JPH05190435A - Electron beam lithography method of semiconductor device - Google Patents

Abstract

PURPOSE:To provide a drawing method wherein a high-accuracy drawing operation can be performed irrespective of whether a defect and a foreign body exist on a wafer or not and whether the distortion of the wafer exists or not and to provide a drawing method wherein a drawing treatment and a calibration treatment for the dislocation of a position are performed with good efficiency. CONSTITUTION:An alignment mark on the surface of a wafer 2 is scanned by using an electron beam 4; its position and its height are detected. Pieces of their data are compared with stored values which have been classified in advance according to the type, the production stage and the inside-the-face distribution of the wafer at the point of time when the pieces of data are detected. When the pieces of data are outside the permissible range of the distribution of the stored values, corresponding stored values are read out instead of them. By means of the values, the deflection amount of the electron beam, the amount of the electron beam, the shape of a photoelectronic face, and the focal position of the electron beam are corrected by controlling a deflector 7, a rotary lens 6 and an object lens 8. A region which is drawn by one electron beam irradiation treatment is decided on the basis of the scanning width of the electron beam in a direction perpendicular to the movement direction of a specimen stand 1. Whenever the drawing operation of the region is finished, a calibration treatment by detecting a mark is executed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique which is particularly effective when applied to a semiconductor manufacturing technique and a fine processing technique using an electron beam. For example, a circuit pattern or the like is transferred onto a wafer whose surface is coated with a resist. It is related to the technology useful for utilizing the technology.

[0002]

2. Description of the Related Art In a manufacturing process of a semiconductor integrated circuit device, an electron beam (electron beam) is used to expose a resist on a wafer coated with the resist to expose a resist on the wafer. Exposure technology is known (for example, Industrial Research Institute Co., Ltd., 1986, 1).
"Electronic Materials", November 18, 1986, separate volume, pages 110-114). With this electron beam exposure technology,
Since the integrated circuit pattern is directly drawn by irradiating an electron beam on the surface of the resist-coated wafer, it is better than the conventional light exposure technology (technology that forms the integrated circuit pattern on a photomask once and transfers it onto the wafer). A fine integrated circuit pattern can be formed.

Therefore, in order to effectively draw a fine integrated circuit pattern by an electron beam, accurate alignment between the electron beam and the sample is premised, and an electron beam (focused beam) is generated based on the design data of the integrated circuit. ) Is controlled by a computer, a high-speed and high-precision fine processing can be realized.

[0004]

However, the present inventors have clarified that the above-mentioned technique has the following problems. That is, since a wafer on which an integrated circuit pattern is drawn has a distortion and a surface step during many steps, a figure (exposure figure) drawn by an electron beam should be corrected according to the step. But,
It is necessary to maintain the alignment accuracy and dimensional accuracy of the exposure patterns that are subsequently superimposed on each other. For this reason, for each unit exposure area (see FIG. 3) formed on the wafer, the computer control of the electron beam drawing apparatus must perform a predetermined correction according to the step difference on each design data to perform fine processing. I won't. However, if many processes are performed and the pattern is overlaid on the underlying pattern on the semiconductor wafer, distortion or steps may occur on the wafer surface while the process is repeated, or foreign matter may adhere to the wafer surface during the work. Or
Alternatively, a mark defect may occur on the wafer during the processing process, and it may not be possible to accurately detect the reference mark formed on the wafer. For this reason, even if the alignment of the wafer and the circuit pattern is corrected based on the mark position detection value, a registration error occurs due to the mark detection error, resulting in a problem that high drawing accuracy cannot be obtained.

Further, when drawing a fine pattern using an electron beam, it is known that the variation of the irradiation position of the electron beam itself gradually increases as the irradiation time increases. In order to suppress this, it is necessary to perform calibration based on the mark position on the stage or the mark position on the wafer whether the irradiation position is a desired position at regular time intervals. In a drawing device using an electron beam, in order to maintain high-speed drawing processing, the time interval for position detection and height detection of a reference alignment mark should be made as short as possible to perform position mark calibration in a short period of time. Is desired. Further, the calibration process needs to be performed a plurality of times while the pattern is entirely written on the wafer, and at what timing these processes are performed determines the drawing accuracy.

Further, when a process of drawing a pattern on the wafer is executed while continuously moving the stage on which the wafer is mounted in one direction, the direction of continuous movement of the stage and the coordinate axis of the underlying pattern already formed on the wafer are executed. It became clear that a rotation step was generated in the direction of. There may be a point at which a coordinate shift of about 500 μm occurs due to this rotation error. Conventionally, in order to correct this rotation error, a method has been adopted in which, after detecting the error, the stage on which the wafer is mounted is minutely rotated by the rotation error. However, when the electrostatic chuck method is adopted as the wafer holding means, it becomes difficult to correct the rotation. In the electrostatic chuck method, the wafer is fixed on the stage by static electricity. However, the holding force of the electrostatic chuck method is temporarily released when the wafer is moved from the room to the vacuum state of the drawing apparatus. No positioning can be done to optimize the rotational position of the.

Further, when a circuit pattern is drawn using an electron beam, the sample is charged when the sample (wafer) is irradiated with the electron beam, and the electric charge at this time causes an electron to be drawn when the circuit pattern is drawn next. The irradiation position of the line shifts from the desired position. Since the shift amount of the irradiation position depends on the current value of the electron beam, it becomes large during the drawing process in which the current value becomes large, which causes a problem in drawing the circuit pattern with high accuracy. In addition, the deflection distortion of the electron beam becomes large at the boundary portion of the deflection area (the boundary portion with the adjacent deflection area) that can be drawn by the one electron beam processing, and high-precision drawing cannot be performed near this boundary portion. Furthermore, the electron beam itself influences each other's electric charge according to Coulomb's law, and the outline becomes unclear during the drawing process, and fluctuations occur in the electron beam itself, which may affect the dimensions and position of the circuit pattern to be drawn. There are variations. These problems become noticeable especially in an extremely fine circuit pattern drawing process in which high accuracy is expected.

