JPH021071A - Method and device for reading graphic - Google Patents

Abstract

PURPOSE:To read a graphic with high accuracy by generating a rectangular signal from a sensor signal by a threshold and generating a signal to indicate the feature of the graphic based on the counted value of clock pulses from the rise point of the rectangular signal to the fall point of the rectangular signal. CONSTITUTION:The graphic to be measured is image-formed on an image sensor 16 by an optical means, converted into an electric signal, and smoothed in an LPF43. In a threshold processing means 44, the rectangular signal is generated by forming this signal based on a prescribed threshold. In this case, the threshold is set to a certain value so that the rise point and the fall point of the rectangular signal can accurately correspond to the graphic to be measured. Counting area command means 45 and 47 respectively detect the rise point and the fall point of the rectangular signal, and counting means 45-51 count the clock pulses from a clock pulse generating means 41, which generate in a period from the rise point to the fall point. Further, a two-stage structure shown in a figure is expressed by separating the structure into an X directional system and a Y directional system. An electric signal value to indicate the feature of the graphic to be measured can be generated based on these counted values.

Description

Detailed Description of the Invention (Field of Industrial Application) The present invention is applicable to, for example, reduction projection exposure equipment, step-and-repeat equipment, PCBn optical equipment, metal exposure equipment, LCD
The present invention relates to a figure reading method and device for highly accurately detecting center coordinates, width dimensions, pitches, etc. of alignment marks, positioning grid patterns, etc. used in exposure devices such as CD exposure devices that require two-way placement.

(Prior art and its problems) Due to the miniaturization of photomask patterns due to the increased integration of semiconductor devices, in order to align the photomask pattern and the wafer, the position of the alignment mark formed on both must be raised higher. Accurately detect,
Alternatively, it is required to detect the width, pitch, etc. of a grid pattern for positioning with higher precision.

To partially meet these demands, a reticle pattern (a pattern that is the original of a photomask pattern) is projected onto a photodiode array via a magnifying optical system.
The optical signal corresponding to the reticle pattern output from the photodiode array is converted into an electrical signal, the electrical signal is quantized (binarized) with a predetermined threshold (voltage), and the quantized signal is converted into an electrical signal. A method is disclosed in Japanese Patent Publication No. 58-5041 in which a quantized pattern is formed by continuously scanning a quantized pattern, and this pattern is used to detect defects in a portion smaller than the minimum dimension of the pattern.
It is known from Publication No. 4.

The resolution in this method is determined by the pixel size of the photodiode array, that is, the number of bits, the bit interval, and the magnification of the magnifying optical system. As the magnification of the enlarging optical system increases, the detection field becomes narrower and becomes impractical, so there is a limit to how high the magnification of the optical system can be increased. For example 1024
In a device using a one-dimensional fixed image sensor with a bit spacing of 28 μm, if you try to secure the field of view of the magnifying optical system over 100 μm, the resolution will be only 0.14 μm, and the resolution will be 0.14 μm.
．． It cannot meet the requirement for positioning accuracy of 1 μm or less.

(Objective of the Invention) The present invention has been made to solve the above problems, and achieves high resolution by using an existing magnifying optical system and a solid-state image sensor to secure a sufficient field of view for the magnifying optical system. The object of the present invention is to provide a figure reading method and device that can read figures with high accuracy.

(Means for Solving the Problems) In order to achieve the second object, the present invention forms an image of an optical signal from a figure to be measured on a solid-state image sensor by an optical means, and converts the imaged optical signal into an image on a solid-state image sensor. In the figure reading method of converting into an electrical signal and reading the characteristics of the figure to be measured from the electrical signal, the electrical signal waveform obtained by scanning the optical signal on the solid-state image sensor is smoothed, and the smoothed A rectangular signal is generated by shaping an electrical signal waveform according to a predetermined threshold value, and the rising and falling points of the rectangular signal are detected, and the period up to the detected rising and falling points of the rectangular signal is 1. Count the clock pulses that occur at a constant period. , an electric signal value representing the characteristics of the ratio measurement pattern is created based on the counted number.

