the beam can be shaped into any desired rectangular shape by controlling the amount of deflection. The exposure pattern is decomposed into rectangles, and the beam. after being rectangularly shaped is deflected toward the irradiation position for exposure. Accordingly, the variable size rectangle method also does not require the use of a mask that is needed in optical lithography.

With the variable size rectangle method, a large rectangular pattern can be exposed in a single shot; accordingly, the exposure efficiently greatly improves when exposing a pattern that can be decomposed into large rectangular shapes, but when exposing small discrete rectangular shapes, sufficient throughput cannot be obtained.

While the single beam method, the BAA method, and the variable size rectangle method described above do not need the use of a mask that is needed in optical lithography, there is proposed a different electron beam exposure method called the block exposure method which uses a mask. In semiconductor devices, and particularly in memories or the like, an area where the same pattern is repeated occupies a substantial portion of the slice.

If a block mask is prepared that has aperture patterns corresponding to such repetitive patterns, the repetitive patterns can be exposed in a single shot. In actual semiconductor devices, there are various kinds of repetitive patterns; therefore, if various aperture patterns corresponding to the various repetitive patterns are prepared and made available for selection, most semiconductor device patterns can be exposed using an available block mask pattern. For a pattern for which the corresponding block mask is not available, exposure is performed using the block mask in combination with the variable size rectangle method or the like. With the block exposure method, since any complicated patterns can be exposed in a single shot as long as the corresponding block mask is available, throughput improves greatly.

However, for semiconductor devices (microprocessors, etc.) that have random patterns for logic or other purposes, the area where the block exposure method can be applied is limited and the throughput cannot be increased satisfactorily. Further, since the block exposure method, like optical lithography, uses a mask, the mask must be produced separated, which increases the lead time before exposure can actually be performed. Moreover, the mask must be managed strictly, because dust on the mask would cause a defect in the exposure pattern.

Accordingly, the time required for the mask management increases the overhead time of the apparatus as in the case of optical lithography, resulting in the problem that the actual throughput does not improve as much as expected. There is also the problem that the cost required for the mask production and management adds to the cost of the product.

Various methods proposed in the prior art for the improvement of the throughput of electron beam exposure have been described above. To improve the throughput, the area that can be exposed per shot must be increased, while reducing the time required for one shot. Reducing the time per shot can be achieved either by reducing the setup time required to set up the beam ready for exposure, such as the exposure pattern shaping and deflection, or by shortening the exposure time per shot by increasing the beam current density per unit area.

The setup time differs from method to method and must be addressed according to the method employed. When the beam current density is increased, the beam is blurred by Coulomb interactions and the resolution drops. The influence of Coulomb interactions is also related to the size of the beam; if the beam size is increased while keeping the beam current density unchanged, there occurs the problem that the resolution drops because of Coulomb interactions.

To solve such problems, T. R. Groves and R. A.

Kendall propose, in J. Vac. Sci. Technol. B16 (6), Nov/Dec 1998, pp. 3168-3173, an electron beam exposure apparatus equipped with a plurality of variable rectangular type beam projection systems (columns). In this apparatus, each column includes an independent electron beam source, a variable rectangular shaping means, and an electrostatic deflection means having a small deflection range. The inventor of the present invention also proposed an electron beam exposure apparatus having a plurality of columns in Japanese Unexamined Patent Publication No. 10-128795 and other literature. In such apparatuses having a plurality of independent columns, the problem of Coulomb interactions described above is alleviated. However, axial distance between the columns cannot be shortened beyond a certain limit, limiting the number of columns that can be provided in one apparatus, and the throughput, therefore, cannot be increased sufficiently. Furthermore, since the axial distance between the columns is large, the variation of the column axial distance due to temperature changes or variations in temperature distribution is relatively large compared with the minimum linewidth of the pattern to be exposed, which can lead to the problem of displacements at pattern joins. At present, electron beam exposure can provide sufficiently good resolution compared with optical lithography notwithstanding the influence of Coulomb interactions, and in a practical apparatus, the improvement of throughput is an issue of greater concern.

As described above, various methods for improving the throughput in electron beam exposure have been proposed, but each method has its own problems. At the present state of technology, the block mask exposure method offers the highest throughput, but the problems, as previously noted, are that it requires a lead time for the preparation of a mask, and that its management is difficult and the overhead time increases, as a result of which the effective throughput does not improve as much as one would expect. In the case of the BAA method and the variable size rectangle method, the lead time for the mask preparation is not needed, but the throughput is low compared with the block exposure method. The BAA method has the further problem that the management of the BAA' increases the overhead time. On the other hand, with the multiple column method it is difficult to improve the throughput sufficiently by itself, and there is the problem that the pattern resolution drops at pattern joins.

SUMMARY OF THE INVENTION It is an object of the present invention to provide an electron beam exposure apparatus that does not require lead time and that can achieve high effective throughput.

To achieve the above object, the electron beam exposure apparatus of the present invention includes shaping means for splitting an electron beam irradiated from a single electron beam source into a plurality of sub-beams and then respectively shaping each sub-beam into a rectangular shape of variable size, the shaping means further including means for respectively deflecting each sub-beam, though over a small range, and the invention uses prior known focusing means and deflection means for focusing the plurality of sub-beams as a whole and for deflecting the beams over a wider range.

More specifically, the electron beam exposure apparatus of the invention comprises a beam source for generating an electron beam; shaping means for shaping the electron beam; deflecting means for changing the irradiation position of the electron beam on a sample; and projection means for focusing the shaped electron beam onto the sample, wherein the shaping means comprises: splitting means for generating a plurality of sub-beams by splitting the electron beam; rectangular shaping means for respectively shaping each of the plurality of sub-beams into a desired rectangular shape; and sub-beam deflection means for respectively moving the irradiation position of each of the plurality of subbeams.

The electron beam exposure apparatus of the invention splits the electron beam generated by a single electron beam source into a plurality of sub-beams, applies shaping to each sub-beam using a variable size rectangle method, and deflects each shaped beam. If the axial distance between the sub-beams is small, the effects of the displacement between the sub-beams can be held small and the problem of displacements at joins does not occur. Moreover, all the sub-beams can be deflected over a wide range using prior known deflection means. As a result, the beam can be split into a large number of sub-beams, and the throughput improves greatly. It is also possible to provide a plurality of electron beam sources and provide for each electron beam source the above-described construction, i. e, the system for splitting the electron beam generated by the electron beam source into a plurality of sub-beams, then applying shaping to each sub-beam using the variable size rectangle method, and deflecting each shaped sub-beam.