The present invention has been made in view of the above circumstances. A first object of the present invention is to reduce the influence of defects on the wafer, foreign matters and distortion of the wafer, thereby enabling highly accurate electron beam writing. It is to provide a drawing method. A second object is to efficiently perform the mark detection operation and the electron beam irradiation operation in order to efficiently perform the electron beam pattern drawing process and the position calibration process in the drawing device in which the stage is continuously moved. It is to provide an electron beam drawing method. or,
A third object is to provide an electron beam drawing method capable of accurately correcting a rotation error of a wafer even in a drawing device adopting an electrostatic chuck method. or,
A fourth object is an electron beam drawing apparatus and a drawing method capable of drawing with high accuracy even when the irradiation position is shifted due to a charged wafer of an irradiated electron beam or its own electric charge. Is to provide. Regarding the above and other objects and novel features of the present invention,
It will be apparent from the description of this specification and the accompanying drawings.

[0009]

The typical ones of the inventions disclosed in the present application will be outlined below. That is, the drawing technique of the present invention is, for example,
When applied to electron beam drawing technology, when detecting multiple alignment mark positions and heights on the wafer, the accuracy of detection of mark positions and heights on the sample surface deteriorates due to distortion on the sample, defects, foreign matter, etc. In order to prevent the above, a plurality of detection data of the sample are subjected to statistical processing such as sample type, manufacturing process, in-plane distribution, etc. to exclude the detection data out of the distribution and use the remaining data for the electron beam (charged beam). And the formation of an electron beam and the beam focus position of the electron beam are aligned to draw a pattern on the sample. Further, when a pattern is written on the wafer while continuously moving the stage on which the wafer is mounted, the writing area on the wafer is orthogonal to the continuous movement direction of the stage depending on the deflection amount of the electron beam (charged beam). It is divided in the direction, and the calibration processing is performed by repeating the detection of the mark position and height and the pattern drawing for each area. Also, when drawing a pattern on the wafer while continuously moving the stage on which the wafer is mounted, the rotation error between the direction of continuous movement of the stage and the direction of the coordinate axis of the underlying pattern formed on the wafer is measured. If the rotation error is small, the electron beam deflection system is corrected, and if the rotation error is small, the electron beam deflection system is corrected and the electron beam shape is rotated. By adding
The drawing is performed by correcting the rotation error. Further, in a method of writing a pattern on a wafer by combining movement of a wafer, deflection scanning of an electron beam, and turning on / off of an electron beam, the dimensions, position coordinates, etc. of the pattern to be drawn on the wafer are centered on a plurality of reference points. It is converted into the coordinates described above, and a plurality of exposures are performed for one drawing point based on the respective coordinates.

[0010]

According to the electron beam writing technique of the present invention described above, when detecting a plurality of alignment mark positions and heights on a wafer, whether the detection data reflects the position and height of the sample, It is possible to eliminate by the statistical processing of the detection data whether or not defects on the sample surface, foreign substances, and distortion of the sample are included. For example, based on the deposition of a specific wiring film or insulating film on the wafer, the distortion of the wafer after etching,
It is possible to obtain the distribution of the mark position and the height and remove the detection data deviated from the distribution. Further, when a pattern is written on the wafer while continuously moving the stage on which the wafer is mounted, the drawing area on the wafer is divided into a direction orthogonal to the continuous movement direction of the stage according to the drift amount of the electron beam. As a result, the mark position and height can be detected for each area and the pattern can be drawn for each area without significantly increasing the drawing time to improve the accuracy. Also, while writing a pattern on the wafer while continuously moving the stage on which the wafer is mounted, the rotation error between the direction of continuous movement of the stage and the direction of the coordinate axis of the underlying pattern formed on the wafer is measured. If the rotation error is small, the electron beam deflection system is corrected, and if the rotation error is large, the electron beam deflection system is corrected and the shape of the photoelectron surface of the electron beam is corrected. By correcting the rotation, the rotation error is corrected and the image is drawn. In the method of writing a pattern on the wafer by combining movement of the wafer, deflection scanning of the electron beam, and on / off of the electron beam, one type of pattern data that defines the dimensions, position coordinates, etc. of the pattern to be written on the wafer. On the other hand, a plurality of coordinates centering on a plurality of reference points are obtained, and multiple exposures are performed with reference to these coordinates. As a result, fluctuations in the electron beam are reduced, and the connection accuracy at the electron beam deflection boundary can be improved.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the accompanying drawings. FIG. 1 is a block diagram showing a configuration of an electron beam exposure apparatus according to the present invention, FIG. 2 is a perspective view showing a configuration of an electron beam apparatus, and FIG. 3 is a relationship between an alignment state of unit exposure regions on a wafer and alignment marks. An explanatory view showing an example, FIG. 4 is a statistical processing flow of mark detection data, FIGS. 5 and 6 are explanatory views of data processing by in-plane distribution of a wafer, and FIGS.
11 is an explanatory diagram showing an example of a process for obtaining a deflection coefficient according to the height of the exposed surface of the exposure target (wafer), and FIG. 11 is an explanatory diagram showing an example of a reference mark detection position when obtaining the deflection coefficient,
FIG. 12 is an explanatory view of a drawing method by the electron beam exposure apparatus of this embodiment, FIG. 13 is an explanatory view showing a rotation error occurring between the coordinate axes of the wafer circuit pattern and the moving direction of the stage, FIG.
Is an explanatory diagram showing a procedure for generating data using the electron beam irradiation method of the second embodiment, and FIG. 15 is an explanatory diagram showing a procedure for electron beam irradiation based on the above data.

