JPH0488625A - Charge particle beam exposure system - Google Patents

Abstract

PURPOSE:To optimize the table speed by a method wherein the stored amount of control data to be fed to at least one means out of a blanking means, a beam formation means, a beam deflection means is monitored so as to change the table speed within the stripes. CONSTITUTION:A data tracking circuit 25 stores the control data for main deflection to monitor the increase and decrease in the data for changing the table speed according to the monitor results. The circuit 25 composed of an FIFO buffer 30 temporarily storing the data loads the X, Y position data for alignment in the main deflection with the coordinate of respective sub-fields in the exposure order from the left side to justify the data to the right. An up/down counter 31 counts up according to the signal of the data processor part 24 loading a pair of data to the buffer 30 while counting down according to the signal of the exposure control unit 23 unloading a pair of data from the buffer 30. The contents of the counter 31 represent the data amount stored in the buffer 30. The table speed specifying circuit 32 inputs seven levels from the counter 31 so as to specify one speed out of four stages of table speed.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Application Field) The present invention mainly relates to a charged beam lithography system for manufacturing masks for semiconductor integrated circuit manufacturing, and particularly relates to a charged beam lithography system that uses a variable shaped beam method and a vector scan method. The present invention relates to a charged beam drawing device that uses a method to draw a pattern while continuously feeding a table.

(Prior Art) In writing using an electron beam, the area that can be written by deflecting the beam is limited to a few square meters at most in order to ensure the required pattern accuracy. Therefore, in order to draw a wider area than this, it is necessary to place the substrate on an XY table and move the table to draw a wider area. There are basically two methods for moving this table:

One method is to stop the table, draw the range that can be covered by deflection, and then advance the table step by step, which is called the step-and-repeat method. The other method is to draw the area that can be covered by deflection in the Y direction while continuously moving the table in one direction of X (or Y), and connect the resulting long and narrow strip-shaped areas by step-feeding the table in the Y direction. This method is called the continuous table feeding method. The former is wasted time because no drawing is performed while the table is moving, so the latter is naturally faster.

Furthermore, in the classification of drawing methods, the following three methods have been put into practical use. This will be explained using FIG. Figure 6 (a
) is the longest-established rusk scan method, and is a sweeping type in which a table is continuously fed and a beam is scanned in a direction perpendicular to the table. Since this method scans areas without figures to be irradiated with the beam turned off, the only way to increase the speed is to increase the scanning speed or widen the scanning width. There are two methods shown in FIG. 6(b): continuous table feed and step φ and repeat, both of which are called vector methods because the beam is moved only where there is a figure to be irradiated in smaller area units. In particular, (bl) is a method of rask scanning at that position, and (b2) is a method of irradiating a beam shaped at that position while feeding it in steps.

By the way, when the table is continuously fed in Fig. 6(b), (bl)
Alternatively, when using the method (b2), the problem is how to determine the table speed. As the area and density of the figure to be irradiated increases, the amount of data and irradiation time required to draw it increase. If the table speed exceeds the data supply capacity, it will naturally become impossible to draw. The ideal speed setting method is to make the table speed as fast as possible without rendering drawing impossible, but in reality this is often difficult, so the following method has been adopted.

(1) Table tracking As shown in Figure 7, the beam is irradiated as shown in a while it can follow the progress of the table, but when it can no longer be followed b
to avoid local overcrowding. but,
If overcrowding continues for a long time, it becomes impossible to draw.

(2) Retry For example, in the table tracking shown in FIG. 7, drawing is interrupted at a good break before drawing becomes impossible, and the table speed is lowered before drawing after the interruption. As the frequency of retries increases, the loss also increases, so when a certain number of retries is exceeded, the table speed may be lowered.

(3) Prediction Method As shown in FIG. 8, the table is sent at a constant speed, but in this method, the degree of overcrowding of each stripe is predicted in advance and a safe speed is set.

Various methods have been implemented for this prediction, but highly accurate predictions (eg, software simulations, etc.) take time, which can offset the effects of this.

(Problems to be Solved by the Invention) As described above, in conventional electron beam lithography apparatuses using a continuous table feeding method, it has been difficult to optimize the table speed, which has been a factor in reducing the lithography throughput. Moreover, such a problem is not limited to electron beam lithography apparatuses, but also applies to ion beam lithography apparatuses that use ion beams.