Furthermore, the present invention focuses an optical signal from a figure to be measured on a solid-state image sensor using optical means, converts the imaged optical signal into an electrical signal, and determines the characteristics of the figure to be measured from the electrical signal. The figure reading device includes a smoothing circuit that smoothes an electrical signal waveform obtained by scanning an optical signal on the solid-state image sensor, and a smoothed waveform output from the smoothing circuit that is shaped by a predetermined threshold value. a threshold processing means for generating a rectangular signal, a counting area specifying means for detecting a rising point and a falling point of the rectangular signal, a clock pulse generating means for generating a clock pulse of a constant period, and a clock pulse generating means for generating a clock pulse of a constant period; and counting means for counting the clock pulses generated during the period from the detected rise time to the fall time of the clock pulse.

Preferably, the figure reading device further includes an arithmetic processing device that calculates a value representing the characteristic of the figure to be measured based on the count output from the counting means.

(Operation) A figure to be measured is imaged on a solid-state image sensor by optical means, and the imaged optical signal is converted into an electrical signal. The electric signal is smoothed by a smoothing circuit, and then converted into a rectangular signal by a threshold processing circuit based on a predetermined threshold value. The threshold value is set to a predetermined value such that the rising and falling points of the rectangular signal correspond accurately to the figure to be measured. The counting area specifying means detects the rising and falling points of the rectangular signal, and the argument means detects the rising points.

The clock pulses from the clock pulse generation means generated during the period up to the falling edge are counted. Based on the counted number of pulses, it is possible to read the characteristics of the figure to be measured with high resolution.

(Example) The present invention will be described in detail below based on the drawings.

FIG. 1 is a schematic diagram showing a step-and-repeat device for creating a master mask from a reticle, to which a graphic reading device 1 of the present invention is applied.

As shown in the figure, the reticle 2 is loaded into a position fine adjustment device 8 having a motor M1, and a dry plate 6 made of a glass substrate whose surface is coated with an emulsion or a film of chrome or the like is loaded with a motor M2. M3. XY direction movement device, Z direction movement device.

It is loaded into a dry plate moving device 7 consisting of a position copying H1m device or the like. The enlarged pattern of the reticle 2 was projected onto a dry plate 6 by a light source 3 consisting of a xenon lamp or a mercury lamp and a lens 4.5, and by moving the position of the dry plate 6 with a dry plate moving device 7, the same pattern was continuously arranged. A mask pattern is created. The figure reading device 1 has alignment marks 9 (
2) is read, its positional coordinates are detected, and based on this, the relative position with respect to the dry plate 6 is adjusted to a desired position by means of a positional fine adjustment device 8.

FIG. 2 is a schematic diagram showing the graphic reading device l.

As shown in the figure, the photoresistor is not sensitive to light (λ
= 546 nm or 632.8 nm).

The light is reflected by a half mirror 12 via a lens 11, and is irradiated via a lens 13 onto a cross-shaped alignment mark 9 provided on the reticle 2. The light reflected from the alignment mark 9 passes through the half mirror 12 via the lens 13, and a part of the light is reflected by the half mirror 14 and passes through the lens 15 to the X-direction image sensor, which is an X-dimensional solid-state imaging device. 16 and an optical image of the alignment mark 9 is formed on the sensor 16. The rest passes through the half mirror 14 and passes through the lens 17 to the Y-direction image sensor 1, which is a Y-dimensional solid-state imaging device.
8 and the optical image of the alignment mark 9 is displayed on the sensor 1.
The image is formed on 8.

FIG. 3(a) is a diagram showing how an optical image 91 of the alignment mark 9 is formed on the light receiving element 161 of the X-direction image sensor 16, and FIG. corresponds to the position of each pixel shown in an enlarged view of the light receiving element 161 of
This is a diagram showing a video signal (?! multiplication signal) scanned and output from the X-direction image sensor 16. A video signal similar to that shown in FIG. 3 is also output from the Y-direction image sensor 18.

An electrical circuit for processing video signals from image sensors 16, 18, such as those shown in FIG. 3, is shown in FIG. That is,
The image sensors 16 and 18 are composed of CODs (charge-coupled devices) having a photosensitive function and a self-scanning function.
A start pulse generator 40 is connected to 6 and 18, and a start pulse for starting scanning of the light receiving elements of the image sensors 16 and 18 is supplied from the generator 4o. A frequency divider 42 is connected to the sensor 16 and supplies the sensor 16 with a pulse obtained by dividing the output from a clock pulse generator 41 that generates a clock pulse of, for example, 600 kHz to, for example, a frequency division ratio of 8 to 75 kHz. The image sensor 16 is
+6 light receiving elements for each input pulse frequency divided to 75kHz
1 performs charge transfer (signal transfer) and outputs the manipulated video signal.