(First Embodiment) First, a drawing procedure by an electron beam drawing apparatus will be described with reference to FIGS.
The sample stage indicated by 1 in the drawing is in the horizontal direction in the X direction and the Y direction.
X- that can move in any direction (for example, the horizontal and vertical directions in FIG. 3)
On the sample stage 1, a semiconductor wafer 2 having an electron beam resist coated on the surface thereof, for example, is placed as an object to be exposed (sample) on the sample stage 1. An electron beam source 3 is provided above the sample table 1, and an electron beam 4 is emitted toward a wafer 2 mounted on the sample table 1. Between the electron beam source 3 and the sample stage 1, an electron beam adjusting device of an electron optical system including a molding machine 5, a rotating lens 6, a deflector 7, an objective lens 8 and the like is provided.

The electron beam 4 emitted from the electron beam source 3 passes through the shaper 5, and the apertures 5a, 5b (FIG. 2) provided therein convert the photoelectron surface into a predetermined shape (rectangular shape). After that, the rotation lens 6 corrects the rotation and the deflector 7 corrects the deflection, whereby an arbitrary position on the wafer (object to be exposed) 2 is irradiated with an electron beam at an arbitrary light amount. And by the objective lens 8,
Focusing of the irradiated electron beam 4 on the surface of the object 2 to be exposed is performed.

At the position above the object 2 to be exposed on the sample table 1, the height of the incident portion of the electron beam 4 on the wafer 2 is measured by detecting the reflected light of the electron beam 4 applied to the wafer 2. Z A measuring device 23 is formed. In addition, at the position above the sample table 1, secondary electrons or scattered light 22 when the electron beam 4 or another light source 21 is irradiated on the mark forming portion on the sample upper surface, are detected, and based on this detection result, A mark detector 24 for detecting the mark position on the exposed object is provided. The Z measuring device 23 and the mark detector 24 are both connected to the control computer 16 and the detection result is obtained by the computer 1.
It will be sent to 6.

The shaper 5 for converting the photoelectron surface into a predetermined shape is connected to a first buffer memory 25 and a control computer 16 which can be accessed at high speed via a shaper controller 9 and a molding signal generator 10. The rotating lens 6 is connected to the calculating unit 11, the first buffer memory 25, and the control computer 16 via the rotating lens control unit 12. The rotary lens controller 12 is connected to the second buffer memory 25 and the control computer 16. In addition, the deflector 7 is provided with the calculation unit 11 via the deflection control unit 14 and the deflection signal generation unit 15.
And the second buffer memory 25, respectively.
On the other hand, the objective lens 8 is connected to the computing unit 11 and the control computer 16 via the objective lens control unit 17 and the second buffer memory 25. A sample table controller 19 that controls the operation of the XY stage of the sample table 1 is connected to the control computer 16.

The control computer 16 is connected to a drawing data storage unit 20 which stores, for example, a large amount of information about a figure to be exposed on the object to be exposed 2, and which is composed of a hard disk device having a large storage capacity. The drawing data storage unit 20 is configured to transfer predetermined graphic information appropriately selected by the control computer 16 to the first buffer memory 18, which can be accessed at high speed, as necessary.
Further, the control computer 16 is provided with an exposure target (wafer) 2
A database 26 is connected to store data regarding the mark position and height of the wafer according to the wafer type (CMOS LSI, bipolar LSI, ...), the manufacturing process in which the exposure is performed, and the in-plane distribution of the wafer. Then, this data is linked to an exposure job program corresponding to each wafer 2 and can be sequentially fetched.

On the other hand, the arithmetic unit 11 includes a buffer memory 18
Based on the graphic information held in the, the shape of the photoelectron surface of the electron beam 4 (square size, etc.) and the deflection amount (control amount of electromagnetic polarization, electrostatic polarization), etc. are calculated, and the various calculation results are calculated. It outputs various control signals according to. The shaping signal generation unit 10 and the shaping device control unit 9 receiving the control signal from the arithmetic unit 11 control the shaping device 5 (control of the shape of the photoelectron surface) based on the signals. Further, the control signals are the rotary lens controller 12, the deflection signal generator 15, and the objective lens controller 1.
Also sent to 7. Of these, the rotary lens controller 12 controls the operation of the rotary lens 6, and the deflection signal generator 15 cooperates with the deflection controller 14 to control the deflector 7. The objective lens controller 17 controls the objective lens 8 (focusing). In this case, the control calculator 16 and the deflection signal generator 15, the calculator 16 and the rotary lens controller 12, and the calculator 16 and the objective lens controller 1
A second buffer memory 25 (storing means) is provided between 7 and the second buffer memory 25 is the same as the Z buffer.
Height data from the measuring device 23 is input.

Then, the detection results (height data) obtained by the Z measuring device 23 are element regions P ₁ , P ₂ ... P _n described later.
The exposure surface height measured by the detection is recognized for each deflection area (for each in-plane distribution) (see FIG. 3), and is stored in the second buffer memory 25 in advance according to the recognized height. The thus-corrected arbitrary correction coefficient is selectively read. The correction coefficient read out in this manner is used for the deflection signal generator 15,
It is sent to the rotary lens controller 12 and the objective lens controller 17 as a signal indicating that fact.