Further, it is preferable that each of the plurality of sub-beams is shaped and deflected independently of each other. However, when the plurality of sub-beams respectively have different current distributions although they are generated by dividing a same electron beam, it may be necessary to use a redundancy that an irradiation of a same area on a sample is divided into a plurality of exposures of different sub-beams. In that case, a same beam control signal is supplied to sub-beams in respective groups which expose a same pattern by adding desired delay times. Further, in this case, it is necessary to reduce a quantity of dose of each sub-beam according to a number of the divisions.

Further, blanker means may be provided for performing control simultaneously for all sub-beams as to whether all the sub-beams together are irradiated or not irradiated onto the sample, or sub-beam blanker means may be provided for controlling each sub-beam independently of each other as to whether each sub-beam is irradiated or not irradiated onto the sample; alternatively, both means may be provided and used in combination. If the sub-beam blanker is provided, each sub-beam can be irradiated independently. When using both means in combination, the common blanker means is used, for example, when changing the deflection amount of a deflector having a large deflection range, such as a main deflector or a sub deflector forming the deflection means, and at other times, the sub-beam blanker means is used.

The splitting means for splitting the beam into closely spaced sub-beams is realized using a substrate having a plurality of first shaping apertures of a prescribed rectangular shape arranged at a prescribed pitch. Using this substrate, a plurality of sub-beams having a prescribed rectangular shape, and arranged at a prescribed pitch, are generated. The rectangular shaping means comprises first shaping deflection means for respectively deflecting each of the plurality of subbeams, a shaping aperture array having a plurality of second shaping apertures of a rectangular shape arranged to correspond with the prescribed pitch, and second shaping deflection means for deflecting back the plurality of sub-beams passed through the plurality of second shaping apertures; using the first shaping deflection means, the plurality of sub-beams are irradiated on the corresponding second shaping apertures in the shaping aperture array, and each sub-beam is shaped into a shape that is defined by the region of overlap between the sub-beam and the second shaping aperture on which the sub-beam is irradiated. The first and second shaping deflection means each comprise two shaping deflection substrates, each including a plurality of apertures arranged to correspond with the arrangement of the plurality of sub-beams, a pair of deflection electrodes formed on both sides of each aperture for forming an electrostatic field, and shield electrodes formed in other positions flanking each of the apertures than the positions where the pair of deflector electrodes are formed, wherein the direction of the electrostatic field formed by the deflection electrode pair on one shaping deflection substrate is oriented at 90 to the direction of the electrostatic field formed by the corresponding deflection electrode pair on the other shaping deflection substrate, and the two shaping deflection substrates are arranged in close proximity to each other.

The sub-beam deflection means comprises two deflection substrates, each including a plurality of apertures arranged to correspond with the arrangement of the plurality of sub-beams, a pair of deflection electrodes formed on both sides of each aperture for forming an electrostatic field, and shield electrodes formed in other positions flanking each of the apertures, than the positions where the pair of deflector electrodes are formed, wherein the direction of the electrostatic field formed by the deflection electrode pair on one deflection substrate is oriented at 90 to the direction of the electrostatic field formed by the corresponding deflection electrode pair on the other deflection substrate, and the two deflection substrates are arranged in close proximity to each other.

In the present invention, it is required that the plurality of closely spaced sub-beams be accurately deflected after they are shaped independently of each other, and it can be said that the fact that the rectangular shaping means and sub-beam deflection means for achieving this have been implemented in integrated form on a single substrate, as described above, is the factor that has made it possible to increase effective throughput in the present invention.

The sub-beam blanker means comprises: a blanker deflection substrate which includes a plurality of apertures arranged to correspond with the arrangement of the plurality of sub-beams, a pair of deflection electrodes formed on both sides of each aperture for forming an electrostatic field, and shield electrodes formed in other positions flanking each of the apertures than the positions where the pair of deflector electrodes are formed; and a shield plate for blanking the plurality of sub-beams deflected by the deflection electrode pairs.

Preferably, the substrate of the splitting means includes a plurality of shaping aperture sets each consisting of the plurality of first shaping apertures, and any one of the plurality of sets is selectively movable into a path of the electron beam. In the substrate of the splitting means, it is only necessary to form the plurality of first shaping apertures, and wiring or the like need not be provided; therefore, a plurality of shaping aperture sets can be provided. Since the substrate of the splitting means is damaged by the irradiation of electron beams, providing the plurality of shaping aperture sets and using them selectively serves to improve serviceability.

Generally, a deflector with a larger deflection range requires a longer setup time. In the prior known apparatus, therefore, a main defector, a sub-deflector, and, if necessary, a sub-sub-deflector, are combined so that the beam can in effect be deflected at high speed over a wide deflection range. In the apparatus of the present invention also, it is desirable that deflectors having different deflection ranges and requiring different setup times be used in combination. in the present invention, however, since each sub-beam can be deflected respectively, and since the main deflector, the sub-deflector, the sub-sub-defector, etc. deflect all the sub-beams together by the same amount, the method of deflection differs according to the range of sub-beam deflection.

In a first case, the sub-beam deflection ranges are contiguous to or overlap with each other. In this case, the deflection range of all the arrayed sub-beams is set as the deflection range of the lowest-order deflection means, and the deflection is performed in combination of a higher-order deflection means having a larger deflection range, as practiced in the prior known apparatus. After the pattern exposure in the deflection ranges of all the sub-beams is completed, the deflection position of the deflection means is moved to the next deflection position, and the same process is repeated.

In the case where the sub-beam deflection ranges are not contiguous to each other, the deflection position of the other deflector (lowest-order deflector) constituting the deflection means is moved by an amount equal to the width of the sub-beam deflection range, to expose the full range of the sub-beam array. The remainder of the process is the same as that in the prior known apparatus.

For example, when the centers of the sub-beam deflection ranges are spaced apart from each other by a distance equal to four times the sub-beam range width, the area between each adjacent sub-beam deflection range can all be exposed by displacing the center position four times.

