JP3920382B2 - Method and apparatus for detecting defects or foreign matter - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a method and apparatus for detecting a defect or foreign matter existing on a substrate such as a semiconductor wafer, a printed circuit board, a TFT liquid crystal display device, or a magnetic disk substrate at high speed and with high accuracy.
[0002]
[Prior art]
In order to ensure the yield of substrates such as semiconductor wafers, printed substrates, TFT liquid crystal display devices, and magnetic disk substrates that are becoming finer or more complex year by year, defects or foreign substances existing on the substrate are detected at high speed and with high accuracy. A device to do is indispensable.
A regular or irregular pattern such as a fine circuit pattern or an information recording groove is generally formed on a substrate such as a semiconductor wafer, a printed circuit board, a TFT liquid crystal display device, or a magnetic disk substrate. Therefore, how to separate and extract defects or foreign substances existing on the substrate from these patterns has been an important development issue.
[0003]
Conventionally, in order to solve this problem, various methods have been devised and put into practical use. For example, in Japanese Patent Application Laid-Open No. 55-149829, in order to separate a pattern on a substrate and foreign matter, a linearly polarized laser is irradiated to the substrate at a shallow angle, and scattered light generated from the pattern and foreign matter is directed upward of the substrate. A method of extracting only a scattered light component generated from a foreign substance by a sensor using an analyzer after being condensed by a provided detection optical system is shown.
[0004]
Japanese Patent Application Laid-Open No. 59-65428 discloses that a linearly polarized laser beam is irradiated to a substrate at a shallow angle in order to separate a pattern on a substrate from a foreign material, and scattered light generated from the pattern and the foreign material is directed above the substrate. After condensing with the provided detection optical system, a spatial filter is provided on the Fourier transform surface of the substrate in the detection optical system to shield a specific diffracted light component having a regular pattern forming angle or arrangement, and scattering generated from foreign matter A method for extracting only light components with a sensor is shown.
[0005]
In the above known example, an attempt is made to extract scattered light from a defect or a foreign substance by focusing on scattered light disturbance and diffraction due to the regularity of the geometric shape on the substrate. Accordingly, when trying to detect finer defects or foreign matters, it has become impossible to achieve sufficient performance. That is, in order to detect finer defects or foreign matters, the sensitivity of the sensor itself for detecting scattered light is increased, the detection pixel size on the substrate is reduced, or the output of the laser beam to be irradiated is increased. Countermeasures have become necessary.
[0006]
As a sensor that can realize a small pixel size with high sensitivity, a linear sensor such as a CCD (Charge Coupled Device) linear sensor, or a TDI (Time Delayed Integration) sensor that can be said to be a modification thereof, is used. As a known example using a linear sensor, Japanese Patent Laid-Open No. 1-158308 is known.
[0007]
However, since the detection dynamic range of a linear sensor is usually narrower than that of a sensor such as a photomultiplier tube, for example, if the sensitivity is increased so that a minute defect or a foreign substance existing on a substrate can be detected, Recognizing large defects or foreign objects by strong scattered light caused by large defects or foreign substances that may exist on the substrate at the same time, a so-called “blooming phenomenon” occurs in which the vicinity of the defects or foreign objects falls into a saturated state. It was difficult to identify the size or position in terms of particle size and standard particles.
[0008]
In order to avoid this, the “anti-blooming circuit” that reduces the detection sensitivity and suppresses the occurrence of the “blooming phenomenon”, or automatically releases the surplus charge when the linear sensor reaches the saturation state. Measures such as introduction of were taken.
[0009]
However, the former, which lowers the detection sensitivity, is a method that goes against high integration of semiconductors. In the latter method, defects or foreign matters that generate strong scattered light that would otherwise exceed the dynamic range of the linear sensor. However, there is a problem that the size of the required position cannot be identified.
[0010]
Further, in detecting foreign matter and defects, there is a case in which a method of inspecting by comparing the same pattern is used with attention paid to the repeatability of the pattern on the substrate. In this case, a subtle difference between the two images to be compared, that is, the reference image and the inspection image affects the detection sensitivity, and a determination threshold value that allows a difference in brightness level due to the difference between the normal inspection image and the reference image is provided. When a difference equal to or greater than the threshold value occurs in the two images, it is determined that there is a foreign object or a defect. In general, a portion having a complicated pattern shape has a large difference between the two images, and a threshold value is often determined based on the large difference. Therefore, even in an area where the difference between the two images is small, a threshold value more than necessary is set, and sufficient detection sensitivity may not be obtained.
[0011]
As a means for solving this, there is a method described in Japanese Patent Application Laid-Open No. 63-32666, but this method is intended for pattern imaging, and there is a problem that the circuit scale becomes large for scattered light detection.
[0012]
[Problems to be solved by the invention]
Normally, in detection of scattered light from a defect or foreign matter on a substrate using a linear sensor, the data after A / D (analog / digital) conversion of the sensor output is, as shown in FIG. A unit of 50 for one pixel is handled as a two-dimensional image extending in the feed direction X of the substrate and the sensor row direction Y of the linear sensor.