FIG. 2 is a perspective view showing the construction of the drawing apparatus of the electron beam apparatus. On an XY stage (sample stage) 1 which is movable in the horizontal plane of the sample shown in FIG. 2, a wafer (exposed object) 1 having an electron sensitive resist or the like coated on its surface is mounted. An electron gun 3 serving as an electron beam source is provided above the sample table 1 and configured to emit an electron beam 4 toward the wafer 2. In the path of the electron beam 4 from the electron beam source 3 to the sample stage 1, an electron optical system including a molding machine 5, a rotating lens 6, a deflector 7, an objective lens 8 and the like is provided, and an electron beam adjusting unit (device) is provided. Is configured.

An operation procedure of wafer alignment / pattern exposure using the electron beam writing apparatus having the above structure will be specifically described with reference to FIG. As shown in FIG. 3, the exposed surface of the semiconductor wafer 2 has a plurality of element regions (unit exposure regions) P ₁ ,
It is divided into P ₂ ... P _n , and each element region P ₁ , P ₂ ...
At a predetermined position of P _n , alignment marks K ₁ , K ₂ ... K _n
Is provided. This alignment mark is for maintaining the overlay accuracy of a plurality of circuit patterns used in each step in an exposure step that is repeated a plurality of times when forming an integrated circuit element on the wafer surface ( (For alignment). In order to actually perform the alignment using this alignment mark, the electron beam 4 or the light 22 from the light source 21 is applied to the alignment mark (K ₁ , K ₂ ... K _n ) and the scattered light at this time is Z. The alignment mark position is three-dimensionally recognized by detecting it by the detector 23 or the mark detector 24. The data thus detected is represented as L _i (x _i , y _i , z _i ), and the data L _i is information on the alignment K _i formed on the wafer 2 (data of the mark forming path; ideal). Whether or not the grid points are correctly reflected is determined by, for example, the flowchart shown in FIG. By comparing the data contents according to the flow charts in this way, alignment mark detection errors that occur when foreign matter adheres to the object to be exposed (wafer) 2 or a base step is generated are eliminated.

In this flowchart, first, step 1
A plurality of alignment marks (K
₁ , K ₂ ... K _n ) are measured, and in step 2, the measurement data is subjected to coordinate conversion to obtain data L _i (x _i , y _i , z _i ). Then, calculates an error [Delta] L _i from the ideal grid point (the value at the design stage) that represents the data L _i and the mark has been該得, in step 3-6 said error [Delta] L _i is the past detection data Based on whether or not it is below the reference value already determined by the statistical processing for each product type, below the reference value determined for each manufacturing process, and below the reference value determined for each lot. Whether or not it is within the range determined for each in-plane distribution of the wafer. Here, each product type refers to a difference in the type of LSI formed on a wafer (for example, bipolar LSI, CMOS ...), and each process refers to a difference in manufacturing process such as aluminum 1 layer, interlayer insulating film, aluminum 2 layer ... Say. Then, when all of these judgments are “Yes”, in step 7, this data L _i is adopted to calculate data representing a new mark position and height. On the other hand, if any of Steps 3 to 6 is “No”, it is further determined in Step 8 whether or not the error measurement ratios of K _{1 to} K _n are the predetermined value (5%) or more. Here, the error measurement ratio means all marks (Fig.
Of each mark), the ratio represents the number of marks whose detection value deviates from the reference value. When the result of this determination is "No", the process returns to step 1 and data detection is performed again, while when "Yes",
The data deviating from the reference value is replaced with the already stored value (step 9), and the subsequent processing is performed.
The criteria used for the determination in steps S3 to S6 are:
This is the data detected by the processing performed up to that point, and this data is classified and stored according to the type of product to be exposed, manufacturing process, lot, and in-plane distribution, and the wafer detection data corresponding to this classification condition. , Is read out as appropriate when new data is detected and is used as a reference for comparison. In this method, as the inspection of the object to be exposed is repeated, a larger amount of data is accumulated, so that the accuracy of the determination standard can be improved over time.

Next, a procedure for eliminating and supplementing abnormal data relating to the detection data L _i based on the in-plane distribution of the object to be exposed (wafer) will be described with reference to FIGS. 5 and 6. As shown in FIG. 5, when the detected data L _i includes the abnormal point data (L _{i + 4} ), the abnormal point data is subtracted from the error ΔL from the ideal grid point of each detected data. (L _{i + 4} ), average value processing is performed on this, and the processed data is replaced with the abnormal point data.

By the way, the element region (unit exposure region) P ₁ ,
When performing exposure processing on the object (wafer) 2 to be exposed based on the graphic information in the drawing data storage section 20 for P ₂ ... P _n , an objective lens corresponding to the height of the exposed surface of the object (wafer) 2 to be exposed Focusing by the control of 8 and correction for the deflection coefficient of the deflector 7 (first correction), the coordinate system forming the element regions P ₁ , P ₂ ... P _n on the exposed object 2 and the electron optical system. It is necessary to correct the deflection coefficient of the deflector 7 for the mismatch of the coordinate systems and the distortion of the exposed object 2 (second correction). Therefore, in this embodiment, the following correction processing is performed after the above-described data processing is completed.

