JPH0594942A - Measuring system for dimension of integrated circuit - Google Patents

Abstract

(57) [Abstract] [Purpose] The target pattern is accurately detected even if there are variations in the drawing position of the alignment mark itself, variations in the exposure position of the pattern, and differences in the orthogonality between the exposure apparatus and the measurement apparatus. [Structure] Position data of each chip drawing from the exposure apparatus 200 in advance in a drawing position data table 11, alignment mark value variations during alignment in an alignment initial value table 10, differences in alignment values and differences in orthogonality between exposure apparatuses 200. Stage orthogonality correction table 1
It is characterized in that the position deviation correction amount calculation unit 9 refers to these tables to correct the stop position of the stage or the beam shift amount when the stage is moved for each chip.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention irradiates a pattern such as metal, semiconductor, or insulator in an integrated circuit such as VLSI with a charged beam and detects reflected electrons or secondary electrons to measure and measure pattern dimensions. The present invention relates to an integrated circuit dimension measuring system for improving pattern alignment accuracy in a pattern measuring / inspecting apparatus for inspecting defects. In particular, it relates to a technique for achieving the full automation of the measurement / inspection with high accuracy and high throughput.

[0002]

2. Description of the Related Art Along with the miniaturization of the pattern size of LSI and the like, a dimension measuring device using a charged beam is used. In this type of device, the stage is moved so that the measurement pattern is located directly below the beam, and the initial setting of the objective lens,
Focusing is performed by fine adjustment, the charged beam is scanned, the secondary electron signal waveform in the direction perpendicular to the measurement pattern is obtained, and the dimension of the measured pattern is measured. Various methods for measuring dimensions from the secondary electron signal waveform include setting the appropriate slice level in the secondary electron signal, binarizing it, and measuring the pattern dimensions from the rising and falling intervals, as well as edge measurement. .A method of approximating each of the baselines with a straight line and measuring the pattern size from the interval of the intersections (Japanese Patent Laid-Open No. 61-
No. 80011) are used.

In a series of operations for measuring such pattern dimensions, the stage is moved to the position of the pattern to be measured, the SEM image is captured, the target pattern is detected, and then measurement and inspection are performed. However, if the accuracy of stage movement is poor, the target pattern will not be within the field of view and measurement will not be possible. In addition, as shown in FIG. 4A, when the position is close to the end of the pattern or when the position near the intersection with another wiring is measured,
As shown in (b), if the stage stop position is deviated, the pattern is not rectangular at the beam scanning position (dotted line).
Pattern width cannot be measured accurately. Such a shift in the stop position of the stage itself may be poor in stop accuracy of the stage itself, but as shown in FIG. 5, the chip position itself may vary during writing. FIG. 5A shows the influence of random variations (factors such as stage stop accuracy).
(B) shows the influence (X
Since the orthogonality between the axis and the Y-axis is mechanically adjusted, it deviates from the strict right angle.) FIG. 5C shows the effect caused by the mounting of the mask on the exposure apparatus rotating with the stage movement axis of the exposure apparatus ( The chip is being rotated and drawn). The actual pattern is a combination of these. If there is such variation in the drawing position itself, even if the stage accuracy is good, the target pattern cannot be detected, and accurate measurement cannot be performed. The same applies to the pattern defect inspection.

Incidentally, as shown in FIG. 6, design coordinates for designating a pattern on a wafer are defined by a wafer coordinate system x _W , y _W with the wafer as a reference. On the other hand, the movement of the wafer mounted on the stage is controlled by the stage coordinate system x _S , y _S with reference to the stage. When the wafer is loaded on the stage, the stage coordinate system and the wafer coordinate system do not match, so in order to correct both, the three parameters of the center coordinate deviation (ΔX, ΔY) and rotation angle θ are measured before measurement. Ask. Wafer alignment calculates these three parameters.