[0013]
In the case of a general fine defect or foreign matter, as shown at 51, a halftone output is obtained. The detected waveform of the linear sensor at this time is shown as in FIG. An A / D (analog / digital) conversion output is extracted as a Y-direction output with a pixel size of 50 as a unit. However, if there is a defect or a foreign object, scattered light is detected and the output voltage of the sensor increases. A defect / foreign object detection voltage Vpeak is obtained.
[0014]
Here, when the voltage at the saturation level 54 of the linear sensor is Vsat and the detection voltage (noise level) in the dark current state when there is no light input is Vnoise, the dynamic range DR of the detection system is given by Equation 1. It is physically impossible to obtain a resolution higher than the dynamic range.
[0015]
DR (dynamic range) = Vsat / Vnoise Equation 1
Therefore, if the laser output is increased to improve the detection sensitivity so that the detection voltage from a minute defect or foreign object becomes equal to or higher than the noise level Vnoise, the dynamic range is increased even if a relatively small defect or foreign object exists. As a result, the defect / foreign object detection voltage Vpeak reaches the saturation level voltage Vsat.
[0016]
When the saturation level voltage Vsat is reached, the electric charge photoelectrically converted by the linear sensor overflows and, as shown in FIG. 9, the so-called saturation level state is reached over a wide area like the periphery 53 of the original scattered light generation position 52. The “blooming phenomenon” occurs.
[0017]
FIG. 10 shows an example of the detection waveform of the linear sensor at this time. That is, as shown in 56, even if a large defect or foreign matter is present, the defect / foreign matter detection voltage Vdefect cannot exceed the saturation level voltage Vsat. It overflows and pushes up the level of surrounding pixels to the saturation level voltage Vsat over a wide area 58 such as 57a and 57b. This is a blooming phenomenon, and conventionally, it has been difficult to identify the size and position of such a defect or foreign matter.
[0018]
On the other hand, in the “anti-blooming circuit”, since this surplus charge is discharged through a line of another system, the surplus charge does not overflow to cause blooming. Therefore, the position of a relatively large defect or foreign object can be identified, but the output level is fixed at Vsat.
[0019]
By the way, the defect or foreign matter size is generally identified by the peak voltage value of the detected scattered light intensity V. As shown in FIG. 11, when a sensor having a wide dynamic range such as a photomultiplier tube is used, the conversion curve 60 is used from the peak voltage value to calculate the size of defects or foreign matters converted using standard particles or the like. Can be obtained. However, in the case of a sensor having a finite dynamic range such as a linear sensor, the peak voltage is fixed at Vsat regardless of the presence or absence of the “anti-blooming circuit”, so that the conversion curve reaches a peak like 61, and as a result The size of the relatively large defect or foreign material could not be identified.
[0020]
Conventionally, the position of the defect or foreign matter is identified from the pixel position where the peak voltage is generated. However, since the peak voltage is distributed over a wide range, it is difficult to identify the position of the defect or foreign matter.
[0021]
Accordingly, an object of the present invention is to detect a defect or foreign matter existing on a substrate such as a semiconductor wafer, a printed board, a TFT liquid crystal display device, or a magnetic disk substrate with high sensitivity. Provided is a method for detecting a defect or a foreign object that can be compatible with the occurrence of blooming of a linear sensor by identifying the size and position of a relatively large defect or foreign object when using a sensor, a TDI sensor, or the like, or an apparatus using the same. There is.
[0022]
Furthermore, an object of the present invention is to correct a level difference between two images to be compared in a detection method for comparing an inspection image and a reference image. It is an object of the present invention to provide an apparatus that prevents an unnecessary deterioration in detection sensitivity by always maintaining an optimal determination state by optimally setting according to the region.
[0023]
[Means for Solving the Problems]
In the present invention, paying attention to the fact that the blooming state that occurs when the “anti-blooming circuit” as shown in FIG. 10 is not used is an overflow of surplus charge, instead of the peak voltage V, The size of the defect or the foreign matter is to be identified based on the total, that is, the total charge amount in the saturation region or the blooming region 58, and the center of gravity position is determined by the pattern of the saturation region or the blooming region. It is intended to identify the position of a defect or a foreign object.
[0024]
That is, the present invention uses a linear sensor such as a CCD linear sensor or a TDI sensor in a state where the “anti-blooming circuit” is not operated, and when the linear sensor is saturated or blooming due to a relatively large defect or foreign matter, the saturation is detected. This is achieved by means for identifying the size of the defect or foreign matter from the total charge amount of the region and identifying the position of the defect or foreign matter from the pattern of the saturated region.
[0025]
The present invention achieves the following by adopting the above-described means.
[0026]
The amount of charge accumulated in each pixel is detected as an output voltage and treated as a gradation after A / D (analog / digital) conversion. Accordingly, the total charge amount E is converted as the sum of the number of gradations in the blooming region 53 in FIG. 9 (ie, the product of the saturation level gradation value and the saturation pixel number) (Equation 2).