That is, the calculation of the deflection coefficient of the deflector 7 according to the height of the exposed surface and the correction thereof are performed according to the following procedure. The height is corrected prior to the exposure operation on the object to be exposed 2. In this case, first, a sample having a high surface 32 and a low surface 31 having different height differences is provided on the sample table 1, and the standard marks M _H of the high surface 32 and the low surface 31 of the sample,
Detect the position of M _L. In this case, in detecting M _H , the position of the sample stage 1 is moved to deflect the electron beam 4 to irradiate M _H as shown in FIGS. 24 (FIG. 1), thereby obtaining standard mark M _H detection results at different deflection positions. On the other hand, when detecting a standard mark M _L, as shown in FIGS. 9 and 10, similarly to the above, detecting the standard mark M _L at different deflection position (by moving the sample stage 1). Then, a plurality of positions of the deflection area while moving the sample stage 1 by a predetermined distance _{m 1, m 2, m 3} , ... m n ( FIG. 1
By detecting the positions of the standard marks M _H and M _L in 1), the amount of distortion corresponding to the change in the height of the exposed surface can be obtained (in this case, the positions m ₁ , m ₂ , m ₃ , It is assumed that the data representing m _n is accurately obtained by the sample table controller 19 based on the feed amount of the sample table 1 (XY stage).

The deflection correction coefficient to be given to the deflection signal generator 15 of FIG. 1 is such that X and Y are X and Y coordinates of deflection, ΔX,
When ΔY is the strain amount obtained from the position of the sample table 1,
Coefficients A ₀ , A ₁ with the left side of the following equations (1) and (2) represented as ₀ ...
_{A 9, B 0, B 1} ... is a B _9. ΔX = A ₀ + A ₁ X + A ₂ Y + A ₃ X ² + A ₄ XY + A ₅ Y ² + A ₆ X ³ + A ₇ X ² Y + A ₈ XY ² + A ₉ Y ³ (1) ΔY = B ₀ + B ₁ X + B ₂ Y + B ₃ X ² + B ₄ XY + B ₅ Y ² + B ₆ X ³ + B ₇ X ² Y + B ₈ XY ² + B ₉ Y ³ (2)

For the deflection correction coefficients A _n ^Z and B _n ^{Z at} an arbitrary sample surface height Z, the data of the standard mark _MH among the standard marks shown in FIGS. , Deflection correction coefficients obtained by applying the equation (2) to A _n ^H and B _n ^H , and the standard mark M
_Assuming that the deflection correction coefficients obtained by using _L are A _n ^L and B _n ^L , A _n ^Z = (A _n ^H −A _n ^L ) Z / H (3) B _n ^Z = (B _n ^H −B _n ^L ) Z / H (4) (where H is the difference in height between the standard marks M _H and M _L ).

On the other hand, the objective lens 8 is corrected for the focal position and astigmatism. In this case, the focus correction current value I _r is determined according to the deflection coordinates X and Y, and is calculated by the following equation (5). I _r = α ₀ + α ₁ X + α ₂ Y + α ₃ X ² + α ₄ XY + α ₅ Y ² + α ₆ X ³ + α ₇ X ² Y + α ₈ XY ² + α ₉ Y ³ (5) where α ₀ is the sample surface height Z. Which changes according to
Required in (6). α ₀ ^Z = α ₀ ^L + (α ₀ ^H −α ₀ ^L ) / H · Z (6)

Similarly, the current value I _{SX for} astigmatism correction is I _SX = β ₀ + β ₁ X + β ₂ Y + β ₃ X ² + β ₄ XY + β ₅ Y ² + β ₆ X ³ + β ₇ X ² Y + β ₈ XY ² + Β ₉ Y ³ I _SY = γ ₀ + γ ₁ X + γ ₂ Y + γ ₃ X ² + γ ₄ XY + γ ₅ Y ² + γ ₆ X ³ + γ ₇ X ² Y + γ ₈ XY ² + γ ₉ Y ³ (7) Where β ₀ and γ ₀ are functions of Z, β ₀ ^Z = β ₀ ^L + (β ₀ ^H −β ₀ ^L ) Z / (8) γ ₀ ^Z = γ ₀ ^L + (γ ₀ ^H −γ ₀ ^L ) Z / H is given as (9).

In the electron beam exposure apparatus of this embodiment, the correction coefficient A _n ^z (A ₁ ^Z , A ₂ ^Z ...
A ₉ ^Z ), B _n ^z (B ₁ ^Z , B ₂ ^Z ... B ₉ ^Z ), and α ₀ ^Z ,
Regarding β ₀ ^Z and γ ₀ ^Z , the height difference H between the high and low surfaces is H.
Alternatively, the sample surface height including the outside thereof is apportioned, and the values corresponding to the respective sample surface heights Z ₁ , Z ₂ , Z ₃ , ... Z _n are calculated in advance as γ ₁ , γ _{2 ,.} Stored in the second buffer memory 25 together with ₃ , β ₁ , β _2, ... β ₃ (constant value),
Further, correction coefficients C _0i , C _1i , C _2i , C _3i , D _0i , D for alignment for each exposure area calculated for each alignment block K ₁ , K _2, ... K _m (see FIG. 3) are provided. _1i ,
D _2i and D _3i are stored in the second buffer memory 25,
This example in the group individual element regions P _1, P ₂ ... second correction coefficient dependent on the exposure surface height Z for each of the exposure areas corresponding to P _n, in accordance with the measured exposure level height Z Selected from the values stored in advance in the buffer memory 25 of the second buffer memory 25
The read signal is given to the deflection signal generator 15, the rotary lens controller 12, and the objective lens controller 17 directly without interposing the control computer 16. By this procedure,
An exposure operation is performed on each unit exposure area.

