JPS5835407A - Signal processor for checking pattern defect - Google Patents

Abstract

PURPOSE:To increase the accuracy in pattern measurement, by comparing a compensation measuring signal, which is obtained by subtracting a signal extracted in a low frequency filter from the defect data signal of the pattern whose phase is adjusted by a space filtering method, with a threshold signal. CONSTITUTION:The laser light signal of the defect of the regular pattern by the space filtering method is transduced into an electric signal by a photoelectric transducer 501 and branched through an amplifier 502. One output is inputted into a comparator 505 through the buffer amplifier 503 and a phase adjusting circuit 504, wherein the phase is adjusted. The other output is inputted into the comparator 505 through the low pass filter 506 and an adding circuit 507, and subtraction for both signals is performed. The difference signal is compared with the threshold signal which is inputted from a reference voltage circuit 508 through the adding circuit 507, and the defect signal 509 is outputted. In this way, the causes of the degradation of the measuring accuracy such as fluctuation in the photoelectric transducer 501 and secular change are alleviated and the accuracy in the pattern measurement is improved.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to defect inspection #C of industrial products having regular patterns such as shadow masks. In order to improve accuracy, this invention relates to a signal processing device that uses a series of electronic circuits to reduce factors that reduce measurement accuracy, such as the physical impact of an optical system, deviations in the shape of an object to be measured, and variations in characteristics of a photoelectric conversion system.

This circuit is a means for improving measurement accuracy in an eccentric signal processing system #C that performs pattern measurement and inspection using an optical measurement system, which is an application field of optical information processing. Concerning the composition.

The main causes of decreases in measurement accuracy in optical meters at 11m1 are physical changes in the optical system that includes the object to be measured, such as optical axis misalignment and vibration, deviations in the shape of the object, and physical changes in the optical system that includes the object to be measured. Changes over time in the characteristics of the opto-electrical conversion emplacement used when converting the obtained optical signal into an electrical signal: 1. Fluctuations in the characteristics due to the influence of ambient temperature Because the ingredients etc. are considered.

Therefore, the present invention is directed to a series of electronic circuits that receive light through an optical measurement system, convert the optical signal into an electrical signal, and then reduce the aforementioned factors that cause a decrease in measurement accuracy. The purpose of this study is to obtain a device that improves the accuracy of pattern measurement.

That is, pattern measurement using an optical measurement system.

In particular, when determining the approximate shape of an aperture in a regular pattern, a laser beam diffraction Burnn space-station wave number filtering method using a laser beam with temporal and spatial herherency is often used.

The present invention improves the recognition rate when detecting the presence or absence and position of defects in a pattern consisting of a regular array of unit apertures from bright spots generated between inverse Fourier transforms using a spatial filtering method (for pattern defect inspection). ) Its purpose is to provide a reliable IG device.

First, the basic configuration of the spatial filtering method is explained in 111J.
Jζ] Let me explain.

Laser oscillator l outputs nine laser beams 2, H remeter 3
The object to be inspected 5 is irradiated with the expanded parallel light 4. The object to be inspected 5 is placed at the front focal point of the 7-layer conversion lens 7, and the 7-layer conversion lens 7 of the object to be inspected is placed on the back focal plane by the light 6 that is diffracted when passing through the object to be inspected. A transformed spectrum appears. Therefore, on the back focal plane of the Fourier transform lens 7, the normal pattern (7
- A negative photographic film in which the intensity distribution of the Fourier transform spectrum (by the Fourier transform lens 7) has been recorded, and an empty frequency filter a are placed. Only the spectrum that is compatible with the defect pattern is absorbed by the spatial frequency a filter 8, and the spectrum that is compatible with the defect pattern is transmitted.

By the way, this spatial frequency filter 8 is rounded and provided on the front focal plane lζ of the inverse 7-Lier transform lens lO, and the light 9 transmitted through the spatial frequency filter 8 is inversely Fourier transformed by the inverse 7-Lier transform lens 10. A spatially filtered inverse 7-Lier transform image, that is, an image of only the defective portion of the inspected object 6, is placed on the back focal point wi (inverse 7-9 transform plane) of the inverse 7-Lier transform lens and is placed on the nine screen H. It becomes strange.

That's all for spatial filtering! This is the basic structure of the method.