[0027]
E (total charge amount) = Vsat (or saturation level gradation value) × saturation pixel number Formula 2
As a result, as shown in FIG. 12, a conversion curve 62 for the total charge amount E and the defect or foreign matter size converted into standard particles is obtained, and Vsat (= Esat: charge amount at saturation) or Vnoise (= Enoise: noise). Compared to the conventional dynamic range based on the component charge amount), it is possible to realize a region 63 having a large dynamic range.
[0028]
Further, in the present invention, in the two-chip comparison process 34 in FIG. 1, an index value such as an average brightness value is obtained from the brightness level in an arbitrary region of the inspection image 32a and the reference image 32b, and each index is obtained. A stable comparison inspection can be performed by estimating the brightness level difference between the two images from the difference in values and applying level adjustment to correct the difference to the inspection image or the reference image.
[0029]
In addition, the detection sensitivity can be prevented from being deteriorated more than necessary by changing the threshold value for comparison according to the level difference fluctuation of the pattern on the substrate. For example, in the case of LSI, the difference between images is relatively small in the cell portion, but the difference between images tends to be large in a complicated pattern portion such as a peripheral circuit portion. Therefore, by providing pattern information in advance and variably giving an optimum threshold value according to the pattern type, the detection sensitivity can be prevented from being deteriorated more than necessary. In addition, the signal level detected according to the pattern type often has a characteristic brightness. In this case, the determination threshold is changed based on the brightness level of the inspection image or the reference image. The detection sensitivity can be prevented from being deteriorated more than necessary.
[0030]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described in detail below.
[0031]
FIG. 6 schematically summarizes the configuration of the defect / foreign particle inspection apparatus on the substrate according to the present invention.
[0032]
A substrate (for example, a semiconductor wafer) 1 on which a circuit pattern 2 having a concavo-convex shape and defects or foreign matter 3 are present is mounted on a stage 4. Illumination light emitted from a laser light source 5 such as a semiconductor laser is converted into parallel light (collimated light) by a collimating lens 6 (required only in the case of a point light source laser such as a semiconductor laser). The light beam is enlarged and deformed, and further irradiated onto the surface of the substrate 1 through the mirror 9, the condensing lens 10, the mirror 11, and the like. Regarding the irradiation direction of the laser beam for illumination, there are variations such as direct irradiation and oblique irradiation. However, since the influence on the contents of the present invention is small, it is treated as oblique irradiation here.
[0033]
If there are irregularities such as defects, foreign matter 3 or circuit pattern 2 on the substrate, scattered light is generated. The scattered light is collected by, for example, an optical system such as the Fourier transform lens 12, the spatial filter 13, and the inverse Fourier transform lens 14 to form an image on the linear sensor 16 on the circuit board 15. The Fourier transform lens 12 and the inverse Fourier transform lens 14 are combined to form an imaging optical system having a constant magnification. And the detection direction also has variations such as direct detection and oblique detection as in the illumination system.
[0034]
The spatial filter 13 blocks scattered light from a highly repetitive pattern such as a memory cell of a semiconductor memory or a pattern directed in a specific direction out of scattered light from a circuit pattern on the substrate 1, thereby preventing defects or foreign matter from This is for extracting scattered light. There are a method in which a spatial filter is placed at the Fourier transform position of the imaging optical system, and a method in which an analyzer is used instead of the spatial filter, as disclosed in Japanese Patent Application Laid-Open No. 55-149829.
[0035]
As the linear sensor 16, a CCD linear sensor, a TDI sensor or the like is used, and the output of the linear sensor is processed by the defect / foreign matter detection circuit 17 and the control computer 18 to obtain the position and size of the defect or foreign matter, and if necessary. The result is displayed, stored, or transferred to the host computer. The control computer 18 also controls the entire inspection apparatus including the stage 4.
[0036]
On the substrate 1, as shown in FIG. 7, a circuit pattern 21 that is repeated for each chip or shot (a plurality of chips or shots are formed on a semiconductor wafer) may be formed. Since the visual field 20 of the linear sensor 16 is smaller than the size of the substrate 1, in order to inspect the entire surface of the substrate, the linear sensor's visual field 20 is relatively scanned as 22 in the XY plane. It is necessary to drive the stage on which 1 is mounted. In the middle of this scanning, the defects existing on the substrate and the scattered light from the foreign material 3 are detected by the linear sensor.
[0037]
FIG. 1 shows details of the first embodiment of the defect / foreign matter detection circuit 17 and the control computer 18 shown in FIG. As shown in FIG. 7, this embodiment is an effective method for inspecting a patterned substrate having a repetition for each chip or shot.
[0038]
The defect / foreign substance detection circuit 17 includes an A / D conversion circuit 30, a delay memory 31, a chip comparison circuit 41, a saturation region processing circuit 42, and a masking process 38. The control computer 18 includes a defect candidate memory 39 and a processing unit (hardware or soft wafer) 40 for calculating the total charge amount and obtaining the size and position of the defect or foreign matter. Further, by providing data display means or storage means using the output of the processing unit 40 as an input signal, the size or position of the detected defect or foreign matter can be displayed or stored.
[0039]
The electric charge due to the scattered light detected by the linear sensor 16 is converted into digital information (gradation value) by the A / D conversion circuit 30, and the information one chip before stored in the inspection image 32 a and the delay memory 31. Is the reference image 32b. In the inspection image 32a, scattered light from the defect or foreign matter 3 and scattered light from the circuit pattern 2 are detected.