In wafer alignment, two or more patterns having known design coordinates are detected, and the stage coordinates are read and calculated. First, by initializing the wafer alignment, a deviation caused by a difference in wafer holding method between the exposure apparatus and the measurement apparatus is input in advance. With this initial value, the stage is moved to the approximate mark position and mark detection is performed. As a mark detection method, for example, there is a method of using a pattern having the same shape as a reference pattern that has been captured in advance as a mark and obtaining the position by pattern matching. The mark detection is first performed at a low magnification, the stage is moved to the detected coordinates, the magnification is made higher, and a two-stage method is performed again to prevent the influence of image distortion at a low magnification from entering. Mark detection detects two or more predetermined marks and calculates an alignment value from the coordinates.

An example is shown in which three marks are detected. Design coordinates (wafer coordinate system) of the mark center are (x _1D , y _1D ),
(X _2D , y _2D ), (x _3D , y _3D ), the measured stage coordinates are (x _1m , y _1m ), (x _2m , y _2m ), (x _3m , y _3m ).
Then, the alignment value can be calculated by the following formula.

(1) When three-point mark is detected ΔX = [x_1m+ X_2m-A (x_1D+ X_2D) cos θ + b (y_1D+ Y_2D) sin θ] / 2 ΔY = [y_1m+ Y_2m-B (y_1D+ Y_2D) cos θ−a (x_1D+ X_2D) sin θ] / 2 θ = tan^-1(A / B) A = a (x_2D-X_1D) (Y_2m-Y_1m) -B (y_2D-Y_1D) (X_2m-X_1m) B = a (x_2D-X_1D) (X_2m-X_1m) + B (y_2D-Y_1D) (Y_2m-Y_1m) A = (C / E)^0.5 , B = (D / E)^0.5 C = [(x_2m-X_1m)²+ (y_2m-Y_1m)²] (y_3D-Y_2D)²-[(x_3m-X_2m)²+ (y_3m-Y _2m )²] (y_2D-Y_1D)² D = [(x_3m-X_2m)²+ (y_3m-Y_2m)²] (x_2D-X_1D)²-[(x_2m-X_1m)²+ (y_2m-Y _1m )²] (x_3D-X_2D)² E = (x_2D-X_1D)² (y_3D-Y_2D)²− (X_3D-X_2D)² (y_2D-Y_1D)² (2) When two-point mark is detected (when one point cannot be detected) ΔX = [x_1m+ X_2m-A (x_1D+ X_2D) cos θ + a (y_1D+ Y_2D) sin θ] / 2 ΔY = [y_1m+ Y_2m-A (y_1D+ Y_2D) cos θ−a (x_1D+ X_2D) sin θ] / 2 θ = tan^-1(A / B) A = (x_2D-X_1D) (Y_2m-Y_1m)-(Y_2D-Y_1D) (X_2m-X_1m) B = (x_2D-X_1D) (X_2m-X_1m) + (Y_2D-Y_1D) (Y_2m-Y_1m) A = [{(x_2m-X_1m)²+ (y_2m-Y_1m)²} / {(x_2D-X_1D)²+ (y_2D-Y_1D)²}]^0.5  Also, using these alignment values, the wafer coordinate system
(X_W , Y_W ) And the stage coordinate system (x_S , Y_S ) Conversion
Is given by the following equation (where 2-point mark detection is
a = b). x_S = Ax_W cos θ-by_W sin θ + ΔX y_S = Ax_W sin θ + by_W cos θ + ΔX x_W = A (x_S −ΔX) cos θ−b (y_S −ΔY) sin θ y_W = A (x_S −ΔX) sin θ + b (y_S −ΔY) cos θ In such a wafer alignment process, the stage
The stop position may be misaligned and the target pattern may not be in the field of view.
Jijiru The reason why this position shifts is that the wafer
Position variation when loading on the
The variation in the drawing position of the ark itself may be mentioned. the latter
Is the position variation when loading on the stage of the exposure system.
Or due to variation in shape such as wafer diameter
It depends on the exposure apparatus. Therefore, exactly
In order to detect the target pattern,
Alignment as well as the range of position variation during scanning
Pattern search considering variations in the position of the mark itself
The range must be expanded. However, there is a
Not only will it take longer to expand the search range of
It is possible to mistake a similar pattern as the target pattern.
is there. For this reason, the target pattern should be set as accurately as possible on the stage.
It is necessary to move to the position. However, in the prior art,
To compensate for variations in the drawing position of the alignment mark itself
I couldn't do it right.