If the centers of the sub-beam deflection ranges are spaced apart from each other by a distance equal to four times the sub-beam range width in both the X-and Y-axis directions, then the exposure should be performed by displacing the center position a total of 16 times.

Usually, the deflection means is constructed by combining a main deflector with a sub-deflector; in this case, the deflection for exposing the area between each adjacent sub-beam deflection range may be performed using the subdeflector, but preferably a sub-sub-deflector, whose deflection range is smaller than that of the subdeflector but whose deflection setup time is shorter, should be provided to perform the deflection.

Preferably, each sub-beam deflection range is set smaller than its maximum deflection range, and a pattern running over a boundary between split deflection ranges is exposed in one shot. This serves to prevent displacements at joins.

Further, when the sub-beam deflection ranges are arranged overlapping each other, provisions are made so that when the exposure of a pattern in a certain sub-beam deflection range is completed, if the exposure in its adjacent sub-beam deflection range is not completed yet, the sub-beam with which the exposure has been completed is used to expose the pattern in its adjacent sub-beam deflection range. This improves the throughput.

BRIEF DESCRIPTION OF THE DRAWINGS Other features, objects and advantages of the present invention will become apparent from the following description of preferred embodiments with reference to the drawings, in which: Figure 1 is a diagram showing the entire construction of an electron beam exposure apparatus according to an embodiment of the present invention; Figure 2 is a diagram showing electron beam paths in an electron optical system in the electron beam exposure apparatus of the embodiment; Figures 3A and 3B are diagrams showing the path of one sub-beam in the electron optical system in the electron beam exposure apparatus of the embodiment; Figures 4A and 4B are diagrams showing a structural example of first and second shaping aperture arrays in the electron beam exposure apparatus of the embodiment; Figure 5 is a diagram showing a deflector array substrate of the electron beam exposure apparatus of the embodiment; Figure 6 is a top plan view of a deflector array substrate; Figure 7 is a diagram showing an aperture, electrode shape, and an electric field formed by electrodes in one aperture unit in the deflector array substrate; Figure 8 shows a side view and enlarged cross sectional views of a first shaping deflector array 4, a second shaping deflector array 5, and a sub-beam deflector array 9 ; Figures 9A to 9D are diagrams for explaining the splitting of a deflection range according to the embodiment ; Figures 10A and 10B are diagrams for explaining the splitting of a deflection range according to the embodiment; and Figure 11 is a diagram showing an aperture, electrode shape, and an electric field formed by electrodes in one aperture unit in the deflector array substrate in an modified example.

DESCRIPTION OF THE PREFERRED EMBODIMENTS Figure 1 is a diagram showing schematically the construction of an electron beam exposure apparatus according to an embodiment of the present invention, Figure 2 is a diagram for explaining electron beam paths in an electron optical system in Figure 2, and Figures 3A and 3B are diagrams for explaining the path of one subbeam in the center.

In Figure 1, reference numeral 1 is an electron gun, 3 is a first shaping aperture array, 4 is a first shaping deflector array, 5 is a second shaping deflector array, 6 is a second shaping aperture array, 7 is a sub-beam blanker, 8 is an iris, 9 is a sub-beam deflector array, 10-1 to 10-9 are magnetic lenses, 11 is a common blanker, 12 is a main deflector, 13 is a sub deflector, 14 is an aberration/Coulomb blurring corrector, 15 is a wafer, 16 is a stage mechanism for moving the wafer 15 by clamping it thereon by vacuum, 17 is a controller for the stage mechanism, 18 is an reflected electron detector used to detect the focusing condition and the position of a reference mark, 19 is a detection signal processing circuit, 20 is a main deflector controller, 21 is a subdeflector controller, 22 is an aberration/Coulomb blurring corrector controller, 23 is a sub-beam deflector array controller, 24 is a common-blanker controller, 25 is a sub-beam blanker controller, 26 is a second shaping deflector array controller, 27 is a first shaping deflector array controller, 28 is an electron optical system controller, 29 is a control computer, 30 is a large scale storage device, 31 is an interface for interfacing the control computer to the various parts of the apparatus, 32 is a network adapter for connection to a host computer, 33 is a computer bus, 34 is a control bus, 35 is a Faraday cup, and 36 is a sub-sub-deflector.

In the figure, solid lines and dashed lines, leading from the electron gun 1 to the wafer 15, indicate the outermost paths and the optical axes, respectively, of the electron beams emitted from both edges of the electron gun.

The basic construction of the electron beam exposure apparatus of the present embodiment is the same as that of the prior known apparatus, and details not shown here are the same as those in the prior known apparatus. For example, the electron beam paths, the wafer 15, the stage mechanism 16, the deflectors, and the corrector are all contained within a cylindrical vacuum chamber. The following description deals only with those parts which characterize the present invention.

The electron beam paths in the electron optical system of the embodiment will be described with reference to Figure 2 and Figures 3A and 3B. Reference numerals 2-1 to 2-8 indicate the axes of the magnetic fields produced by the respective magnetic lenses 10-1 to 10-8.