[0040]
The inspection image 32a and the reference image 32b are sent to the 2-chip comparison circuit 41 and the saturation region processing circuit 42, respectively. In the two-chip comparison circuit 41, a non-coincidence image 35 is obtained by extracting a non-coincidence portion between the two images by the two-chip comparison process 34 according to the procedure shown in FIGS.
[0041]
FIG. 2 shows an example of a detected waveform of the reference image 32b. There are few defects or foreign objects in this image. That is, the height change of the detection waveform 25 of the reference image 32b is due to scattered light generated by the circuit pattern on the substrate. Since the circuit pattern has repeatability for each chip or shot, the level of the detection waveform 25 can be deleted as a fixed value. Therefore, a threshold waveform 27 having a level higher than that of the detection waveform 25 can be obtained.
[0042]
On the other hand, FIG. 3 shows an example of a detected waveform of the inspection image 32a. In this image, a local peak value 28 is generated in the detection waveform 26 due to scattered light due to defects or foreign matter. Therefore, by comparing the detected waveform 26 with the threshold waveform 27 obtained in FIG. 3, the peak 29 due to a defect or a foreign substance as shown in FIG. 4 can be extracted. For example, a mismatch image 35 is obtained.
[0043]
In the saturation region processing circuit 42, images 33a and 33b are obtained by binarization processing, and a common saturation region 36 is extracted from the two images. By performing an expansion process on this data, mask data 37 for the masking process 38 is created.
[0044]
From the mismatched image data 35 and the mask data 37, pixels that are candidates for defects or foreign matters are obtained, and charge values (gradation values) and position information are stored in the defect candidate memory 39 for all the pixels.
[0045]
During or after the inspection, the contents of the defect candidate memory 39 are read. If the saturation level has not been reached, the size and position of the defect or foreign matter are identified from the peak value (tone value) and the position. . On the other hand, when the saturation level is reached, as shown in FIG. 9, the total charge amount in all the saturated regions 53 that are in contact with each other is obtained, and a defect or a foreign substance actually exists from the center of gravity of the saturation region. Coordinates close to the location 52 are obtained and treated as the size and position of the defect or foreign matter.
[0046]
Next, FIG. 5 shows details of the second embodiment of the defect / foreign matter detection circuit 17 and the control computer 18 shown in FIG.
[0047]
The defect / foreign substance detection circuit 17 includes an A / D conversion circuit 30 and a comparator 44. The control computer 18 includes a defect candidate memory 39 and a processing unit (hardware or soft wafer) 40 for calculating the total charge amount and obtaining the size and position of the defect or foreign matter.
[0048]
The electric charge due to the scattered light detected by the linear sensor 16 is converted into digital information (gradation value) by the A / D conversion circuit 30, and compared with a preset threshold value 43 and a comparator 44. When the threshold value 43 is exceeded, charge values (gradation values) and position information are accumulated in the defect candidate memory 39 for all the pixels as defects or foreign matter candidates.
[0049]
During or after the inspection, the contents of the defect candidate memory 39 are read. If the saturation level has not been reached, the size and position of the defect or foreign matter are identified from the peak value (tone value) and the position. . On the other hand, when the saturation level is reached, as shown in FIG. 9, the total charge amount in all the saturated regions 53 that are in contact with each other is obtained, and a defect or a foreign substance actually exists from the center of gravity of the saturation region. Coordinates close to the location 52 are obtained and handled as the size and position of the defect or foreign matter.
[0050]
This second embodiment is also effective for inspection of a substrate having no pattern (for example, inspection of a mirror wafer).
[0051]
Further, the processing of the processing unit 40 from the defect candidate memory 39 is performed by performing a convolution integration process with a template of an arbitrary size in addition to the method for obtaining the detected center-of-gravity coordinates and obtaining the center of gravity of the processing result (that is, It is also possible to obtain the center of gravity by weighted average processing). In the case of the present embodiment, there is an effect that the position can be accurately calculated even when the detected image has some directionality such as due to illumination conditions.
[0052]
In FIG. 2 and FIG. 3, the case where the illustrated Vnoise is almost the same level has been described. However, the Vnoise (the voltage indicated by the solid line in FIG. 2) including the brightness of the background of the substrate is the same substrate. It can happen that it also varies from place to place. In this case, the threshold value 27 (the voltage indicated by the dotted line in FIG. 2) obtained from FIG. 2 is not an optimum value in FIG. 3, but a value obtained by correcting the difference between Vnoise in FIG. 2 and FIG. The value corrected to the threshold value obtained from the above should be applied to FIG. Therefore, although details will be described later, as shown in FIG. 13, the average value of the brightness level of the image in an arbitrary small region is obtained for the reference image and the inspection image, and the correction value is calculated from the difference between the average values of the two images. By finding and correcting the threshold value, the threshold value can always be kept optimal.