The output terminal of the X-direction image sensor 16 is connected to a low-pass filter 43, and the output terminal of the low-pass filter 43 is connected to a threshold processing circuit 44.

The threshold processing circuit 44 includes a comparator 4
41 and a switching transistor 442.

The output terminal of the threshold processing circuit 44 is connected to the clock pulse terminal CP of a D-flip-flop (hereinafter referred to as rD-FFJ) 45 and to the clock pulse input terminal CP of another D-FF 47 via an inverter 46. be done. A constant voltage Vcc is input to both data input terminals of the D-FFs 45 and 47. The ζ output terminal of D-FF45 is connected to the AND circuit 48, and the D-FF47
The ζ output terminal of is connected to another AND circuit 49. The remaining input terminals of each of the AND circuits 48, 49 are connected to the clock pulse generator 41, e.g.
A clock pulse of 0kllz is provided. The output terminals of AND circuits 48 and 49 are pulse counters 50 and 49, respectively.
The output terminals of the pulse counters 50 and 51 are connected to an arithmetic unit 52.

The processing circuit for the output signal of the Y-direction image sensor 18 has the same configuration as the above-described circuit for processing the output signal of the X-direction image sensor 16, so a description thereof will be omitted.

Next, the operation of the electric circuit constituting the graphic reading device 1 shown in FIG. 4 will be explained with reference to FIGS. 5 to 7.

Optical image 91 of alignment mark 9 as shown in FIG.
When a pulse is input to the image sensor 16 from the start pulse generator 40 while the image is being formed on the light receiving element 161 of the X-direction image sensor 16, the image sensor 1
6 scans the light-receiving element 161 with a 75 ktlz pulse from the frequency divider 42 to generate a step-like voltage consisting of a voltage according to the amount of light received by each pixel of the light-receiving element 161, as shown in FIG. 3(b). Outputs a waveform video signal. The horizontal axis in FIG. 3(b) indicates the elapsed scanning time based on the 75 kHz pulse from the frequency divider 42, and the light receiving element in FIG. 3(a). 6
It also corresponds to the position of each pixel of 1.

The correspondence relationship between the pixels in FIG. 3(a) and the video signal is shown in an enlarged manner in FIG. That is, the pixel 16 where the light-receiving element 161 is formed
When the optical image 91 of the alignment mark 9 is to the right of the two-dot chain line shown above 1a, the output voltage shown by each pixel of the light receiving element 161 is a signal waveform corresponding to the light amount of the optical image 91 shown by the solid line. It exhibits S1. In other words, for pixels arranged on the right side of the pixel 161a, the amount of light irradiated is greater as the pixels are closer to the right (in Fig. 3, closer to the center of the alignment mark), so the voltage value is higher for the pixels closer to the right. On the other hand, on the left side of pixel 161a, since no pixel is irradiated with light, there is no output voltage, and furthermore, pixel +
In 61a, only a part of the pixel is irradiated with light, so
It shows the voltage value according to the irradiated area.

The video signal Sl output from the image sensor 16 is smoothed by the low-pass filter 43, resulting in a smooth signal waveform S2 shown by the dashed line in FIG. However, in this case, the horizontal axis in FIG. 5 represents the elapsed scanning time. The signal S2 is supplied to the inverting terminal of the comparator 441 of the threshold processing circuit 44, and a predetermined threshold voltage VTIIX is supplied to the non-inverting terminal of the comparator 4. Therefore, only when the signal S2 is equal to or lower than the threshold voltage Vn+x, the comparator 441 outputs a high-level voltage, and the high-level voltage causes the switching transistor 442 to become non-conductive, and the switching transistor 442, which is the output terminal of the threshold processing circuit 44, The emitter terminal of 442 is at a high potential.

The output signal S3 of the threshold processing circuit 44 with respect to the signal S2 and the threshold voltage VTIIX is shown as shown in FIG. 6(b). By appropriately setting the threshold voltage VTIIX, the rising and falling positions of the output signal S3 can be adjusted to the optical fiJI91 of the alignment mark 9 in FIG.
Both ends α2. They can be made to correspond accurately to Q3.

The signal S3 having a rectangular waveform is a D-FF with up-edge operation.
45, and the constant voltage Vcc as shown in Figure 6(d) is input to the D terminal, so
) When the rectangular signal S3 rises as shown in FIG. 6(e), the output S4 of the Q terminal falls. Although a detailed explanation will be omitted, the configuration is such that a rising pulse is input to the Cp terminal when the start pulse generator 40 generates the start pulse, and at that time, the D terminal input voltage is temporarily maintained at zero level. As a result, after the start pulse is generated, the Q terminal output maintains a high level until the rectangular signal S3 rises. This also applies to the DFF 47.