On the other hand, Figure 2 shows an example of a pattern consisting of a regular arrangement of rectangular units (orifices). is defined as However, in the defect inspection device using the spatial filtering method described in Section 1.
Only the defective portions Di to D4 of 2) appear on the re-diffraction image as points of brightness corresponding to the defect size. Note that the aml (1) shows a normal pattern.

The image of only the defective portion of the inverse image of the pattern to be inspected 5 appearing on the inverse Fourier transform plane is detected by the photodetector array U as shown in the fourth panel. (transferred parallel to the optical axis at right angles to the transition layer),
There is no change in the Fourier transform pattern klI on the am point plane of the Fourier transform lens 7.

Therefore, when the pattern to be inspected 5 is translated in parallel, only the spectrum that resonates with the normal pattern among the 7-lier transform spectra of the pattern to be inspected is transferred to the spatial filter 8.
and the spectrum where the defect pattern <am is transmitted. Thus, due to the parallel movement of the pattern to be inspected 5, the inverse 7-Lier transformed image moves in the opposite direction of optical axis symmetry.
> An image of only the defective portion of the inverse image of the pattern to be inspected 5 moves in the direction opposite to the movement direction of the pattern to be inspected 5, sandwiching the optical axis ν.

In FIG. 1, if the direction of movement of the pattern to be inspected 5 is perpendicular to the plane of the paper, and the direction in which the seven detector arrays 11 on the inverse Fourier transform plane are arranged is parallel to the plane of the paper, then
The defect image moving through the inverse Fourier transform crosses the 7-degree detector array U at right angles and is detected as an output signal of the defect image. Also, as the pattern to be inspected 5 moves, its movement This defect inspection is carried out on devices that output distance information! ! By installing it in the device, it is possible to detect the defects in the pattern to be inspected and to know the positions of the nine defects. Assuming that the X direction is the direction perpendicular to the movement direction of the pattern to be inspected, and the Y direction is parallel to the direction of movement of the pattern to be inspected, the defect position component in the X direction is the defect position component on the inverse Fourier transform surface. Ryuzo-jXIRgJ
Therefore, it can be detected by the position of the unit photodetector in the direction of the 7-axis photodetector array.It can be detected by the position of the unit photodetector in the direction of the 7-dimensional array, and that in the Y direction is determined by the position scale on the fixed base and main body of the pattern to be inspected. This allows for position displacement detection, eccentricity, etc. using . Or a first-class drive shaft partial rotary engine for moving the pattern to be inspected;
Detection is possible by means such as providing a - holder. In addition to these, since the pattern to be inspected moves at a constant speed, the defect position component in the Y direction can also be obtained by counting the time with the electric point set at the time of movement and opening.

The photodetector array is a rectangular array of 7 units with 11 holes, and the output signal obtained by the behavior of the pattern to be inspected corresponds to the maximum allowable accumulation of defects. It has a threshold; it is binarized (by a comparator), and defect detection is performed.

As described above, a signal corresponding to partial defects in the pattern to be inspected can be obtained. However, in an optical measurement system, even a normal aperture may have a minute bright spot on the aperture side due to causes such as misalignment of the optical axis, vibration, and minute errors in the shape of the aperture side.

In addition, the light receiving section is affected by fluctuations in sensitivity over time of the light-to-electrical change element 1, such as a photomultiplier tube in a phototube, and noise components caused by the output signal Ef thereof.

When using a photodetector aperture array as shown in FIG. 4, in which the conditions for determining the tolerance range for rounding and defects are determined by the bright spot on both sides of the aperture side, the threshold value and the 11 *It is possible to detect true pattern defects by comparing the signals, but the fixed signal may include error components due to various factors as described above, fluctuation components, noise components, rounding, etc. High measurement accuracy could not be expected.

One aspect of the present invention is to use a signal control circuit as shown in FIG. 5, which is an embodiment of the present invention, in order to reduce the factors that cause the measurement accuracy to deteriorate as described above.

The basic configuration shown below will be explained below.

The amount of light received through the optical system of the light-to-electricity converter 501 (laser light 511 detected by the photodetector) K
A corresponding electrical signal 512 is issued. Amplifier 502 roaring light → electrical converter signal (51) Adjust the level appropriately. Signal (51
3) When amplified to the level, it becomes a collection.