[0053]
In the basic circuit configuration of FIG. 13, the threshold value of the threshold generator 103 is added to the input reference image 32b and the added reference image is compared with the inspection image 32a by the comparator 104. The inconsistent image 35 is to be obtained. At this time, the inspection image input to the comparator 104 is an inspection image corrected by the difference between the average values of the brightness levels of the two images. In FIG. 13, the two-chip comparison circuit 41 increments the output value of the threshold generator 103 with respect to the reference image 32b by the adder 102, and detects the difference from the inspection image 32a by the comparator 104. . The corrector 105 obtains each pixel data by the latch means 107, calculates the sum by the adder 108, and calculates the average level of neighboring pixels for the reference image 32b by the divider 109. Similarly, the average level is calculated by the corrector 106 for the inspection image 32a. If the difference between the reference image and the inspection image is detected from the two average levels using the subtractor 110 and the inspection image 32a is corrected by the subtractor 111, the above function can be achieved. Although the average value has been described as being obtained in one dimension, it goes without saying that the same applies to any two-dimensional area.
[0054]
In FIG. 13, the output of the threshold generator 103 can be varied according to the difference in the area on the substrate. For example, as shown in FIG. 7, in an LSI, a chip area that forms a section of a semiconductor wafer is determined in advance, and a pattern type such as a memory cell part or a peripheral circuit part is determined in each chip. ing. Therefore, as shown in FIG. 14, the coordinate signal 112 may be input to the threshold generator 103, and the threshold generator 103 may output an optimum threshold signal using a lookup table (LUT) 113. For example, the look-up table may be set so that the threshold value is set high in the region of the peripheral circuit portion where the pattern is complicated and the detection signal changes in a complicated manner, and is set low in the memory cell portion. The coordinate signal 112 is related to a signal for moving the stage 4. Therefore, the threshold value is automatically changed to an optimum value as the stage 4 moves.
[0055]
Further, the threshold generator 103 of FIG. 13 can be set so as to correspond to the brightness level of the reference image as shown in FIG. In the optical system as shown in FIG. 6, in the region where the repeatability in the detection visual field is strong like the memory cell portion, the output light amount is reduced by the spatial filter 13, and therefore the background detection signal level, That is, the background detection signal level, that is, Vnoise is large in a region where Vnoise is small and the repeatability in the field of view is weak like the peripheral circuit portion. Therefore, the threshold value may be changed in response to Vnoise corresponding to the brightness level of the reference image.
[0056]
As another method, an inspection trial is performed prior to the inspection of actual foreign matters and defects in a plurality of chips, and a correlation distribution of signal levels at corresponding points in the chip of the reference image and the inspection image is obtained. Even if there is a difference in the brightness of both images within the range of the correlation distribution area, this is a difference that appears statistically on the trial inspection, and a value exceeding this correlation distribution is compared with the reference image in the inspection image. When it occurs, it can be determined that there is a foreign object or a defect.
[0057]
As a specific configuration, as shown in FIG. 16, the two-dimensional addition memory 114 correlates using the signal levels of the reference image 32b and the inspection image 32a as addresses. After performing this for an arbitrary image, the threshold value generation circuit 115 corresponds to each signal level of the reference image (over the entire area from the dark part to the bright part of the image), and the correlation distribution region. The threshold value of the LUT 113 is initialized based on an appropriate increment value exceeding. The correlation distribution in the bright area of the image and the dark area of the image tends to show that the distribution is broad in the bright area but not in the same pattern. Therefore, the detection sensitivity of dark areas is increased. In the actual inspection, the inspection can be performed in the same manner as the LUT 113 in FIG. According to the present embodiment, an optimum threshold value can always be obtained for foreign matters and defects, so that the detection sensitivity can be improved.
[0058]
FIG. 17 shows another specific embodiment of the two-chip comparison process 34 in the present invention. In FIG. 17, the two-chip comparison process 34 includes shift registers 212 a to 212 c, serial in parallel out shift registers 213 a to 213 c, LUT 214, subtraction circuits 216 a to 216 k, minimum value detection circuits 218 and 219, and a switch 220. The reference image 32b is delayed and shifted using shift registers 212a and 212b that delay one scanning line of the image and serial-in / parallel-out shift registers 213a to 213c, and has various delay amounts and shift amounts. Multiple reference image outputs can be obtained, and the outputs are synchronized with the inspection image 32a.
[0059]
Next, by subtracting the outputs of the shift registers 213a to 213c and the output of the shift register 213d by the subtraction circuits 216a to 216i, the reference image 32b is shifted by ± 1 pixel in the X and Y directions with respect to the inspection image 32a. A difference image from the obtained image is obtained. Here, the difference image is obtained by subtracting the reference image 32b from the inspection image 32a, and when the difference image is a negative value, the value 0 is output. On the other hand, as for the output of the shift register 213d, the difference from the LUT 214 in which a value is set in advance is obtained by the subtraction circuit 216k as in FIG. . If the value of 216k is negative, the output of 220 is zero. Here, 216j calculates a pixel difference that is predicted to be in phase with the inspection image and the reference image, and outputs a value of 0 if this difference is negative.