The electron beam emitted from the electron gun 1 is first converged by the magnetic field 2-1 and then irradiated onto the first shaping aperture array 3. The first aperture array 3 comprises many rectangular apertures arranged in an array, as will be described later, and the electron beam passing therethrough is split into many sub-beams. The electron beam passing through an aperture 42 in the center also emerges as a sub-beam, as shown in Figure 3. The group of sub-beams is converged by the magnetic field 2-2 and, halfway through the converging process, enters the first shaping deflector array 4. The first shaping deflector array 4 is mounted at a position where the magnification ratio of the electron beam becomes the same as that with the first shaping aperture array 3. The same is true of the second shaping deflector array 5, the second shaping aperture array 6, the sub-beam blanker 7, and the sub-beam deflector array 9. The first shaping deflector array 4 also contains apertures, as in the first shaping aperture array 3, and deflection electrodes are formed on both sides of each aperture so that the beam can be deflected by a desired amount in the X-and Y-axis directions perpendicular to the optical axis (Z axis). Figure 3A shows the case where the sub-beam is not deflected by its corresponding first shaping deflector 4-1, and Figure 3B shows the case where the sub-beam is deflected by its corresponding first shaping deflector 4-1. After being converged, the group of sub-beams passes through the magnetic field 2-3 and enters the second shaping deflector array 5. In the second shaping deflector array 5, the deflector corresponding to each sub-beam reverses the deflection performed by the corresponding deflector in the first shaping deflector array 4, thus reversing the path back to the original orientation. The second shaping deflector array 5 is immediately followed by the second shaping aperture array 6, in which the degree of overlapping between each sub-beam and its corresponding aperture varies according to the amount of deflection applied by the corresponding deflector in the first shaping deflector array 4. As shown in Figure 3A, when the sub-beam is not deflected by the first shaping deflector 4-1, one half of the sub-beam passes through the aperture. On the other hand, when the sub-beam is deflected by the first shaping deflector 4-1 as shown in Figure 3B, most of the sub-beam passes through the aperture. Reference numeral 5-2 shows the deflection being applied by the second shaping deflector array 5-1; as can be seen, the direction of the deflection is reversed from that of the deflection applied by the first shaping deflector 4-1. If the direction of deflection is reversed from that shown in Figure 3B, the beam passing through the aperture becomes narrower. Such deflection is applied in the X-and Y-axis directions to shape the beam into various rectangular shapes.

The group of sub-beams are then converged by the magnetic field 2-4 and fall on the sub-beam blanker 7. when a sub-beam is deflected by its corresponding deflector in the sub-beam blanker 7, as shown in Figure 3A, the sub-beam is blocked by the iris 8. On the other hand, when the sub-beam is not deflected by the deflector in the sub-beam blanker 7, the sub-beam passes through the iris 8, as shown in Figure 3B. In this way, on/off control can be performed on each individual sub-beam to irradiate or not irradiate the sub-beam onto the wafer 15. After passing through the magnetic field 2-5, the group of sub-beams enters the sub-beam deflector array 9.

As shown in Figures 3A and 3B, the irradiation position on the wafer 15 changes depending on whether deflection is applied or not applied by the corresponding deflector in the sub-beam deflector array 9. The group of subbeams is then focused onto the wafer 15 through the magnetic fields 2-6,2-7, and 2-8.

Figures 4A and 4B show a structural example of the first and second shaping aperture arrays 3 and 6. As shown in Figure 4A, the shaping aperture array is formed from a thin plate of a silicon wafer or the like. An aperture area 41 where an array of apertures are to be formed is reduced in thickness by etching or like technique, and square apertures 42 are opened in the aperture area 41 by etching. In the present embodiment, for example, 20 x 20, i. e., a total of 400 apertures, each 15 Am square, are formed at a pitch of 60 Wm. Since a 60: 1 demagnified image of the shaping aperture array is projected at the plane of the wafer 15, if the image is projected undeformed, the sub-beams, each focused into a 0.25-pm square shape, will be arranged at a pitch of 1.0 Fm on the wafer 15.

As shown, 5 x 5, i. e., a total of 25 aperture areas 41 are provided. Since the apertures 42 are simply opened and no wiring is needed, the plurality of aperture areas 41 can be provided on the shaping aperture array substrate. The shaping aperture array substrate, especially the first shaping aperture array 3 which receives electron beams over its entire surface, requires replacement or other maintenance work according to the condition of use, because the apertures may be deformed due to heating etc. after long periods of use. By providing the plurality of aperture areas 41, and by constructing the system so that each one of the aperture areas 41 can be selectively positioned in the electron beam paths by the moving mechanism not shown, the replacement cycle can be prolonged, facilitating maintenance.

Figure 4B is a diagram showing a cross-sectional structure of the shaping aperture array; in this example, one aperture array area 41 is shown. Reference numeral 43 is a silicon (Si) substrate, 44 is a boron-diffused insulating layer, 46 is silicon, and 47 is a protective metal film. The apertures 42 are formed after further reducing the thickness of the aperture area 41 by etching.

Figure 5 is a deflector array substrate 50 used to construct the first shaping deflector array 4, the second shaping deflector array 5, the sub-beam blanker 7, and the sub-beam deflector array 9. This substrate 50 is also formed from a thin plate of a silicon wafer or the like, and a deflector array area 51 where a deflector array is formed is reduced in thickness, as in the case of the aperture area 41 of the shaping aperture array.

Reference numeral 52 shows signal electrode pads for supplying signals to be applied to the deflector electrodes formed in the deflector array area 51, and 53 indicates a GND electrode pad.

Figure 6 is a top plan view of the deflector array area 51. As shown, square apertures 56 are arranged corresponding to the arrangement of the apertures 42 in the shaping aperture array. Along two opposing sides of each aperture 56 are arranged a positive electrode 53 and a negative electrode 54, while a shield ground electrode 55 is formed on each of the other two sides.

Accordingly, in the present embodiment, 20 x 20, i. e., a total of 400 square apertures 56 are formed at a pitch of 60 pm. Each aperture 56 is about 25 am square, which is slightly larger than each aperture 42 in the shaping aperture array.

Figure 7 is a diagram showing the aperture, the electrode shape, and an electric field formed by the electrodes in one aperture unit 57. As shown, the positive electrode 53 and the negative electrode 54 are symmetrical to each other, the center portions forming parallel electrodes with both ends slightly curved. The shield ground electrodes 55 are shared with the respective adjacent aperture units. It is shown that a uniform electric field, with electric lines of force in the center portion being parallel and equidistant, is formed by the electrodes with the above-described shape.

If the amount of deflection provided by the deflector array substrate 50 is not accurate, a beam of a desired shape may not be obtained or the exposure position may become displaced; therefore, extremely high accuracy is demanded of the deflector array substrate 50 in terms of the amount of deflection it provides. In view of this, each aperture 56 is formed in a square shape to ensure the formation of a uniform electric field. Each aperture 42 in the shaping aperture array has the shape shown at reference numeral 58; the sub-beam shaped through the aperture 42, therefore, quite easily passes through the aperture 56 and is deflected accurately by the uniform electric field.

Turning back to Figure 5, many wiring lines for connecting the signal electrode pads 52 and GND electrode pads 53 to the positive electrodes 53 and negative electrodes 54 in the deflector array area 51 are formed in multiple layers in areas between the deflector array area 51 in the center and the surrounding signal electrode pads 52 and GND electrode pads 53.