It plays a role as a buffer for the optical to electrical converter 301 in the previous stage. The noise of relatively high frequency components contained in the optical-to-electrical converter 501 (frequency components sufficiently higher than the frequency of the signal to be detected) is used to limit the high-range medium power frequency of the amplifier 9502 to an appropriate level. It can be removed by

A signal 513 (hereinafter referred to as a "measurement signal") output from the amplifier 502 is divided into two systems. The low-pass filter 506 detects a sufficiently lower frequency component than the frequency of the ζ filter detection signal, which is caused by the shape deviation of the object to be measured.

; Since the input stage of the comparator 505 is a differential amplifier, the input signal 518 is input to the reference signal input side of the comparator S05 via the adder circuit 507, and the measurement signal 515 is input to the other input terminal of the comparator. By inputting , both signals are 515°! If the amplitude characteristics and phase characteristics of $18 are equal, it is possible to remove low frequency components caused by shape deviations of the object to be measured.

However, since the low frequency component signal to be removed is extracted using the low-pass filter 506, a phase lag occurs with respect to the measurement signal 515. For this reason, the comparator 50
Measurement signal 515 and low-pass filter 5 at the 5-power terminal
A phase difference occurs between the two signals of 518 that have passed through 06.

In order to avoid this problem, a phase adjustment circuit is inserted into the measurement signal 5x51H or the low-pass filter 506.In the present invention, since the phase delay caused by the 4 low-pass filters 508 is a problem, the low-pass filter When inserted into the measurement signal 506 side, the phase adjustment circuit 5 is set to the leading decimal phase, and when inserted into the measurement signal 515 side, the phase adjustment circuit 5 is set to the synchronized phase.
04 may be provided. However, in the latter case, it is necessary to adopt a phase adjustment circuit 504 that has characteristics that do not affect the low frequency band that the company wants to eliminate, but also the phase adjusted signal 515. be. In any case, it is desired that the phase gap 111111504 has a characteristic in which the amplitude characteristic does not change and only the phase characteristic can be observed.

The buffer amplifier 503 is rounded to provide a hysteresis characteristic to the comparator 505 due to discrimination problems.

This method prevents the signal component of one of the comparator input terminals (by feeding back a part of the comparator output signal] from being input to the other input terminal of the comparator via the low-pass filter side. It plays a role in improving the characteristics = Therefore, the buffer amplifier 503 can be fed back and inserted before the terminal on the other side. In this embodiment, the buffer amplifier 503 is placed before the phase adjustment circuit 504. Even if it is provided between the comparator 50s and the phase adjustment circuit 504, the above function can be achieved.

In addition, when the comparator 505 is manually operated, both signals 515
, 518 also need to match. For this purpose, the gain of the buffer amplifier 503 is adjusted, the adder circuit 507 is given a gain, and the resistance that determines the weighting of each terminal is adjusted. The low-pass filter 506 is implemented as an active filter and its gain is adjusted. Alternatively, it is possible to insert an attenuator into the higher level system and adjust the level.

The reference voltage generation circuit 508 generates a rounded voltage that determines a threshold value for defect control. The DC voltage level generated by the reference voltage generation circuit 508 is applied to the reference signal 518 of the comparator 505 via the addition circuit 507.

Thus, 514 is the output of the buffer amplifier 503.

516 is the output of the low-pass filter 506, 519 is the output of the comparator 505 which is a defective signal, and 509 is its output terminal.

ζ Here, and O are the actual values in the examples. Normal defect inspection performed by spatial filtering is performed from a mask pattern of several microns to several tens of microns to several hundreds of microns.
A wide range of .mu. metal mesh crystals is available, but here, as an example, a case where the unit slot of a 51 kl law pattern like the 411th one is 300 .mu.x soo is shown. The light source uses a Kutori-N@ laser with a low output level. The inverse Fourier transform image obtained through the optical measurement system is detected by a 1-width TOO deflector array 11) and sent to a photomultiplier tube (light→
Electrical conversion @Sol.