The present invention has been made in consideration of the above circumstances, and its purpose is to provide a charged beam lithography apparatus that can optimize the table speed and improve the lithography throughput.

Structure 1 of the Invention C (Means for Solving the Problems) The gist of the present invention is to change the table speed within the stripe, and in order to change the table speed, blanking means, beam shaping means, and beam deflection are used. The objective is to monitor the degree of processing of control data sent to means, etc.

That is, the present invention includes blanking means for turning on and off a charged beam, beam shaping means for shaping the cross-sectional shape of the beam, and beam deflection means for positioning the shaped beam on a sample surface. In a charged beam drawing apparatus that draws a desired pattern on the substrate while continuously moving an X-Y table placed thereon in one direction, the blanking means while monitoring the current position of the table;
a drawing control unit that sends control data to the shaping means and the deflection means in a scheduled order; a data processing section that supplies control data to this unit; and a data processing section that is arranged between the drawing control unit and the data processing section, A storage section stores control data to be sent to at least one of the ranking means, the shaping means, and the deflection means in chronological order, and monitors the amount of control data that remains in this storage section, and monitors the amount of control data that remains in this storage section. A table speed setting section for changing the table speed in relation to the table speed is provided.

(Operation) The above-mentioned table tracking attempt is naturally necessary in order to avoid producing defective masks. The prediction method explained in FIG. 8 is an effort to increase the table speed as much as possible within a safe range, but it is still far from ideal. This is because it is generally not the case that pattern density occurs more in the X direction in Figure 8 and less in the Y direction, and even if the safe maximum speed is predicted for each stripe, for example If the densest pattern is distributed like a, all stripes will have the same speed.

In order to make a big step forward from the prediction method that has such drawbacks, it is necessary to change the table speed even within the stripe, and the problem is how to change the speed based on what.

According to the present invention, the table speed is changed based on the amount of control data staying in the storage section (the amount of staying data), so in areas where the pattern density is high, the amount of staying data increases and the table speed is slowed to reduce the pattern density. In the lower part, there is less retained data and the table speed can be increased.

This makes it possible to make the table speed as fast as possible without exceeding the supply capacity of the data processing section.

(Example) Hereinafter, details of the present invention will be explained with reference to illustrated embodiments.

FIG. 1 is a schematic configuration diagram showing a continuous table feed variable shaped beam type electron beam lithography apparatus according to an embodiment of the present invention. Note that this figure shows only the parts necessary for explaining the present invention.

In the figure, reference numeral 10 denotes an electron optical lens barrel, and within this lens barrel 10, an electron gun 11, deflectors 13, 14, 15, 16, various lenses (not shown), etc. are installed. Electrons emitted from the electron gun 11 pass through a magnetic lens (not shown), which is composed of, for example, five stages, and are irradiated onto a substrate to be processed 23 mounted on an XY table 12. The deflector 13 is a blanking electrode for turning on and off the beam, and is driven by a blanking driver 14.

Data specifying the shot time of the beam is input to the driver 14, and at the same time as the driver 14 inputs this data, it turns on the beam using the blanking electrode 13, and turns it off after the specified time has elapsed.

The deflector 15 is a shaping deflector for shaping the beam, and D
It is driven by a molding deflection driver 16 made of AC, AMP, etc. Data giving the X and 7 dimensions for shaping the beam into a rectangle or triangle is input to the driver 16.

Deflector 17 is a sub-deflector and deflects the beam by approximately 50 μm.
It can be positioned within an area of ×50 μm and is driven by the sub-deflection driver 18. The deflector 19 is a main deflector that can position the beam within an area of approximately 2 mm x 2 + * m,
It is driven by the main deflection driver 20. Driver 18.2
0 is input with data for positioning the beam in the X and Y directions, but the actual beam position moves in the direction of the sum of the vector given by the main deflection and the vector given by the sub deflection. Further, 21 is a laser length measuring device for measuring the coordinates of the X-Y table 12, and 22 is a motor for driving the table. Note that there are actually two laser length measuring devices 21 and two motors 22, one for the X direction and one for the Y direction.

23 is a drawing control unit that drives the X-Y table 12 based on various control data given from the data processing section 24, and controls the molding deflection driver 16 and the main deflection driver while grasping the current position of the table 12. 20. Control data is passed to the sub-deflection driver 18 and the blanking driver 14 in a predetermined order. The data processing unit 24 is a computer 24
a and a data expansion unit 24b. Reference numeral 25 denotes a data tracking circuit (table speed setting section) which is a feature of this embodiment, and this circuit 25 will be described later.