After the wafer alignment, the stage is moved to the target pattern position based on the design coordinates. However, the stage position deviates from the target pattern due to the following factors. The stop accuracy of the stage itself is low,
The design position and the actual position of the pattern are misaligned (deviation when drawing), the orthogonality of the coordinate system in which the pattern is drawn is misaligned with the orthogonality of the measurement system, and the beam is affected by the magnetic substance. Can be bent. This can be prevented by mounting a laser interferometer, measuring coordinates, and performing foot-back control to improve stopping accuracy. In addition, has a value peculiar to the device, and if the displacement amount is measured in advance and tabulated, and the position of the stage or the beam is displaced by the displacement amount according to the coordinates of the stage, The deviation can be prevented.

[0009]

However, in the conventional measuring apparatus, the wafer alignment is performed and the deviation between the stage coordinate system and the wafer coordinate system and the rotation amount are corrected when the stage is moved. The position is not corrected. This is because in a semi-automatic or manual device in which an operator measures and measures a target pattern, the stage stop accuracy is usually poor and the correction amount is smaller than the stop accuracy. However, in the case of fully automatic measurement, if the position is not corrected for each chip or pattern, the target pattern is low even if the stage stop position is made highly accurate using a laser interferometer. There is a problem that it takes time to detect it because it is only within the field of view at a magnification, or a detection error occurs. That is, the above deviation of the pattern position itself could not be corrected.

An object of the present invention is to measure pattern width using a charged beam even if there are variations in the drawing position of the alignment mark itself, variations in the exposure position of the pattern, differences in the orthogonality between the exposure apparatus and the measurement apparatus, etc. , The target pattern will be accurately detected.

[0011]

A dimension measuring system for an integrated circuit according to the present invention reads mark drawing position data of each wafer and drawing position correction data of each chip from an exposure device, and reads the mark position data to a measured pattern position. When the stage moves, the stop position of the stage or the beam shift amount is corrected for each chip.

Regarding the movement of the stage to the wafer alignment mark, the average alignment value in the exposure apparatus that draws the measurement wafer and the average value when the wafer is aligned in the measurement apparatus are calculated from the initial alignment table. Read. The correction value of the orthogonality of each exposure apparatus and the measuring apparatus is read from the stage orthogonality correction table prepared in advance. The alignment mark position data of the measurement wafer is read from the exposure device. Calculate the deviation between the alignment value at the time of writing and the average alignment value in the exposure device, add this value to the average alignment value on the device side, calculate the alignment value, and use this alignment value , The wafer coordinates of the target mark are converted into the coordinates of the stage coordinate system in which the orthogonality is corrected. By the above procedure, the wafer alignment is performed, the wafer coordinate system and the stage coordinate system are converted, and the stage is moved to the wafer alignment mark using the stage coordinate system.

Further, regarding the movement of the stage to the measurement point, the position information of the measurement pattern is read from the measurement condition file. The design coordinates are converted into stage coordinates based on the wafer alignment result. The amount of chip displacement to be measured is read from the drawing position data prepared in advance. The position of the measurement pattern is calculated by adding the correction value of 1 to the stage, and the stage is moved. The stage is moved to the measurement point by the above procedure.

[0014]

In the present invention, the drawing position correction data for each chip is read from the exposure device as the data for drawing. Therefore, even if the drawing position of each chip varies, the deviation can be obtained from the exposure device and corrected when the stage moves, so that the stage can be accurately moved to the target pattern position. Therefore, it is possible to narrow the search range of the stage or the SEM image for detecting the pattern.