The light is received by an amplifier 102). Inspection of each slot portion of the layered beam is carried out by scanning the deflector array U and then capturing the inverse 7-layer transformed image with the photomultiplier tube. Scanning speed is lO ~ 300
In the range of 1 degree per second. At this time, the signal debris wL number range obtained in response to the defect is 200- to IK & 11 degrees. The frequency components generated are mainly below Hz. Therefore, the cutoff frequency of the low-pass filter sO6 is selected to be around 30Hs.

FIG. 6 is a schematic diagram of another embodiment of the present invention.

This embodiment shows a system in which the measurement signal is input to the comparator 505 after removing frequency components sufficiently lower than the signal to be detected. Therefore, the highest value 5 is received from the reference voltage generation circuit 508.
17 is applied to the reference voltage input side of the direct @* amplifier.

Note that 601 is a subtracter that outputs the output 6 of the 1lrIA path between 1 phase.
A low-pass filter output 516 is subtracted from 11 and a corrected measurement signal 612 is sent to the comparator 505.

FIG. 7 shows an example of signal waveforms of each part in the embodiment shown in FIG. 5 of the present invention. -Figure (a) is.

Output signal waveform 81 of light to electricity exchanger 501! , the same figure (6
) is the output signal waveform 513 . Same figure (
o) is a low pass filter! Output signal waveform 616 of S06.
Same figure (Xunha.

Between the output signal waveform 518 of the adder circuit 607 and the phase! Output waveform of 1 circuit 504! The signal waveform showing the difference from 15 and the output signal waveform 51 of the
It is G.

Thus, according to the present invention, when an industrial product having a standard 1 AWkA turn is inspected for defects, sufficient correction is made for deviations in the shape of the object to be measured and changes in the physical surface of the optical system. Characteristic changes over time of photomultiplier tubes
Characteristic fluctuations due to the influence of the lS ambient temperature and the adverse effects of noise etc. are also reduced. Due to these correction effects, p,
Measurement accuracy can be further improved compared to conventional test mains.

[Brief explanation of the drawing]

15th! ! l is the basic configuration diagram of the spatial filtering method,
Figure 2 is a diagram showing an example of a pattern consisting of a regular array of rectangular unit openings, and Figure 3 (6) shows defects in the pattern.
Figure 4E is an explanatory diagram of the image of the defective part of the inverse image of the pattern to be inspected that appears on the inverse Fourier transform plane, Figure 5 is a block diagram of an embodiment of the present invention, and Figure 6 is a diagram showing the present invention. FIG. 7, a block diagram of another embodiment, is a diagram showing signal waveforms of the embodiment @1. 1... vf oscillator, 2... laser beam, 3...
・Collimator, 4...Parallel light, 5...Object to be inspected, 6
...Diffraction multiplication table, 7...7-Lie transformation lens, 8.
...Spatial frequency filter, 9...Light transmitted through spatial frequency filter 8, 10...Inverse 7-Rie interchangeable lens, 1
1... Photodetector array (screen), 1
2... Optical axis, 13... Unit photodetector, 14
... Slot, 501 ... Optical to electrical converter, 502
...Amplifier, 5σ3...Buffer amplifier, 504.
...Phase adjustment circuit, 505...Comparator. 506...Low pass filter, 507...Addition circuit, 5
08... Reference voltage generation circuit, 509... Defect signal 5
19 output end. 511... Laser light detected by photodetector array 11, 513... Measurement signal, 515, 612
. . . Corrected measurement signal, 517 . . . Threshold voltage signal for defect control, 601 . . . Subtractor.

Claims (1)

[Claims] An optical-to-electrical converter that receives the pattern defect information signal with light using a laser and converts it into an electrical signal, and a rounding height that amplifies the electrical signal and removes noise in frequency components that are sufficiently higher than the frequency of the detection pane signal. The amount of amplification that limits the power frequency to an appropriate level, the low-pass filter that extracts the low frequency component caused by the shape deviation of the running object from the measurement signal of the output of this amplifier, and the extracted component in the low-pass filter R- a phase adjustment circuit for compensating for a phase delay of a signal to be measured; a reference voltage generation circuit for generating a threshold voltage signal for defect control; and a reference voltage generation circuit for generating a threshold voltage signal for defect control; A signal layer device for pattern defect inspection, comprising: a comparator for comparing a corrected constant signal obtained by subtracting the corrected signal with the threshold voltage signal;