[0060]
The outputs of the subtraction circuits 216a to 216i are input to the minimum value detection circuit 218, and the minimum values of 216a to 216i are output. Here, as described above, the subtraction circuits 216a to 216i output a value of 0 when a negative value is obtained. The outputs of the minimum value detection circuit 218 and the switch 220 are input to the minimum value detection circuit 219, whichever is smaller.
[0061]
In the inspection of the two-chip comparison method, an unintended signal output is generated due to a slight positional deviation of an image or a pattern generation condition. In particular, the pattern edge portion may be erroneously output as a foreign matter or a defect. According to the method, since the erroneous determination can be reduced by shifting the image, there is an effect that a stable inspection can be performed.
[0062]
In the above description, the method and the apparatus are inspected by detecting scattered light from defects and foreign matters existing on the substrate. However, not only the detection of the scattered light but also the reflected light from the substrate is imaged. The fact that this can be inspected by detecting this is also a matter of course from the configurations and functions of FIGS. 13 to 16.
[0063]
【The invention's effect】
As described above, according to the present invention, when detecting defects or foreign matters existing on a substrate such as a semiconductor wafer, a printed circuit board, a TFT liquid crystal display device, or a magnetic disk substrate, a narrow dynamic range in which blooming or the like is likely to occur is obtained. Using a linear sensor, it is possible to identify the size and position of a relatively large defect or foreign object as if it were a sensor with a wide dynamic range, realizing a highly sensitive and highly accurate defect or foreign object inspection device. The manufacturing yield of the substrate manufacturing process can be improved.
[Brief description of the drawings]
FIG. 1 is a processing function block diagram showing a first embodiment of the present invention.
FIG. 2 shows an example of a detection waveform of a reference image in the first embodiment of the present invention.
FIG. 3 shows an example of a detection waveform of an inspection image in the first embodiment of the present invention.
FIG. 4 shows extraction of a peak due to a defect or foreign matter in the first embodiment of the present invention.
FIG. 5 is a processing function block diagram showing a second embodiment of the present invention.
FIG. 6 is a schematic view mainly showing a configuration of an optical system according to an embodiment of the present invention.
FIG. 7 is a diagram illustrating the relationship between the visual field of a substrate and a linear sensor.
FIG. 8 is a diagram illustrating an example of a normal detection waveform.
FIG. 9 is an enlarged view of a captured image showing a state when blooming occurs.
FIG. 10 is a diagram illustrating an example of a detected waveform when saturation or blooming occurs.
FIG. 11 is a diagram showing the relationship between the size of a conversion defect or foreign matter and the intensity of scattered light due to a peak voltage in a conventional method.
FIG. 12 is a diagram showing the relationship between the size of a conversion defect or foreign matter according to the present invention and the intensity of scattered light due to a peak voltage.
FIG. 13 is a diagram for explaining an embodiment of the two-chip comparison circuit of the present invention.
FIG. 14 is a diagram for explaining an embodiment of a two-chip comparison process of the present invention.
FIG. 15 is a diagram for explaining another embodiment of the two-chip comparison circuit of the present invention.
FIG. 16 is a diagram for explaining another embodiment of the two-chip comparison circuit of the present invention.
FIG. 17 is a diagram for explaining another embodiment of the two-chip comparison circuit of the present invention.
[Explanation of symbols]
1 Substrate
3 Foreign matter
5 Semiconductor laser oscillator
12 Fourier transform lens
13 Spatial filter
14 Inverse Fourier transform lens
16 Linear sensor
17 Defect / foreign matter detection circuit
18 Control computer
30 A / D conversion
31 Delay memory
32a Inspection image
32b Reference image
41 2-chip comparison circuit
42 Saturation region processing circuit
50 pixels
52 Blooming area (saturated area)
60-62 conversion curve
102 Adder
103 Threshold generator
104 comparator
108 Adder
109 Divider
110 Subtractor
113 Look-up table (LUT)

Claims (21)

Irradiate the substrate with laser light,
Collecting scattered light generated from the substrate ;
The focused scattered light is imaged on a linear sensor,
Said charge information according to the obtained scattered light by the linear sensor into digital information, to determine the presence or absence of a defect or foreign matter on the substrate based on the converted digital data,
By determining the size or position data of the defect or foreign object when the defect or foreign object is present,
The detection process for detecting defects or foreign matter on the substrate,
In the saturated state or blooming state of some pixels of the linear sensor that occurs when scattered light generated from a relatively large defect or foreign matter on the substrate is imaged onto the linear sensor, a linear sensor output is saturated The size of the defect or foreign matter is identified from the total charge amount caused by the defect or foreign matter obtained by the product of the number of pixels in the state and the digital information value to be saturated, and the saturated region pattern caused by the defect or foreign matter is identified. A method for detecting a defect or foreign matter, wherein the position of the defect or foreign matter is identified from the center of gravity . Means for irradiating the substrate with laser light;
Means for collecting the scattered light generated from the substrate by irradiating the laser beam,
And emphasizing means than the scattered light generated from a normal circuit pattern scattered light from the defect or foreign matter on the substrate,
Means for imaging the scattered light the focused on the linear sensor,
And the linear sensor for changing the scattered light into charge information,
Means for converting the charge information by the obtained scattered light by the linear sensor into digital information,
Means for determining the presence or absence of a defect or foreign matter on the basis of the converted digital data,
And means for obtaining the data of the size or location of the defect or foreign matter when the defect or foreign matter is present,
A detecting device for detecting a defect or foreign matter on the substrate,
In the saturated state or blooming state of some pixels of the linear sensor that occurs when scattered light generated from a relatively large defect or foreign matter on the substrate is imaged onto the linear sensor, a linear sensor output is saturated The size of the defect or foreign matter is identified from the total charge amount caused by the defect or foreign matter obtained by the product of the number of pixels in the state and the digital information value to be saturated, and the saturated region pattern caused by the defect or foreign matter is identified. A device for detecting a defect or foreign matter, characterized by comprising means for identifying the position of the defect or foreign matter from the center of gravity . In claim 2,