[0015]

FIG. 1 is a diagram showing an example of the configuration of a dimension measuring system of the present invention. The inspection / measurement apparatus 100 includes an electron optical lens barrel (not shown) for irradiating a target sample with an electron beam, a stage (not shown) for moving the position of the sample, and a scanning electron microscope (hereinafter, SEM). 1) control computer 2, secondary electron detection system for taking in signals such as backscattered electrons or secondary electrons 3, signal processing / measuring unit 4 for processing it, lens barrel control system 5, lens barrel control table 6, power supply Interface 7, stage control system 8, positional deviation correction amount calculation unit 9
Is equipped with. 10 is an alignment initial value table, 1
Reference numeral 1 is a drawing position data file, and 12 is a stage orthogonality correction table. The lens barrel control system 5 refers to the lens barrel control table 6 to calculate how much current should be applied to each electromagnetic lens (voltage for an electrostatic lens) according to each irradiation condition, and to calculate each lens. The current is passed to control the beam. In the stage control system 8, in order to accurately stop the stage at a target coordinate, feedback is given to the stage movement while reading the current position coordinate with a laser interferometer, a linear scale, or the like. The signal / process measuring section 4 obtains a signal image captured in synchronization with the scanning of the beam from the secondary electron detection system 3 to detect a target pattern, measure dimensions, detect defects and the like. The above configurations except 9 to 12 are the ones that are also used in the normal SEM1. Reference numeral 200 denotes an exposure device.

In the present invention, in order to improve the positioning accuracy of the stage, a stage displacement correction amount calculation unit 9 is added. The positional deviation correction amount calculation unit 9 obtains positional data for drawing each chip from the exposure apparatus 200 in advance, and tabulates it as a drawing position data table 11 in order to correct positional deviation due to positional variations among the chips. In addition, the variation of the mark position at the time of alignment is displayed in the alignment initial value table 10, and the exposure apparatus 2
The stage orthogonality correction table 12 is provided to correct the difference in alignment value and the difference in orthogonality depending on 00. Using each of these tables 10 to 12, variations in the drawing position of the alignment mark itself, variations in the pattern exposure position, the exposure apparatus 200 and the inspection / measurement apparatus 100.
Correct the deviation due to the difference in the orthogonality of.

The pattern size in the wafer drawn by the exposure apparatus 200 is measured by the inspection / measurement apparatus 100. At this time, the drawing result data at the time of exposure is transferred to each table 10 to 12 connected to the stage positional deviation correction amount calculation unit 9 in the control computer 2 of the inspection / measurement apparatus 100 before the measurement. The data to be transferred includes the detection position information (or drawing position shift information) of the alignment mark and the drawing position data of each chip.

The control computer 2 of the inspection / measurement apparatus 100 registers the measurement conditions before the measurement. This may be designated each time, or a measurement condition created in advance may be called. The measurement conditions include the coordinate / shape of the pattern used in the wafer alignment, the position coordinate of the pattern for measuring the dimensions, and the irradiation condition of the beam used in these measurements. The position of the stage is moved based on this measurement condition. At this time, the position is corrected with reference to the drawing data by the position shift correction amount calculation unit 9. Hereinafter, a method of improving the accuracy of positioning by the dimension measuring system of the present invention, that is, a method of correcting the positional deviation during drawing will be described.