The control computer 16 of the electron beam exposure apparatus displays data on the object to be exposed (wafer), that is, the type of wafer,
A database 26 having information on the manufacturing process, starting lot, in-plane distribution, etc. is connected so that these data can be linked to the job data for exposure control of the object to be exposed, and the actual mark of the object to be exposed can be obtained. By discriminating the detected data according to the flow shown in FIG. 4 based on the extracted data, it is possible to surely improve the overlay accuracy of the exposure on the exposed object.

FIG. 12 is an explanatory view showing an example of a drawing method by the electron beam exposure apparatus. In this drawing method, the XY stage (sample stage) 1 of the electron beam drawing apparatus (FIG. 2) is continuously moved only in the Y direction (vertical direction in FIG. 12). This is because in the above drawing apparatus, the XY stage is made of a material (ceramic) other than metal in order to prevent overcurrent. In this case, the weight of the stage becomes large, and the stage is moved in both XY directions. This is because it is necessary to improve (enlarge) the function of the entire device in order to continuously move the device.

In carrying out the drawing method of continuously moving only in one direction (Y direction), the exposure surface of the wafer 2 is vertically divided into a direction orthogonal to the continuous moving direction of the stage (vertical direction in the figure) to form a plurality of drawing areas. Control is performed (exposure) by using the drawing operation in each drawing area as one operation unit. Note that the operation time that the beam device can continuously draw in one operation is constant, and therefore, the drawing area set as one operation unit is smaller than the range that can be sufficiently drawn within the operation time. Set to area. Then, the position of the mark (alignment mark) on the wafer and the mark height are detected each time one operation unit is completed. And
Then, the next drawing operation is performed, and the series of processes is repeated. In this case, between the mark detection and the drawing operation, the XY stage 1 is stepwise moved in the direction orthogonal to the continuous movement direction thereof, and the mark detection and the pattern drawing operation in the area are alternately repeated. .. In this embodiment, the operation (drawing / calibration) from the point M _{1 to the} point E ₁ is set as one operation unit, and the drawing area (the area indicated by the diagonal lines in FIG. 12) is twice as wide as the drift amount of the electron beam ( The width (width in the direction orthogonal to the stage movement direction) of the two grids in FIG. 4)
Has been decided.

To detect the alignment mark (alignment), the mark portion (formed in the wafer's describe region) of the sample is irradiated with an electron beam, and the mark detector 24 shown in FIG. Or secondary electrons) are detected (first, mark detection is performed along the path M1 indicated by the one-dot chain line in the figure). The mark position data detected by the mark detector 4 is stored in the database (mark position data storage unit) 26 shown in FIG. At this time, if the relative position between the electron beam to be drawn and the sample (wafer) 2 drifts, the position accuracy of the drawing pattern deteriorates. Therefore, in this drawing method, the electron beam 4 and the sample (wafer) 2 are not removed every fixed time. A correction (calibration process) for the drift of the relative position with respect to is performed.

Alignment mark (alignment mark)
Since K ₁ , K _2, ... Are formed in the peripheral portion of the describe area of the wafer 2 as shown in FIG. 3, it is not possible to detect the alignment mark and the drawing mark at the same time. Therefore, the mark detection for each of the divided areas is performed, and then the circuit pattern in the area is drawn (performed along a path E1 indicated by a two-dot chain line in the figure). In this drawing method, the mark position detection processing and the drawing processing are performed for each of the divided drawing areas of the wafer, which can reduce the stage movement time required to switch between the mark position detection processing and the drawing processing. it can. As a result, the deviation between the electron beam 4 and the sample (wafer) 2 is efficiently corrected, and the drawing accuracy of the circuit pattern on the wafer is further improved.

FIG. 13 schematically shows the continuous movement direction of the stage and the mechanical rotation error of the pattern on the wafer. In the semiconductor manufacturing line, if the mechanical rotation error of the pattern on the wafer is not managed, the coordinates may shift by a maximum of about 500 μm / 100 mm depending on the location. Therefore, in the drawing method of the present embodiment, these errors are eliminated by the X- of the stage.
A rotation angle difference θ between the coordinate axis in the Y movement direction (for example, the X axis) and the coordinate axis of the base pattern formed on the wafer 2 (X axis).
(FIG. 13) is measured, and the rotation error is corrected at an arbitrary drawing point F (x, y) based on the measured value θ to obtain the corrected coordinate P (x, y). In measuring the detection value θ, at least two alignment marks formed on the wafer surface are detected and compared with each other.

This rotation error θ deteriorates the positional accuracy of the drawing pattern. This is because in drawing with an electron beam, the square photoelectron surfaces formed by the apertures shown in FIG. 2 are continuously arranged to form a desired pattern, and thus the square error becomes larger as the rotation error θ increases. This is because a large step is generated at the junction between the photoelectron surfaces, and if a straight line is to be drawn, the photoelectron surfaces are arranged in a stepwise manner and a step is generated. In the present embodiment, when the value θ is small, the rotation error is corrected only by adding the correction to the deflection system of the electron beam 4 (the correction amount of the rotation error is added to the deflector 7 in FIG. 1), and θ Is large, correction is performed by the deflector 7 and the rotation lens 6 for the electron beam is also corrected to correct the rotation error in the shape (square) of the photoelectron surface itself. In this way, by correcting the photoelectron surface itself according to the rotation error, the rectangular photoelectron surface is angle-corrected with respect to the coordinate system in which the rotation error occurs, and the figure (pattern ) Does not cause a step. By performing the correction according to the rotation error in this way, the mechanical rotation between the continuous movement direction of the stage (X and Y directions in the figure) and the pattern on the wafer (coordinates on the wafer, x and y directions in the figure) Regardless of the error θ, it is possible to always maintain high pattern drawing accuracy.