An apparatus for detecting a defect or foreign matter, wherein the linear sensor that converts the scattered light into charge information is a charge coupled device. In claim 2,
An apparatus for detecting a defect or foreign matter, wherein the linear sensor that converts the scattered light into charge information is a TDI sensor. In claim 2,
The means for determining the presence or absence of defects or foreign matter based on the converted digital information obtains a reference image which is an inspection image and an image free of defects or foreign matter based on the converted digital information,
Further, a mask image and a mismatch image are obtained from the inspection image and the reference image,
An apparatus for detecting a defect or foreign matter, characterized in that the non-matching image is masked with the mask image. In claim 2,
An apparatus for detecting a defect or foreign matter, wherein the means for determining the presence or absence of a defect or foreign matter based on the converted digital information is to compare the converted digital information with a set value. In claim 1,
A method of detecting a defect or foreign matter, comprising displaying or storing data on the size or position of the detected defect or foreign matter. In claim 2,
An apparatus for detecting a defect or foreign matter, comprising means for displaying or storing data on the size or position of the detected defect or foreign matter. In claim 1,
A method of detecting a defect or foreign matter, wherein the position of the defect or foreign matter is obtained by obtaining an appropriate weighted average of the pattern of the saturated region caused by the defect or foreign matter. In claim 2,
The means for obtaining the position of the defect or foreign matter from the pattern of the saturated region caused by the defect or foreign matter is to obtain an appropriate weighted average of the pattern of the saturated region caused by the defect or foreign matter, or Foreign matter detection device. Means for irradiating the substrate with laser light or focused light; means for collecting scattered light generated from the substrate by the irradiation means;
And emphasizing means than the scattered light generated from a normal circuit pattern scattered light from the defect or foreign matter on the substrate,
Means for imaging the collected scattered light on a detector;
A detector that converts the scattered light into charge information;
Means for converting charge information by scattered light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means obtains a first index value indicating the brightness of the entire reference image from the brightness in an arbitrary area of the reference image, and the inspection image from the brightness in an arbitrary area of the inspection image. Means for obtaining a second index value indicating overall brightness; means for obtaining a predetermined correction coefficient from the difference between the first and second index values; and brightness of an inspection image or reference image using the correction coefficient. Means for correcting the height level, and means for adding a threshold value to the reference image or inspection image,
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter . Means for irradiating the substrate with laser light or focused light; means for collecting scattered light generated from the substrate by the irradiation means;
And emphasizing means than the scattered light generated from a normal circuit pattern scattered light from the defect or foreign matter on the substrate,
Means for imaging the collected scattered light on a detector;
A detector that converts the scattered light into charge information;
Means for converting charge information by scattered light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means obtains a first index value indicating the brightness of the entire reference image from the brightness in an arbitrary area of the reference image, and the entire inspection image from the brightness in the arbitrary area of the inspection image Means for obtaining a second index value indicating the brightness of
Means for calculating a correction value from the first and second index values; means for adding a threshold value to the reference image or inspection image; and means for changing the threshold value using the calculated correction value; Have
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter . Means for irradiating the substrate with laser light or focused light; means for collecting scattered light generated from the substrate by the irradiation means;
And emphasizing means than the scattered light generated from a normal circuit pattern scattered light from the defect or foreign matter on the substrate,
Means for imaging the collected scattered light on a detector;
A detector that converts the scattered light into charge information;
Means for converting charge information by scattered light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means includes means for calculating a correction value based on pre-given area information on the substrate related to brightness, means for adding a threshold value to the reference image or inspection image, and the calculated Means for changing the threshold value using the correction value,
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter . Means for irradiating the substrate with laser light or focused light; means for collecting scattered light generated from the substrate by the irradiation means;
And emphasizing means than the scattered light generated from a normal circuit pattern scattered light from the defect or foreign matter on the substrate,
Means for imaging the collected scattered light on a detector;
A detector that converts the scattered light into charge information;
Means for converting charge information by scattered light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Anda determined Mel means data size or location of the defect or foreign matter when the defect or foreign matter is present,
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means includes a means for detecting a brightness level of a reference image or an inspection image, a means for adding a threshold to the reference image or the inspection image, and a threshold using the detected brightness level. And means for changing
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter . Means for irradiating the substrate with laser light or focused light; means for collecting scattered light generated from the substrate by the irradiation means;
And emphasizing means than the scattered light generated from a normal circuit pattern scattered light from the defect or foreign matter on the substrate,
Means for imaging the collected scattered light on a detector;
A detector that converts the scattered light into charge information;
Means for converting charge information by scattered light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means includes means for obtaining a correlation distribution of brightness at corresponding locations of the reference image and the inspection image, means for calculating a correction value based on the distribution status of the correlation distribution, and the reference image or the inspection image. Means for adding a threshold value, and means for changing the threshold value using the calculated correction value,