The sub-beam blanker 7 can be constructed using only one deflector array substrate 50 shown in Figures 5 to 7, since its function is to deflect the beam in only one direction ; on the other hand, for the first shaping deflector array 4, the second shaping deflector array 5, and the sub-beam deflector array 9, two deflector array substrates 50 are used because the beam must be deflected in two directions perpendicular to the axis of the electron optical system.

Figure 8 is a diagram showing the structure of the first shaping deflector array 4, the second shaping deflector array 5, and the sub-beam deflector array 9.

As shown in Figure 8, two deflector array substrates 50 are arranged in close proximity to each other. In the deflector array area 51 of one substrate 50, the apertures 56 are formed in the substrate, and the positive electrodes 53, the negative electrodes 54, and the shield ground electrodes 55 are formed on one side of the substrate. The signal electrodes pads 52 and the GND electrode pads are likewise formed on the one side of the substrate. In the deflector array area 51'of the other substrate 50', the apertures 56'are formed in the substrate, and the positive electrodes 53', the negative electrodes 54', and the shield ground electrodes 55.'are formed on the other side of the substrate. The signal electrodes pads 52'and the GND electrode pads are likewise formed on the one side of the substrate. The two substrates 50 and 50'are arranged with their nonelectrode sides facing each other and with the apertures 56 aligned with the corresponding apertures 56', as shown in the diagram. Since electrodes, etc. are not formed on the facing sides, the two substrates can be placed very close to each other.

The positive electrodes 53 and negative electrodes 54 are oriented at 90 to the positive electrodes 53'and negative electrodes 54'. Therefore, by applying voltage between the positive and negative electrodes 53 and 54 and between the positive and negative electrodes 53'and 54', electric fields are formed in directions shown by reference numerals 61 and 61', so that the sub-beam passing through the corresponding apertures 56 and 56' can be deflected in directions oriented at 90 to each other. That is, a deflector is realized that can deflect the sub-beam independently in two directions, the X-and Y-axis directions, perpendicular to the optical axis.

In the present embodiment, the number of apertures 56 and 56'and their pitch are the same in both the X and Y directions. Therefore, each deflector is constructed by fabricating two deflector array substrates 50 using the same fabrication process, and by arranging them back to back with their axes oriented at 90 to each other.

As earlier described, in the present embodiment, the first shaping deflector array 4, the second shaping deflector array 5, the sub-beam deflector array 9, and the sub-beam blanker 7 are each arranged at a position where the magnification ratio of the electron beam becomes the same as that of the first shaping aperture array 3. This makes it possible to use the same arrangement of apertures for each of the substrates used to construct the deflectors. Therefore, the first shaping deflector array 4, the second shaping deflector array 5, and the sub-beam deflector array 9 are each constructed using two deflector array substrates 50 fabricated using the same fabrication process, and the sub-beam blanker 7 is also constructed using the same substrate. This serves to reduce the effects of errors introduced in the fabrication process.

Turning back to Figure 1, the first shaping deflector array controller 27, the second shaping deflector array controller 26, the sub-beam deflector array controller 23, and the sub-beam blanker controller 25 generate driving signals to be applied to the signal electrodes on the first shaping defector array 4, the second shaping defector array 5, the sub-beam blanker 7, and the sub-beam deflector array 9, respectively.

In the apparatus of the present embodiment, the main deflector 12 as an electromagnetic reflector, the sub deflector 13 as an electrostatic deflector, and the sub sub-deflector 36 as an electromagnetic deflector together constitute a common deflection means. The size of the deflection range decreases in the order of the main deflector 12, the sub deflector 13, and the sub subdeflector 36, while the deflection speed (representing the length of the deflection setup time) decreases in the order of the sub sub-deflector 36, the sub deflector 13, and the main deflector 12. In the illustrated construction, the sub sub-deflector 36 is disposed outside of the sub deflector 13, but it can be disposed above of the sub deflector 13, in which case the sub subdeflector 36 can be constructed as an electrostatic deflector.

The above description has dealt with the construction of the electron beam exposure apparatus according to the present embodiment; other parts not specifically described above are fundamentally the same as those in the prior known apparatus.

Next, referring to Figures 9A to 9D and Figures 10A and 10B, a description will be given of how the deflection range is split according to the present embodiment.

As previously described, in the prior known electron beam exposure apparatus, deflectors of different characteristics are combined so that the beam can in effect be deflected at high speed over a wide deflection range. In practice, the throughput is further increased by continuously performing exposure while correcting the amount of stage movement using a sub defector, etc. while the stage is being moved. The present embodiment employs fundamentally the same method. In the present embodiment, however, since each sub-beam can be deflected independently, and since the main deflector, the sub deflector, the sub sub-defector, etc. deflect all the sub-beams together by the same amount, the method of deflection differs according to the range of sub-beam deflection. The method of deflection also differs when the deflection ranges of adjacent sub-beams either overlap or are contiguous with each other than when they are spaced apart from each other. The present embodiment will be described by assuming that the deflection ranges 79 of adjacent sub-beams are spaced apart from each other, as shown in Figure 10.

Figure 9A shows an arrangement of chips (dies) 70 formed on the wafer 15. Since each chip 70 is larger than the deflection range of the electron beam exposure apparatus, the stage must be moved to expose the entire chip 70. Two methods can be used: a method called stepand-repeat in which the stage is moved and then stopped to expose a pattern within the deflection range and, upon completion of the exposure, the stage is moved again and then stopped to expose the adjacent area, and a continuous movement method in which a portion of a pattern is exposed as it is moved into the deflection range while correcting the amount of stage movement using a sub deflector, etc while the stage is being moved.

Either method can be employed in the present invention, but for convenience of explanation, the following description is given by taking the step-and-repeat method as an example.

As shown in Figure 9A, the chips in the same column are exposed in sequence by changing the amount of movement of the stage in only one direction (fixed in the X direction and moved in steps only in the Y direction) by the width of a first deflection range (appropriately set within the maximum deflection range) corresponding to the main deflection range of the electron beam exposure apparatus. The region of the above width exposed at this time is called a frame 71. When the exposure of one frame 71 is completed, the next frame is exposed by moving the stage in the opposite direction, as shown in Figure 9B. Reference numeral 72 indicates the direction of stage movement. In the illustrated example, the width of the first deflection range 73 of the electron beam apparatus is one-third of one side of each square chip; therefore, the entire area of one chip can be exposed in nine step-and-repeat operations, and within one frame, one chip is exposed in three steps.