(Second Embodiment) FIG. 14 is a drawing in which the sample stage (X-Y stage) 1 of the drawing apparatus is moved stepwise in the XY direction when drawing is performed using the electron beam apparatus of FIG. It is explanatory drawing which shows the method of conversion of the drawing data (exposure coordinate system) by the method (drawing method of 2nd Example). In this drawing method, in order to specify a drawing area (exposure coordinate point) to be drawn on the wafer, first, coordinates (Fx, Fy) in a field (unit exposure area) on the drawing apparatus side and four reference values centered on the coordinates (Fx, Fy) Coordinates (Fx ₊ Δ, Fy ₊ Δ), (Fx
_{- Δ, Fy + Δ),} (Fx - Δ, Fy - Δ), (Fx + Δ,
Fy _- Δ) is set. Suppose now that a predetermined pattern coordinate (Px, Py) is drawn based on the coordinate (Fx, Fy). In this case, first, four reference coordinates (F
x ₊ Δ, Fy ₊ Δ), (Fx _- Δ, Fy ₊ Δ), (Fx
_- delta, Fy _- centered on _{Δ), - Δ), (} Fx + Δ, Fy
Four modified coordinates of the pattern coordinates (Px, Py) are obtained. In this case, the four coordinates are (Px _- Δ, Py
_{- Δ), (Px + Δ} , Py - Δ), (P + xΔ, Py
₊ Δ), (Px _- Δ, Py ₊ Δ). And this 4
Based on the pattern data (coordinates) of each type, the drawing processes are superposed four times. In actual drawing, four converted pattern coordinates are calculated for many pattern coordinates. 14
Of the data (a) to (d) of the coordinates (Fx, Fy) are X
When a predetermined amount of movement + Δ is shifted in the Y and Y directions (Fx ₊ Δ, Fy ₊ Δ), −Δ in the X direction and + Δ in the Y direction.
When shifted (Fx _- Δ, Fy ₊ Δ), ₊ in the X direction
When shifted by Δ in the Δ and Y directions (Fx ₋ Δ, Fy
_- delta), X-direction, when obtained by both -Δ shifted in the Y direction _{(Fx + Δ, Fy - Δ} ) of the original pattern data (Px,
The converted value of the pattern data (coordinates) after the shift for Py) is shown. In this example, the coordinate (Px, Py) of the pattern data is shifted in the opposite direction by the amount that the shift direction of the center coordinates (Fx, Fy) of the field has moved.
The drawing data, which is displaced by the amount of ± Δ shift in the field coordinates, is corrected in the opposite direction. The above shift amount Δ is
The maximum electromagnetic deflection width and the maximum electrostatic deflection width of the electron beam drawing apparatus in both the X and Y directions are set to 1/2 or less of the total (drawable range). In order to superimpose the processing, the drawing point after one movement is always in the drawing range.

FIG. 15 schematically shows the above-mentioned drawing method, and four reference coordinates (F) in which the shift amount is shifted ± Δ in the X and Y directions in one field (unit exposure area).
x, Fy), (Fax, Fay), (Fbx, Fb
y), (Fcx, Fcy), (Fdx, Fdy) are provided, and one drawing point (Px, P
y) are represented as different coordinates. Then, the drawing process is repeated four times with each of the above four reference coordinates as the center.

In this drawing method, deflection control is performed by combining deflection of the electron beam and ON / OFF of the electron beam. The electron beam 4 is deflected by a combination of electromagnetic deflection and electrostatic deflection (electromagnetic deflection has a maximum beam deflection width of about 5 mm, and electrostatic deflection has an electron beam deflection width of about 50 to 500 μm).
Therefore, the drawable range of the drawing apparatus with the XY stage fixed is the sum of these two deflection widths. Therefore, as described above, the amount of movement (shift amount Δ) of the XY stage is set to the electromagnetic deflection width and the electrostatic deflection width of the electron beam 4 when performing the four-time overlapping exposure for one pattern data. By shifting 1/2 of the sum, the exposure of four times is efficiently performed. 4 in 1 drawing area
When the exposure process is performed twice, fluctuations of the electron beam 4, particularly obscuration of the electron beam 4 at the boundary between the drawing regions, is averaged, and accurate irradiation is possible.

As described above, in the present embodiment, when detecting a plurality of alignment mark positions and heights on the wafer, the mark positions and heights on the sample surface are affected by the strain on the sample, defects, foreign matters, and the like. In order to prevent deterioration of detection accuracy of the sample, the detection data deviated from the distribution is excluded from the plurality of detection data of the sample by statistical processing such as product type, manufacturing process, in-plane distribution, etc. Since the deflection of the (charged beam) and the formation of the electron beam and the beam focus position of the electron beam are aligned to draw a pattern on the sample, the above-mentioned method is used when detecting a plurality of alignment mark positions and heights on the wafer. It is possible to judge whether the detection data reflects the position and height of the sample, or includes defects on the sample surface, foreign matter and sample strain, and eliminate them. Further, when a pattern is written on the wafer while continuously moving the stage on which the wafer is mounted, the writing area on the wafer is orthogonal to the continuous movement direction of the stage depending on the deflection amount of the electron beam (charged beam). Since the calibration processing is performed by dividing the area into directions and repeating the detection of the mark position and height and the pattern drawing for each area, the calibration processing can be performed without significantly increasing the drawing time. Also, when drawing a pattern on the wafer while continuously moving the stage on which the wafer is mounted, the rotation error between the direction of continuous movement of the stage and the direction of the coordinate axis of the underlying pattern formed on the wafer is measured. If the rotation error is small, the electron beam deflection system is corrected, and if the rotation error is small, the electron beam deflection system is corrected and the electron beam shape is rotated. Is added to correct the rotation error and perform highly accurate drawing. Further, in a method of writing a pattern on a wafer by combining movement of a wafer, deflection scanning of an electron beam, and turning on / off of an electron beam, the dimensions, position coordinates, etc. of the pattern to be drawn on the wafer are centered on a plurality of reference points. Since the plurality of exposures are performed on one drawing point based on the respective coordinates after the conversion, the fluctuation of the electron beam is reduced and the connection accuracy at the electron beam deflection boundary can be improved.