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter . Means for irradiating a laser beam or focused beam over the substrate, and means for focusing onto the detector of the reflected light from the substrate,
A detector that converts the reflected light into charge information;
Means for converting charge information by reflected light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means obtains a first index value indicating the brightness of the entire reference image from the brightness in an arbitrary area of the reference image, and the inspection image from the brightness in an arbitrary area of the inspection image. Means for obtaining a second index value indicating overall brightness; means for obtaining a predetermined correction coefficient from the difference between the first and second index values; and brightness of an inspection image or reference image using the correction coefficient. Means for correcting the height level, and means for adding a threshold value to the reference image or inspection image,
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter . Means for irradiating a laser beam or focused beam over the substrate, and means for focusing onto the detector of the reflected light from the substrate,
A detector that converts the reflected light into charge information;
Means for converting charge information by reflected light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means obtains a first index value indicating the brightness of the entire reference image from the brightness in an arbitrary area of the reference image, and the entire inspection image from the brightness in the arbitrary area of the inspection image Means for obtaining a second index value indicating the brightness of
Means for calculating a correction value from the first and second index values; means for adding a threshold value to the reference image or inspection image; and means for changing the threshold value using the calculated correction value; Have
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter . Means for irradiating a laser beam or focused beam over the substrate, and means for focusing onto the detector of the reflected light from the substrate,
A detector that converts the reflected light into charge information;
Means for converting charge information by reflected light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means includes means for calculating a correction value based on pre-given area information on the substrate related to brightness, means for adding a threshold value to the reference image or inspection image, and the calculated Means for changing the threshold value using the correction value,
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. to the total amount of charge to identify the size of the defect or foreign matter from the detection of the defect or foreign material, characterized in that the position of the defect or foreign matter from the centroid of the pattern of the saturation region due to the defect or foreign matter having a means for the constant apparatus. Means for irradiating a laser beam or focused beam over the substrate, and means for focusing onto the detector of the reflected light from the substrate,
A detector that converts the reflected light into charge information;
Means for converting charge information by reflected light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means includes a means for detecting a brightness level of a reference image or an inspection image, a means for adding a threshold to the reference image or the inspection image, and a threshold using the detected brightness level. And means for changing
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter . Means for irradiating the substrate with laser light or focused light beam;
Means for collecting scattered light generated from the substrate by the irradiation means;
And emphasizing means than the scattered light generated from a normal circuit pattern scattered light from the defect or foreign matter on the substrate,
Means for imaging the collected scattered light on a detector;
A detector that converts the scattered light into charge information;
Means for converting charge information by scattered light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means includes means for obtaining a correlation distribution of brightness at corresponding locations of the reference image and the inspection image, means for calculating a correction value based on the distribution status of the correlation distribution, and the reference image or the inspection image. Means for adding a threshold value, and means for changing the threshold value using the calculated correction value,
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter . Means for irradiating the substrate with laser light or focused light beam;
Means for collecting scattered light generated from the substrate by the irradiation means;
And emphasizing means than the scattered light generated from a normal circuit pattern scattered light from the defect or foreign matter on the substrate,
Means for imaging the collected scattered light on a detector;
A detector that converts the scattered light into charge information;
Means for converting charge information by scattered light obtained by the detector into digital information;
Based on the converted digital information to obtain the reference image is a free image inspection image and defect or foreign material, and means for determining the presence or absence of defects or foreign matter from the difference of the test and reference images,
Means for obtaining data on the size or position of the defect or foreign matter when the defect or foreign matter is present , and
A detecting device for detecting a defect or foreign matter on the substrate,
The determination unit includes a comparison processing unit that compares the inspection image with a reference image;
The comparison processing means is means for shifting the reference image by a predetermined number of pixels in the X and Y directions, and means for obtaining a first difference image of each of the reference image and the inspection image shifted by the predetermined number of pixels. And means for obtaining a minimum value difference image from among the first difference images, and when the inspection image exceeds a preset threshold value, a first image of the inspection image and a reference image in phase with the inspection image Means for obtaining a difference image of 2; and means for outputting a smaller difference image of the minimum difference image and the second difference image;
When the pixel has reached saturation or blooming due to the presence of a defect or foreign matter in the inspection image, it is caused by the defect or foreign matter obtained by the product of the number of saturated pixels and the saturated digital information value. A defect or foreign matter detection device comprising: means for identifying a size of a defect or foreign matter from a total charge amount to be detected, and identifying a position of the defect or foreign matter from a center of gravity of a pattern of a saturated region caused by the defect or foreign matter .

Images