As shown in Figure 9C, the first deflection range 73 is split into second deflection ranges 75 (in the illustrated example, 35 ranges) each corresponding to a sub deflection range. With the deflection position of the main deflector 12 fixed at the center of one second deflection range 75, the amount of deflection in the sub deflector 13, the sub sub-deflector 36, and the sub-beam deflector array 9 is changed to expose a pattern within the second deflection range 75. When the exposure of the pattern within that second deflection range 75 is completed, the deflection position of the main deflector 12 is moved and fixed at the center of the next second deflection range 75, and the same process is repeated.

When this process has been performed on all the second deflection ranges 75 within the first deflection range 73, the exposure of that first deflection range 73 is complete, and the exposure process is repeated for the next first deflection range 73 in Figure 9B. Reference numeral 74 is the locus showing the change of the main deflection position.

As shown in Figure 9D, each second deflection range 75 is split into third deflection ranges (in the illustrated example, 16 ranges). With the deflection position of the sub deflector 13 fixed at the center of one third deflection range 77, the amount of deflection in the sub sub-deflector 36 and the sub-beam deflector array 9 is changed to expose a pattern within the third deflection range 77. When the exposure of the pattern within that third deflection range 77 is completed, the deflection position of the sub deflector 13 is moved and fixed at the center of the next third deflection range 77, and the same process is repeated. When this process has been performed on all the third deflection ranges 77 within the second deflection range 75, the exposure of that second deflection range 75 is complete, and the exposure process is repeated for the next second deflection range 75 in Figure 9C. Reference numeral 76 is the locus showing the change of the sub deflection position.

Figures 10A and 10B are diagrams showing how the exposure progresses in each third deflection range 77.

Reference numeral 79 indicates each sub-beam deflection range of the sub-beam deflection array 9. There are > 20 = 400 sub-beams, as earlier stated, and the sub-beam deflection ranges are each 0.25 pm square on the wafer and are spaced 1.0 pm apart from each other. Each third deflection range 77 is split into 400 fourth deflection ranges 82 each of which is further split into 16 fifth deflection ranges each corresponding to one sub-beam deflection range. After changing the amount of deflection of the sub sub-beam deflector 36 so as to be positioned at the center of each fifth deflection range 83 as shown in Figure 10B, 400 fifth deflection ranges 83 are exposed using the sub-beam deflector array 9. When the exposure is completed, the amount of deflection of the sub sub-beam deflector 36 is changed so as to be positioned at the center of the next fifth deflection range 83, as indicated by the locus in Figure 10B, and the same process is repeated. When the 16 fifth deflection ranges have been exposed for each sub-beam, the exposure of all the fourth deflection ranges 82, that is, the exposure of the third deflection range 77, is complete.

In this example, the deflection range of the sub sub-deflector 36 should be made to cover at least 4 x 4 fifth deflection ranges 83, that is, 1/80 of the deflection range of the sub deflector 13 (which corresponds to one second deflection range 75).

In each sub-beam deflection range 79, each sub-beam is independently shaped into a rectangle 81 by the first shaping deflector array 4 and the second shaping deflector array 5, and is irradiated for exposure after being deflected by the sub-beam deflector array 9 as shown by reference numeral 80 in accordance with the exposure position. To expose the rectangular shape a plurality of times within one sub-beam deflection range 79, the same process is repeated an equivalent number of times. For example, in the example at the left, the rectangular shape is exposed once; at the center, the rectangular shape is exposed twice; and at the right, the rectangular shape is exposed four times. As shown, the lower left corner of the rectangle is set as the reference position; even after shaping, the lower left corner of the rectangle is held at the same position, and in this condition, the beam is deflected so that the lower left corner of the rectangle is moved to the desired position.

As described above, in the present embodiment, the stage is moved so that the center of the first deflection range 73 is aligned with the optical axis, then the deflection position of the main deflector 12 is set at the center of one second deflection range 75, the deflection position of the sub deflector 13 is set at the center of one third deflection range 77, and the deflection position of the sub sub-deflector 36 is set so that each sub-beam deflection range is aligned with one of the fifth deflection ranges 83 into which each of the fourth deflection ranges 82 constituting the third deflection range 77 is subdivided; in this condition, the rectangular sub-beams shaped by the first shaping deflector array 4 and the second shaping deflector array 5 are deflected by the sub-beam deflector array 9 for exposure. When the exposure of each sub-beam deflection range is completed, the deflection position of the sub sub-deflector 36 is moved to the next fifth deflection range 83, and the same process is repeated. This process is repeated 16 times to complete the exposure of the third deflection range 77. Next, the deflection position of the sub deflector 13 is moved to the center of the next third deflection range 77, and the same process is repeated. This process is repeated 16 times to complete the exposure of the second deflection range 75. Further, the deflection position of the main deflector 12 is moved to the center of the next second deflection range 75, and the same process is repeated. This process is repeated 36 times to complete the exposure of the first deflection range 73. Then, the stage is moved in the Y direction to perform the exposure of the next first deflection range 73 in the same manner, and the exposure process is repeated until the exposure of one frame is completed.

Next, the stage is moved in the X direction to perform the same processing for the next frame. In this way, all the patterns on the wafer 15 are exposed.

Next, the exposure process to be performed in the electron beam exposure apparatus of the present embodiment will be described.

First, each unit is adjusted. In this adjustment, each unit is set at the optimum condition and, at the same time, data concerning differences between sub-beams in the sub-beam related units are collected. The electron gun 1 and the magnetic lenses 10-1 to 10-9 are adjusted using the electron optical system controller 28.

Further, the main deflector 12, the sub-deflector 13, and the sub-sub-deflector 36 are adjusted, and data concerning their deflection amounts are collected. These adjustments are the same as those performed in the prior known apparatus. The first shaping aperture array 3 and the second shaping aperture array 6 are each provided with a plurality of aperture areas 41, as previously described, and one of the aperture areas 41 is selected.