In the integrated circuit size measuring system of the present invention, position correction is performed when the stage is moved to the alignment mark for wafer alignment and when the stage is moved to the measurement pattern. [Movement of Stage to Wafer Alignment Mark] FIG. 2 is a diagram showing a procedure of moving the stage to the alignment mark position according to the present invention. Note that (S1) to (S9) indicate each step. The average alignment value (shift amount, rotation amount) of the exposure apparatus 200 on which the measurement wafer is drawn is read from the alignment initial value table 10 (S1). This is a table in which a plurality of wafers drawn by the corresponding apparatus are obtained in advance as a result of wafer alignment performed by the measurement system. This value is used as an initial setting value (default value) for wafer alignment. The average value of the alignment values obtained on the drawing apparatus side is also tabulated in this table. The correction values of the orthogonality of the exposure apparatus 200 on which the measurement wafer is drawn and the orthogonality of the inspection / measurement apparatus 100 are read from the stage orthogonality correction table 12 (S2). this is,
The result obtained from the result (positional coordinate) of detecting the same pattern in the chip on a plurality of wafers drawn by the corresponding device is tabulated in advance. This is because the degree of orthogonality between the X-axis and the Y-axis is mechanically adjusted, so that it deviates from a strict right angle, and is for correcting different amounts in individual devices. The alignment mark position data (drawing data or mark detection result) of the wafer to be measured is read (S3).
In the LSI process, in order to draw a pattern in accordance with a pattern already formed, the position of the mark formed in the wafer is detected and the position is corrected. The pattern is formed based on the mark detection position at this time. The detection position information of each alignment mark at this time or the drawing position data thereof is received from the exposure apparatus 200, and the alignment value at the time of drawing is calculated (the alignment value at the time of drawing may be directly received) (S4). .. It is better to receive this data from the exposure apparatus 200 in advance and convert the drawing position data into a file because the processing time becomes shorter, but it may be received on the spot. Based on the value from, the predicted position with the alignment mark is calculated. That is, the deviation (the parallel shift, the rotational deviation) between the alignment value at the time of writing and the average alignment value of the drawing apparatus is calculated. This value is added to the average alignment value of to obtain the alignment. Using this value, the wafer coordinates are converted into stage coordinates (S5).

The stage is moved to the calculated position (S6), the mark is detected, and its coordinates are stored (S7). By repeating this processing, the deviation between the wafer coordinates and the stage coordinates is calculated (S8, S9). In calculating the predicted position of the mark position, when detecting the second and subsequent marks, the next mark position should be calculated with reference to the mark position of the first point and considering the estimated rotation amount from here. You can [Movement of Stage to Measurement Point] FIG. 3 is a diagram showing a procedure of moving the stage to a measurement position according to the present invention. In addition, (S1
1) to (S19) indicate each step. After the wafer alignment is completed (S11), the design coordinates (wafer coordinates) of the position to be measured are read from the measurement condition file (S12). This position information consists of the coordinates of the pattern to be measured within the chip and the chip position to be measured. From this information, the coordinates are converted into coordinates with the center of the wafer as the origin (S13). Based on the wafer alignment result, the above design coordinates are converted into stage coordinates (x _S , y _S ) (S14). The conversion formula is as described in the prior art. At this time,
It is calculated with reference to the correction value of the orthogonality between the exposure apparatus 200 and the inspection / measurement apparatus 100. If the orthogonality is offset by δ,
This means that the Y axis is inclined by δ with respect to the X axis.
In this case, the result obtained by the above equation may be replaced as follows.

X = x + y sin δ, y = y cos δ The shift amount of the position to be measured is calculated based on the drawing position data. Since a light exposure device is usually used, the chips are collectively exposed. In this case, the amount of deviation within the chip may be considered to be uniform, so by reading the stage stop position deviation data on the exposure device side during drawing of each chip,
It can be the amount of deviation from the design position. Although there is a possibility of a rotational deviation as shown in FIG. 5C, this value can be detected during wafer alignment. That is, first,
When each mark is detected, the amount of rotation of the image on the screen is calculated. Next, the shift amount of the axis of the wafer coordinate system is calculated based on the position coordinates of each mark when a plurality of marks are detected. The difference between the two is the amount of rotation of the chip itself. The correction value of is added to calculate the position of the target pattern (S15).

The stage is moved to the above calculated position value (S1
6) Perform measurement and inspection (S17). All the measurements are performed by repeating this process (S18, S19). The same purpose can be achieved by shifting the beam instead of moving the stage. Further, if the coordinates are read after the stage is stopped and the beam is shifted by the amount of deviation from the target position, the accuracy can be further improved.