Since the signal S4 is input to the AND circuit 48 and a 600 kHz clock pulse [FIG. 6(f)] is also input from the clock pulse generator 41, the AND circuit 4
8 becomes a pulse train S5 as shown in FIG. 6(g). The left end in FIG. 6 is the start pulse generator 40.
A start pulse is generated from
6 indicates the point in time when scanning started. Although detailed explanation is omitted, the number of pulses in the pulse train S5 is counted by the pulse counter 50, which is cleared when the start pulse is generated.
The count number na is output from the counter 50 to the arithmetic unit 52.

The size (bit interval) of one pixel of the light receiving element +61 of the image sensor 16 is p (for example, 28 μm), and the light receiving element +6
The frequency of the pulse signal for signal transfer between each pixel in FIG. 1 is ft (for example, 75 kHz), and the frequency of the clock pulse shown in FIG.
Iz), the scanning movement pressure per clock pulse m
m is m = p (f s/ f 2) (e.g. 3.5
μm). Therefore, the product m·na of the output value na of the pulse counter 50 and the m is calculated from the left position f4 Q 1 of the signal transfer start bit of the light receiving element 161 of the image sensor 16 to the alignment mark shown in FIG. 3(a). 9 represents the distance to the right end position Q2 of the optical image 91 of No. 9. Furthermore, since the position of the incoming direction image sensor IG is fixed, the position 111τQ1 can be specified as the coordinates, and therefore the coordinates of the position Q2 can be specified by the above equation.

Furthermore, the rectangular signal S3 in FIG. 6(b) passes through the inverter 46 and becomes a signal S6 as shown in FIG.
It is supplied to the CP terminal of F47.

Since the constant voltage Vcc shown in FIG. 7(b) is supplied to the D terminal of the D-FF 47, the signal shown in FIG. 7(C) is output to the Q terminal of the D-FF 47. AND
Since the signal S7 and the clock pulse from the clock pulse generator 41 shown in FIG. 7(d) are input to the circuit 49, the output of the AND circuit 49 is as shown in FIG. 6(e). This becomes a pulse train S8. The left end in FIG. 7 also shows the point in time when a start pulse is generated from the start pulse generator 40 and the X-direction image sensor 16 starts scanning. The number of pulses in the pulse train S8 is omitted from detailed explanation, but is counted by a pulse counter 51 which is cleared when the start pulse is generated, and the count number nb is outputted from the counter 51 to the arithmetic unit 52.

The product m-nb of the count number nb and the scanning movement distance m per l clock pulse is the distance between the position MQ1 and the left end position Q3 of the optical image 91 of the alignment mark 9, as shown in FIG. 3(a). It is also possible to specify the coordinates of position Q3 in the same way as described above.

Furthermore, if the arithmetic unit 52 is used, the input count value n
a, nb to position Q2. shown in FIG. Since the distance between the midpoint of Q3 and the position Qx, that is, the position Q1 as described above, can be specified, the value Mx, which is the midpoint coordinate, can be calculated using the following equation.

It becomes possible to move the reticle 2 in the incoming direction using the position fine adjustment device 2 in accordance with the determined midpoint coordinates and place the reticle 2 at a desired position in the X direction.

Above, the action based on the X-direction image sensor 16 has been explained, but the action based on the X-direction image sensor 18 is also
In other words, the center point coordinates in the X direction are determined in the arithmetic unit 52, and the reticle 2 is placed at a desired position in the X direction.
allows you to adjust the movement.

Note that the optical image 9 of the alignment mark 9 shown in FIG.
1, the cross-shaped wide part (Q2-Q3) is the light receiving element! 61, but of course it is also possible to detect a narrow portion.

Furthermore, according to the device of the present invention, the count values na, nb
The ite arithmetic unit 52 is located at position 92. Distance between Q3 11
iWx is calculated by the following formula, Wx=m (na-nb
) =-(2) It is possible to determine iJ (=J method) of the alignment mark 9 in the X direction from this determined distance Wx. In exactly the same way, the width (dimension) in the X direction can also be determined based on the output of the X direction image sensor 18.

Furthermore, position α2. Since the distance between Q3 can be detected with high accuracy, it is possible to obtain, for example, the two width dimensions of a diffraction grating or mask pattern, and furthermore, the pitch of a grating pattern.