The first shaping aperture array 3, the second shaping aperture array 6, the first shaping deflector array 4, the second shaping deflector array 5, the sub-beam blanker 7, the sub-beam deflector array 9, and the iris 8 are adjusted for alignment by using an alignment jig or the like. At this time, data concerning the characteristics of each individual deflector for sub-beam deflection are also collected and stored. Further, the aberration/Coulomb blurring corrector 14, etc. are adjusted and their data are collected. For the above adjustments and data collection, the reflected electron detector 18, the Faraday cup 36, etc. are used.

Based on the thus collected data, correction data are set up in the main deflector controller 20, the sub deflector controller 21, the aberration/Coulomb blurring corrector controller 22, the sub-beam deflector array controller 23, the common-blanker controller 24, the subbeam blanker controller 25, the second shaping deflector array controller 26, the first shaping deflector array controller 27, the electron optical system controller 28, etc.

The control computer 29 creates exposure information for each writing frame from"LSI chip writing data'and "wafer layout and exposure condition information"stored in the large scale storage device 30. At this time, exposure information is created for each split exposure range, as described with reference to Figures 9A to 9D and Figures 10A and 10B.

For exposure, with the common blanker 11 set to blank the entire beam and the sub-beam blanker also set in the blanking position, the wafer 15 is clamped onto the stage 16, and the stage 16 is moved, as previously described with reference to Figures 9A to 9D and Figures 10A and 10B, to set the deflection positions of the main deflector 12 and the sub deflector 13. In this condition, the blanking condition of the common blanker 11 is released. Then, the deflection position of the sub sub-deflector 36 is set, and the exposure is started.

Each sub-beam is independently controllable for its shape and deflection position and sequentially exposes patterns in the respective ranges, but as previously noted, beam blurring occurs due to Coulomb interactions.

This blurring is corrected using the aberration/Coulomb blurring corrector 14, but it will not be desirable if the amount of current becomes large for the sub-beams as a whole or if the current changes widely. Some sub-beam writing ranges contain a large number of patterns, and some contain only a few patterns or no pattern at all.

Basically, the number of shots is determined by the number of patterns contained in the range to be exposed; the number of shots may be large or may be small, depending on the range to be exposed. In view of this, in a range where the number of shots is small, the order of shots is adjusted so as to make the maximum current amount per shot as small as possible and to reduce the change of the current amount between shots.

For example, the range (3,1) at the left of Figure 10A contains a large pattern of one shot, and the range (5,1) at the center contains relatively small patterns of two shots, while the range (m, 1) at the left contains small patterns of four shots. In the illustrated example, the exposure is performed in a number of steps by exposing, for example, only one pattern in (m, 1) in the first shot, one pattern in (5,1) and one pattern in (m, 1) in the second shot, the remaining one pattern in (5,1) and one pattern in (m, 1) in the third shot, and the pattern in (3,1) and the remaining one pattern in (m, 1) in the fourth shot. Actually, such exposure is performed for the exposure ranges of the 400 sub-beams.

This serves to reduce the maximum current amount per shot while also reducing the change of the current amount between shots. with the exposure performed as described above, the exposure of patterns in 400 fifth deflection ranges 83 is completed. Thereafter, the same exposure process is performed repeatedly by moving the stage and changing the deflection positions of the sub sub-deflector 36, the sub deflector 13, and the main deflector 12, until the all patterns on the wafer 15 have been exposed.

The electron beam exposure apparatus has been described above according to the embodiment of the present invention, but various modifications are also possible.

For example, the positive electrodes 53 and negative electrodes 54 formed on the deflector array substrate 50 of the embodiment have been described as having the shape shown in Figure 7, but these may be formed as parallel electrodes like the ones shown in Figure 11. In this modification, however, since the range where a uniform electric field can be formed becomes smaller, the size of the aperture unit 57, and hence the pitch of the subbeams, must be increased if the same sub-beam size is to be used also here. In the example of Figure 7, the aperture pitch is 1/4 of the beam size and sub-beam utilization efficiency is 1/16, while in the example of Figure 11, the aperture pitch is 1/6 of the beam size and sub-beam utilization efficiency is 1/36. Though the subbeam utilization efficiency drops by about one half, it is sufficient for practical purposes.

In the embodiment, the sub sub-beam deflector 36 is provided in addition to the main deflector 12 and sub deflector 13, and the discretely arranged sub-beam deflection ranges 79 are moved, but the sub sub-beam deflector 36 may be omitted and the sub-beam deflector 13 may be configured to perform this deflection.

Further, in the embodiment, the sub-beam deflection ranges 79 have been described as being spaced apart from each other as shown in Figure 10A, but the sub-beam deflection ranges 79 may be made contiguous to each other by increasing the amount of deflection of each set of deflectors in the sub-beam deflector array 9. In that case, the position change performed by the sub sub-beam deflector such as shown in Figure 10B need not be performed, but the third deflection range 77 is exposed by changing only the deflection position of each subbeam.

Furthermore, each sub-beam deflection range 79 may be set smaller than its maximum deflection range, with provisions made to expose a pattern running over a boundary between split deflection ranges in one shot.

This reduces the possibility of displacements at joins.

When the sub-beam deflection ranges are arranged overlapping each other, provisions may be made so that when the exposure of a pattern in a certain sub-beam deflection range is completed, if the exposure in its adjacent sub-beam deflection range is not completed yet, the sub-beam with which the exposure has been completed is used to expose the pattern in its adjacent sub-beam deflection range. This improves the throughput.

Further, when performing the exposure in a number of steps using several sub-beam shots, an upper limit may be preset for the total current amount of the sub-beams as a whole, and when the current value exceeds this upper limit even if the exposure process is divided into a number of steps, the number of shots may be set larger than the largest number of patterns in the sub-beam deflection ranges. In this case, the exposure time increases because of increased number of shots, but since such a situation does not occur often, degradation of the effective throughout is negligible.

As described above, according to the present invention, since very many exposures being performed by the variable size rectangle method proceed concurrently, the throughput improves greatly, and throughput comparable to or higher than that achieved by the block exposure method can be obtained. Furthermore, since the lead time for block mask preparation, as required in the block exposure method, is not needed nor is the block mask management, the overhead time is reduced and the effective throughput is further increased.