Although the above description has been directed to the measurement of the width of the pattern, the present invention is also effective when a defect inspection is performed with a charged beam at a high magnification.

[0024]

As described above, in the dimension measuring system of the present invention, the drawing position correction data for each wafer and each chip is read from the exposure apparatus so that the stage can be accurately moved to the measured pattern position. .. Therefore, the search range of the target pattern can be narrowed during wafer alignment, and the target pattern can be accurately detected in a short time. In particular, it is possible to reduce a mistake in erroneously recognizing a similar pattern in the vicinity as a target pattern. Further, at the time of measurement, even if an image is captured at a high magnification used for direct measurement, the target pattern is within the visual field. For this reason, it is possible to save time and effort for measuring at a high magnification after searching for a measurement pattern at a low magnification as usual, and then moving the stage or shifting the beam. Therefore, it is possible to reduce the detection error of the target pattern and increase the number of layers in the measurement.

[Brief description of drawings]

FIG. 1 is a diagram showing a configuration of an embodiment of an integrated circuit dimension measuring system according to the present invention.

FIG. 2 is a flowchart showing a procedure of wafer alignment according to the present invention.

FIG. 3 is a flow chart for explaining a stage moving (positioning) procedure to a measurement point.

FIG. 4 is a diagram for explaining the influence of poor positioning in a conventional scanning microscope.

FIG. 5 is a diagram showing an influence when there is a positional deviation of a pattern to be measured itself.

FIG. 6 is a diagram showing a relationship between a wafer coordinate system and a stage coordinate system.

[Explanation of symbols]

 1 Scanning Electron Microscope 2 Control Calculator 3 Secondary Electron Detection System 4 Signal Processing / Measurement Section 5 Lens Tube Control System 6 Lens Tube Control Table 7 Power Interface 8 Stage Control System 9 Position Error Correction Amount Calculation Section 10 Alignment Initial Value Table 11 Drawing position data file 12 Stage orthogonality correction table 100 Inspection / measurement device 200 Exposure device

Claims (3)

[Claims] 1. A pattern for each chip of a wafer processed by an exposure apparatus, which has a lens barrel for irradiating a charged beam, a stage for moving a sample, and means for detecting backscattered electrons or secondary electron signals.・ In the inspection / measuring device to be measured, the drawing position data of the mark of each wafer and the drawing position correction data of each chip are read from the exposure device, and the stage stop position or the beam shift amount when the stage is moved to the measured pattern position. A dimension measurement system for an integrated circuit, which corrects each chip. 2. An average alignment value in an exposure apparatus that draws a measurement wafer and an average value when the corresponding wafer is subjected to wafer alignment by the measurement apparatus are read from an alignment initial value table. The correction value of the orthogonality between the exposure apparatus and the measuring apparatus is read from the stage orthogonality correction table prepared in advance. The alignment mark position data of the measurement wafer is read from the exposure device. Calculate the deviation between the alignment value at the time of writing and the average alignment value in the exposure device, add this value to the average alignment value on the device side, calculate the alignment value, and use this alignment value , The wafer coordinates of the target mark are converted into the coordinates of the stage coordinate system in which the orthogonality is corrected. 2. The integrated circuit according to claim 1, wherein the wafer alignment is performed by the above procedure, the wafer coordinate system and the stage coordinate system are converted, and the stage is moved to the wafer alignment mark using the stage coordinate system. Dimensional measurement system. 3. The position information of the measurement pattern is read from the measurement condition file. The design coordinates are converted into stage coordinates based on the wafer alignment result. The amount of chip displacement to be measured is read from the drawing position data prepared in advance. The position of the measurement pattern is calculated by adding the correction value of 1 to the stage, and the stage is moved. The dimensional measurement system for an integrated circuit according to claim 1, wherein the stage is moved to the measurement point by the above procedure.