[0041]

The effects obtained by the typical ones of the inventions disclosed in the present application will be briefly described as follows. That is, even if a defect, a foreign substance, or a distortion is generated on the wafer surface, highly accurate drawing processing can be performed. Further, since the wafer drawing process and the calibration process can be efficiently performed, the processing speed is increased as a whole of the drawing process and the operation efficiency is improved. Further, even if a rotation error occurs between the direction of continuous movement of the XY stage and the direction of the coordinate axis of the underlying pattern formed on the wafer, this can be corrected to perform accurate drawing. .. Since multiple exposures are performed for one drawing point, fluctuations of the electron beam are reduced and drawing accuracy at the electron beam deflection boundary portion is improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an electron beam exposure apparatus according to the present invention.

FIG. 2 is a perspective view showing a configuration of an electron beam device.

FIG. 3 is an explanatory diagram showing an example of a relationship between an alignment state of unit exposure areas on a wafer and alignment marks.

FIG. 4 is a statistical processing flowchart of mark detection data.

FIG. 5 is an explanatory diagram of data processing based on in-plane distribution of a wafer.

FIG. 6 is an explanatory diagram of data processing based on in-plane distribution of a wafer.

FIG. 7 is an explanatory diagram showing an example of a method of obtaining a deflection coefficient according to the height of an exposed surface of an object to be exposed (wafer).

FIG. 8 is an explanatory diagram showing an example of a method of obtaining a deflection coefficient according to the height of an exposed surface of an object to be exposed (wafer).

FIG. 9 is an explanatory diagram showing an example of a method of obtaining a deflection coefficient according to the height of an exposed surface of an object to be exposed (wafer).

FIG. 10 is an explanatory diagram showing an example of a method of obtaining a deflection coefficient according to the height of an exposed surface of an object to be exposed (wafer).

FIG. 11 is an explanatory diagram showing an example of reference mark detection positions when obtaining a deflection coefficient.

FIG. 12 is an explanatory diagram of a drawing method by the electron beam exposure apparatus of this embodiment.

FIG. 13 is an explanatory diagram showing a rotation error generated between the coordinate axis of the wafer circuit pattern and the moving direction of the stage.

FIG. 14 is an explanatory diagram showing a procedure for generating data used when executing the electron beam irradiation method of the second embodiment.

15 is an explanatory diagram showing a procedure of electron beam irradiation based on the data shown in FIG.

[Explanation of symbols]

1 sample stage (X-Y stage) 2 wafer 4 electron beam 6 rotating lens 7 deflector 8 objective lens 12 rotating lens control unit 14 deflection control unit 16 control computer 17 objective lens control unit 20 drawing data storage unit 23 Z measuring device 24 mark detector 24 26 database P _1, P ₂ ... P _n element region K _1, K ₂ ... K _n alignment marks M _H, M _L standard mark

Claims (4)

[Claims] 1. A semiconductor wafer surface is scanned with an electron beam to detect at least one of the positions of a plurality of alignment marks formed on the wafer or the height of the wafer surface, and the detected data is used as data. Compared with a stored value pre-classified according to at least one of a stored value pre-classified according to the wafer state at the time of detection and a stored value pre-classified according to at least one of the distribution values of the stored data of the detected data. When it is determined that the electron beam is deviated, the stored value corresponding thereto is read out, and the stored value is used instead of the deviated data, and the deflection amount of the electron beam, the shape of the photoelectron surface of the electron beam, and the electron beam An electron beam drawing method, characterized in that at least one of the focal positions is corrected to expose the resist applied on the wafer surface. 2. An electron beam drawing method for exposing a resist coated on a wafer by irradiating the wafer with an electron beam while moving an XY stage on which the wafer is mounted in one direction. The sample was divided into a plurality of exposure regions based on the electron beam scanning width in the direction orthogonal to the moving direction of the sample stage, and formed on the wafer surface every time the exposure process for one of the divided exposure regions was completed. An electron beam drawing method characterized by detecting at least one of a position and a height of an alignment mark and correcting the alignment deviation based on the detection result. 3. A rotation error between a continuous movement direction of an XY stage on which a wafer is mounted and a coordinate axis of a pattern formed on the wafer is measured, and a deflection amount of an electron beam is corrected or corrected based on the rotation error. An electron beam drawing method characterized in that at least one of correction of a rotation angle of a shape of a photoelectron surface of an electron beam is performed. 4. A combination of electron beam deflection control and step movement in at least one of the X-direction and the Y-direction of an XY stage on which a wafer whose surface is coated with a resist is mounted is combined to obtain a desired wafer surface. In an electron beam drawing method for exposing a pattern, a drawing area smaller than a drawable range is formed by electron beam deflection irradiation around a predetermined reference point, and at least two drawing reference points are formed in this drawing area. An electron beam drawing method characterized in that the predetermined reference points are sequentially aligned with the two or more drawing reference points and two or more electron beam exposure processes are performed on the same pattern.