This achieves an electron beam exposure apparatus that can be used in mass production processes of LSIs, enabling highly integrated LSIs to be mass produced at low cost.

Claims (12)

1. An electron beam exposure apparatus comprising: a beam source for generating an electron beam; shaping means for shaping said electron beam; deflecting means for changing the irradiation position of said electron beam on a sample; and projection means for focusing said shaped electron beam onto said sample, wherein said shaping means comprises: splitting means for generating a plurality of sub-beams by splitting said electron beam; rectangular shaping means for shaping each of said plurality of sub-beams into a desired rectangular shape; and sub-beam deflection means for moving the irradiation position of each of said plurality of subbeams.

2. An electron beam exposure apparatus as claimed in claim 1, said rectangular shaping means shapes at least some of said plurality of sub-beams independently of each other into a desired rectangular shape, and said sub-beam deflection means moves the irradiation position of at least some of said plurality of sub-beams independently of each other.

3. An electron beam exposure apparatus as claimed in claim 1 or 2, further comprising blanker means for controlling said electron beam as to whether said electron beam is irradiated or not irradiated onto said sample.

4. An electron beam exposure apparatus as claimed in any one of claims 1 to 3, further comprising sub-beam blanker means for controlling each of said plurality of sub-beams independently of each other as to whether said each sub-beam is irradiated or not irradiated onto said sample.

5. An electron beam exposure apparatus as claimed in any one of claims 1 to 4, wherein said splitting means is a substrate having a plurality of first shaping apertures of a prescribed rectangular shape arranged at a prescribed pitch, and said plurality of sub-beams are a plurality of beams of a prescribed rectangular shape arranged at said prescribed pitch, and wherein said rectangular shaping means includes: first shaping deflection means for deflecting each of said plurality of sub-beams; a shaping aperture array having a plurality of second shaping apertures of a rectangular shape arranged to correspond with said prescribed pitch, wherein each of said plurality of sub-beams deflected by said first shaping deflection means is irradiated on a corresponding one of said plurality of second shaping apertures; and second shaping deflection means for deflecting back said plurality of sub-beams shaped through said plurality of second shaping apertures.

6. An electron beam exposure apparatus as claimed in claim 5, wherein said first and second shaping deflection means each comprise two shaping deflection substrates, each including a plurality of apertures arranged to correspond with the arrangement of said plurality of sub-beams, a pair of deflection electrodes formed on both sides of each aperture for forming an electrostatic field, and shield electrodes formed in other positions flanking each of said apertures than the positions where said pair of deflector electrodes are formed, and wherein the direction of the electrostatic field formed by said deflection electrode pair on one shaping deflection substrate is oriented at 90 to the direction of the electrostatic field formed by the corresponding deflection electrode pair on the other shaping deflection substrate, and said two shaping deflection substrates are arranged in close proximity to each other.

7. An electron beam exposure apparatus as claimed in claim 5, wherein the substrate of said splitting means includes a plurality of shaping aperture sets each consisting of said plurality of first shaping apertures, and any one of said plurality of sets is selectively movable into a path of said electron beam.

8. An electron beam exposure apparatus as claimed in claim 1 or 2, wherein said sub-beam deflection means comprises two deflection substrates, each including a plurality of apertures arranged to correspond with the arrangement of said plurality of sub-beams, a pair of deflection electrodes formed on both sides of each aperture for forming an electrostatic field, and shield electrodes formed in other positions flanking each of said apertures than the positions where said pair of deflector electrodes are formed, and wherein the direction of the electrostatic field formed by said deflection electrode pair on one deflection substrate is oriented at 90 to the direction of the electrostatic field formed by the corresponding deflection electrode pair on the other deflection substrate, and said two deflection substrates are arranged in close proximity to each other.

9. An electron beam exposure apparatus as claimed in claim 4, wherein said sub-beam blanker means comprises: a blanker deflection substrate which includes a plurality of apertures arranged to correspond with the arrangement of said plurality of sub-beams, a pair of deflection electrodes formed on both sides of each aperture for forming an electrostatic field, and shield electrodes formed in other positions flanking each of said apertures than the positions where said pair of deflector electrodes are formed; and a shield plate for blanking said plurality of sub-beams deflected by said deflection electrode pairs.

10. An electron beam exposure apparatus as claimed in claim 1 or 2, wherein said deflection means includes main deflection means and sub deflection means whose deflection range is smaller than the deflection range of said main deflection means, and wherein a main deflection range corresponding to the deflectable range of said main deflection means is split into a plurality of sub deflection ranges each corresponding to the deflectable range of said sub deflection means, each of said sub deflection ranges is split into a plurality of sub-beam deflection ranges each corresponding to the deflectable range of said sub-beam deflection means, with the deflection positions of said main deflection means and said sub deflection means held fixed, exposure is performed within each of said sub-beam deflection ranges by varying the deflection position of said sub-beam deflection means, exposure is performed within each of said sub deflection ranges by repeating the exposure within each of said sub-beam deflection ranges while varying the deflection position of said sub deflection means, and exposure is performed within each of said sub deflection ranges by repeating the exposure within each of said main deflection ranges while varying the deflection position of said main deflection means.

11. An electron beam exposure apparatus as claimed in claim 10, wherein said plurality of sub-beam deflection ranges are spaced apart from each other.

12. An electron beam exposure apparatus as claimed in claim 11, wherein said deflection means further includes sub sub-deflection means whose deflection range is smaller than the deflection range of said sub deflection means and wider than the pitch at which said plurality of subbeam deflection ranges are arrayed, each of said sub deflection ranges is split into a plurality of total sub-beam deflection ranges each corresponding a range outside said arrayed plurality of deflection ranges, with the deflection positions of said main deflection means, said sub deflection means, and said sub sub-deflection means held fixed, exposure is performed within each of said sub-beam deflection ranges by varying the deflection position of said sub-beam deflection means, exposure is performed within each of said total sub-beam deflection ranges by repeating the exposure within each of said sub-beam deflection ranges while varying the deflection position of said sub subdeflection means, and exposure is performed within each of said sub deflection ranges by repeating the exposure within each of said total sub-beam deflection ranges while varying the deflection position of said sub deflection means.