An overview of the processing and flow of graphic data will be explained with reference to FIG. The graphic data for manufacturing the mask is prepared on the disk of the computer 24a, and as shown in FIG. It has a geometric expression format. When drawing starts, the computer 24a transfers the first stripe data from the disk to the data development unit 24b, and positions the XY table 12 at the P coordinate. The data development unit 24b generates various control data while performing subfield division as shown in FIG. 2(b) and shot division as shown in FIG. 2(e). In the subfield division shown in FIG. 2(b), the area is divided into areas that can be irradiated with the sub-deflection, and control data for positioning a certain area of the figure with the main deflection is generated in accordance with the drawing order. The shot division shown in Fig. 2 (C) divides the figure in each subfield specified by subfield division into shots, which are the minimum units, and uses control data that specifies the shape and size of beam shaping, and sub-deflection to each. Control data for positioning each shot and control data for designating each shot time to the blanking driver 14 are generated in a predetermined shot order.

The drawing control unit 23 in FIG. 1 receives these control data, and controls each driver 14 and 23 to irradiate a predetermined shot at a predetermined position while monitoring the current position of the table.
16, 18, and 20, the control data is sent at a predetermined timing. At this time, even if control data for a shot to be irradiated to a certain position is supplied to the drawing control unit 23, the table 12 has already passed that position at that time, and if table tracking cannot cope with the problem, drawing becomes impossible. Therefore, it is desirable to make the table speed as fast as possible without exceeding the supply capacity of the data processing section 24.

Next, the data tracking circuit 2, which is a feature of this embodiment, will be described.
5 will be explained. This data tracking circuit 25
In this method, main deflection control data is retained and its increase/decrease is monitored, and the table speed is changed based on the monitoring results.

FIG. 3 shows a specific configuration of the data tracking circuit 25. In the figure, reference numeral 30 denotes a FIFO buffer for temporarily storing main deflection control data, and this FIFO buffer 30 has an X mark for positioning the main deflection to the coordinates of each subfield in the drawing order.

Y position data is loaded from the left and aligned to the right. 31 is an up/down counter, which is counted up by the signal LO^D by which the data processing unit 24 loads one pair of data into the FIFO buffer 30, and by the signal UNLOAD by which the drawing control unit 23 unloads one pair of data.
count down. Therefore, the contents of the counter 31 represent the amount of data remaining in the FIFO buffer 30. 32 is a table speed designation circuit, from counter 31 to R
1, D'1. f! 2. f)'2. ... Input the seven levels of fI4 and set the table speed s1 to 4 stages. S2. S3
+ Specify one of s4.

FIG. 4 shows the operation of the table speed designation circuit 32. Hysteresis is provided to prevent the speed specification from hunting when the level changes. In this way FI
When the amount of data remaining in the FO buffer 30 increases, the table speed is increased, and when the amount of data remains in the FO buffer 30, the table speed is decreased. An automatic adjustment function operates, thereby realizing efficient high-speed drawing. Actual numerical examples are S, -5is/see, S2
= 20■/sea, S 3-351m/sec,
S 4-50mm/sec, and the distance required for acceleration/deceleration between table speed levels is approximately 21,
The level width was determined to be 3 mm, which is 1.5 times this value. 3■
The maximum amount of main deflection data that can be entered is 1200 at the address of the FIFO buffer 30, so the actual level width is set to 10.
24 addresses, and the capacity of the FIFO buffer 30 is 81
92 addresses x (data length). Also, level 4
(i) The main deflection data up to 4) is stored in the FIFO buffer 3.
After storing it at 0, Table 12 was started, and the initial speed was set to 84.

FIG. 5 shows another example of the data tracking circuit 25.
The IFO buffer 30 and up/down counter 3 are the same as in the case of FIG. 3, but the following configuration is adopted in place of the table speed designation circuit 32 as means for setting the table speed. That is, the output and minimum speed of the D/A converter 35 to obtain an analog signal proportional to the amount of retained data are set by a computer, and the output of the D/A converter 36 to obtain the analog signal is added by the operational amplifier 37. The output is used as a signal to command the table speed.