As described above, in the apparatus of the present invention, for example, a one-dimensional solid-state image sensor with 1024 bits and a bit interval of 28 μm is used, and the resolution of the conventional technology is 0.000 μm while securing a field of view of 100-odd μm with the magnifying optical system. .14 μm, whereas
If the frequency division ratio of the frequency divider 42 is, for example, 8, the resolution of the apparatus according to the embodiment of the present invention is 0.0175 μm (=0.14 μm
/8).

(Effects of the Invention) As detailed above, according to the present invention, an optical signal from a figure to be measured is imaged on a solid-state image sensor by an optical means,
In the figure reading method and apparatus for converting the imaged optical signal into an electric signal and reading the characteristics of the figure to be measured from the electric signal, the optical signal on the solid-state image sensor is scanned by a smoothing circuit. The smoothed waveform output from the smoothing circuit is shaped by a predetermined threshold value by a threshold processing means to generate a rectangular signal, and the rectangular signal is Detecting the rising and falling points of the rectangular signal, generating a clock pulse of a constant period by the clock pulse generating means, and counting the clock pulses generated during the period up to the detected rising and falling points of the rectangular signal. Since the electric signal value representing the characteristics of the ratio measurement figure is generated based on the count value output from the counting means, the existing optical system and solid-state Capable of reading graphics with high resolution and precision that exceeds the resolution of the image sensor. Furthermore, by calculating the electrical signal values in a processing unit, values such as the width, pitch, and coordinates of the characteristic quantities of the figure to be measured can be obtained, which can be used for positioning correction and shape reading in exposure equipment equipped with a figure reading device. Shape inspection and shape reading using equipment etc. can be performed with high precision.

[Brief explanation of the drawing]

FIG. 1 is a schematic diagram showing a step-and-repeat device including a pattern reading device to which the present invention is applied, FIG. 2 is a schematic diagram showing the pattern reading device shown in FIG. 1, and FIG.
An explanatory diagram showing the positional relationship between the alignment mark and the light receiving element in the figure and the voltage value detected for each pixel of the light receiving element, FIG. 4 is an electric circuit diagram forming a part of the figure reading device shown in FIG. 2, Figure 5 is an enlarged view of a part of Figure 3, Figure 6
7 and 7 are timing charts of signals in each part of the electric circuit shown in FIG. 4. 1... Graphic reading device, 9... Figure to be measured (alignment), 16. 18.161...Solid-state image sensor (
image sensor, light receiving element), 41... clock pulse generation means, 43... smoothing circuit (low pass filter)
, 44... Threshold processing means, 45.47... Counting area command means (D-FF), 48, 49, 50.51
... Counting means (AND circuit, pulse counter), 52
...Arithmetic processing unit.

Claims (1)

[Scope of Claims] 1. An optical signal from a figure to be measured is imaged on a solid-state image sensor by an optical means, the imaged optical signal is converted into an electrical signal, and the image of the figure to be measured is converted from the electrical signal. In the figure reading method for reading the characteristics, an electrical signal waveform obtained by scanning an optical signal on the solid-state image sensor is smoothed, and the smoothed electrical signal waveform is shaped by a predetermined threshold value. generating a rectangular signal, detecting the rising and falling points of the rectangular signal, counting clock pulses generated at a constant cycle during the period up to the detected rising and falling points of the rectangular signal; A figure reading method, characterized in that an electric signal value representing a feature of the ratio measurement figure is created based on the counted value. 2. Image reading that images an optical signal from a figure to be measured on a solid-state image sensor using optical means, converts the imaged optical signal into an electrical signal, and reads the characteristics of the figure to be measured from the electrical signal. The apparatus includes a smoothing circuit that smoothes an electrical signal waveform obtained by scanning an optical signal on the solid-state image sensor, and a smoothed waveform output from the smoothing circuit that is shaped by a predetermined threshold value into a rectangular shape. a threshold processing means for generating a signal, a counting area specifying means for detecting the rising and falling points of the rectangular signal, a clock pulse generating means for generating a clock pulse of a constant period, and the detection of the rectangular signal. and counting means for counting the clock pulses generated during the period from the rising edge of the clock pulse to the falling edge of the clock pulse. 3. The figure reading device according to claim 2, further comprising an arithmetic processing unit that calculates a value representing a feature of the figure to be measured based on the count output from the counting means.