Although the examples shown in FIGS. 3 and 5 both monitor the retention amount of main deflection data, the same objective can be achieved by monitoring sub-deflection data. However, in the case of secondary deflection, the response of the table is slow despite the large amount of data, so it is difficult to use a large capacity FI.
This is disadvantageous because it requires an FO buffer. Furthermore, in some cases, it is possible to draw one mask without changing the molding deflection data or blanking data, and if the mask is used repeatedly, it cannot be used as is if a method is adopted that does not generate data. However, if data representing the number of retained shots can be derived, a similar function can be realized without sacrificing the spirit of the present invention.

As described above, according to this embodiment, the data tracking circuit 25 retains the control data in the FIFO buffer 30 and monitors its increase/decrease, and the table speed designation circuit 32 changes the table speed.
The table speed can be set as fast as possible without exceeding the supply capacity of 4. In other words, the table speed can be optimized and the drawing throughput can be improved. The present embodiment is particularly effective when drawing a reticle for a reduction type stepper with few repeated patterns in the continuous table feed direction. In addition, even for wafer direct writing in which many repeating patterns are arranged, a highly responsive table with a small weight can be used.
Since it is sufficiently effective, it is also applicable to electron beam direct writing.

Note that the present invention is not limited to the embodiments described above. In the embodiment, an electron beam lithography apparatus using a continuous table feed variable shaped beam method has been described, but the present invention can also be applied to a continuous table feed vector scan method. Furthermore, the present invention can be applied not only to electron beams but also to ion beam drawing apparatuses using ion beams.

Further, the configuration of the electron optical lens barrel is not limited to the same as shown in FIG. 1, but can be changed as appropriate according to specifications. Also,
In addition, various modifications can be made without departing from the gist of the present invention.

[Effects of the Invention] As described in detail above, according to the present invention, the table speed is adjusted within a stripe by monitoring the amount of control data that is retained to be sent to at least one of the blanking means, the beam shaping means, and the beam deflection means. By changing this, it is possible to optimize the table speed and realize a charged beam lithography apparatus that can improve the lithography throughput.

[Brief explanation of the drawing]

FIG. 1 is a schematic configuration diagram showing a table continuous feed variable shaped beam type electron beam drawing apparatus according to an embodiment of the present invention, FIG. 2 is a schematic diagram for explaining a method of drawing figures with the above apparatus, FIG. 3 is a block diagram showing a specific configuration of the data tracking circuit used in the above device, FIG. 4 is a schematic diagram for explaining the operation of the data tracking circuit, and FIG. 5 shows a modified example of the data tracking circuit. The block diagram and FIGS. 6 to 8 are diagrams for explaining a conventional electron beam lithography apparatus, respectively. 10... Electron optical lens barrel, 11... Electron gun, 12... X-Y table, 13.15, 17.19... Deflector, 14... Blanking driver, 16.18 .20... Deflection driver, 21... Laser length measuring device, 22... Drive motor, 23... Drawing control unit, 24... Data processing unit, 24a... Computer, 24b. ...Data expansion unit, 25...Data tracking circuit, 30...FIFO buffer, 31...Up/down counter, 32...Table speed designation circuit, 35.36...D/A conversion vessel. Applicant's agent Patent attorney Takehiko Suzue Figure 3 Figure 4 Figure 6 Factor

Claims (3)

[Claims] (1) Blanking means for turning on and off the charged beam;
It is equipped with a beam shaping means for shaping the cross-sectional shape of the beam, and a beam deflection means for positioning the shaped beam on the sample surface, while continuously moving an X-Y table on which a substrate to be processed is placed in one direction. In a charged beam drawing apparatus for drawing a desired pattern on the substrate, a drawing control unit sends control data to the blanking means, shaping means, and deflection means in a scheduled order while monitoring the current position of the table; A data processing section that supplies control data to this unit, and a data processing section that is arranged between the drawing control unit and the data processing section, and that is arranged to supply control data to at least one of the blanking means, the shaping means, and the deflection means over time. A storage unit that stores data in column order, and a table speed setting unit that monitors the amount of accumulated data, which is the amount of control data that remains in this storage unit, and changes the table speed in relation to the amount of accumulated data. A charged beam lithography device featuring: (2) The charged beam lithography according to claim 1, wherein the table speed setting unit grasps the amount of retained data at a plurality of levels, and changes the table speed in relation to changes in the levels. Device. (3) The table speed setting unit determines the table speed by grasping an amount approximately proportional to the amount of retained data and linearly converting each of the amounts.
The charged beam lithography device described.