JP2004061447A - Substrate inspection method and selection means of substrate inspection method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a substrate inspection method for improving detection precision of a defect, and to provide a selection method of the substrate inspection method. <P>SOLUTION: The substrate inspection method detects the defect on the basis of a plurality of detection brightness values and a plurality of comparison brightness values determined to detect diffusion light from a detection position on a substrate and a comparison position corresponding thereto with a plurality of detectors arranged in different directions under dark field illumination. The method comprises a first synthetic process for synthesizing the plurality of the detection brightness values to produce one detection brightness synthetic value; a second synthetic process for synthesizing the plurality of the comparison brightness values to produce one brightness synthetic value; and a defect judging process for judging the presence of the defect at the detection position on the basis of the detection brightness synthetic value and the comparison brightness synthetic value. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a substrate inspection method for detecting a defect on a substrate and a method for selecting a substrate inspection method.
[0002]
[Prior art]
2. Description of the Related Art As a method for detecting a defect such as dust or a pattern collapse on the surface of a semiconductor substrate, a method using dark field illumination is known. In the conventional substrate inspection method under dark field illumination, the following processing is performed.
[0003]
That is, first, under dark-field illumination, scattered light from the inspection position on the substrate and the corresponding comparison position are detected by a plurality of detectors arranged in different directions, and a plurality of inspection luminance values and a plurality of comparison luminance values are detected. Get the luminance value. Then, a luminance value difference between the luminance value for inspection and the luminance value for comparison is obtained for each detector.
[0004]
Here, in the inspection under dark field illumination, irregularities on the substrate surface are also sensitively detected in addition to defects, so if metal grains exist on the surface, such as a substrate after metal deposition or metal etching, It is necessary to distinguish between defects and grains.
[0005]
Typically, the number of detectors that detect scattered light from the grains is less than the number of detectors that detect scattered light from defects. For example, when scattered light is detected by four detectors, scattered light from grains is often detected by one to three detectors due to its shape, while scattered light from defects is three to four. Are often detected by the detector.
[0006]
Therefore, conventionally, based on the number of detectors from which scattered light is detected and the above-described luminance value difference, defects are detected separately from grains.
[0007]
[Problems to be solved by the invention]
However, in the above-described conventional board inspection method, since the intensity of scattered light from the grain increases as the grain increases, most of the grain is detected as a defect, and the accuracy of defect detection decreases. There was a fear.
[0008]
The present invention has been made in order to solve the above-described problems, and an object of the present invention is to provide a substrate inspection method and a method for selecting a substrate inspection method capable of improving the accuracy of defect detection.
[0009]
[Means for Solving the Problems]
The inventor has conducted intensive studies to achieve the above object. As a result, since the scattered light from the defect has relatively low directionality, there is not much difference in the luminance value detected by each detector. Focused on the fact that there was a large difference in the luminance values detected by the detectors because of high directionality. Then, by combining a plurality of brightness values and generating one brightness combined value, such as averaging a plurality of brightness values detected by a plurality of detectors, the brightness combined value based on the scattered light from the grains is reduced. It was found that the luminance composite value based on the scattered light from the defect could be maintained high. Therefore, it has been found that the defect can be detected with high sensitivity by performing the defect determination based on the luminance composite value. The present invention is based on such findings.
[0010]
The substrate inspection method according to the present invention includes a plurality of inspection positions obtained by detecting scattered light from an inspection position on a substrate and a corresponding comparison position by a plurality of detectors arranged in different directions under dark-field illumination. This is a board inspection method for detecting a defect based on an inspection luminance value and a plurality of comparison luminance values. This method includes a first synthesizing step of synthesizing a plurality of inspection luminance values to generate one inspection luminance synthesis value, and synthesizing a plurality of comparison luminance values to generate one comparison luminance synthesis value. And a defect determination step of determining the presence or absence of a defect at the inspection position based on the inspection luminance composite value and the comparison luminance composite value.
[0011]
In this inspection method, a plurality of inspection luminance values are combined to generate one inspection luminance combination value, and a plurality of comparison luminance values are combined to generate one comparison luminance combination value. According to the luminance composite value generated in this manner, the value can be reduced when the luminance is based on the scattered light from the grains, and can be maintained high when the luminance is based on the scattered light from the defect. Therefore, by determining a defect based on the luminance composite value for inspection and the luminance composite value for comparison, the defect can be detected with high sensitivity, and the accuracy of defect detection can be improved.
[0012]
In the board inspection method according to the present invention, in the first synthesizing step, a plurality of inspection luminance values are averaged to generate an inspection luminance composite value, and in the second synthesizing step, the plurality of comparison luminance values are averaged. It is preferable to generate a comparative luminance composite value in this way.
[0013]
The method for selecting a substrate inspection method according to the present invention includes a luminance value group for detecting scattered light from an inspection region and a comparison region on a substrate under dark-field illumination to obtain a luminance value group for inspection and a luminance value group for comparison. Obtaining the difference between the luminance value for inspection and the luminance value for comparison at a position corresponding to each of the inspection region and the comparative region, based on the luminance value group for inspection and the luminance value group for comparison, A standard deviation calculation step of generating and calculating a standard deviation of the luminance value difference based on the luminance value difference group, and one inspection method from a plurality of inspection methods for detecting a defect on the substrate according to the standard deviation. And a selecting step of selecting.
[0014]
In this selection method, it is possible to improve the defect detection accuracy by selecting an optimal inspection method for detecting a defect on a substrate according to the standard deviation of the luminance value difference, that is, the magnitude of the noise. It becomes.
[0015]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the description of the drawings, the same elements will be denoted by the same reference symbols, without redundant description. Also, the dimensional ratios in the drawings do not always match those described.
[0016]
FIG. 1 is a diagram schematically illustrating a configuration of a substrate inspection apparatus (hereinafter, also simply referred to as an inspection apparatus) 10 for performing a substrate inspection method according to an embodiment of the present invention. FIG. 1A shows the configuration of the detection unit 20 of the inspection device 10, and FIG. 1B shows the configuration of the data analysis unit 30 of the inspection device 10.
[0017]
The detection unit 20 includes a light source 22 and a plurality of detectors # 1 to # 4, as shown in FIG. The light source 22 irradiates a laser beam to scan a predetermined region on the semiconductor substrate W. The laser light emitted from the light source 22 is incident on the semiconductor substrate W almost vertically and is scattered on the substrate W. The plurality of detectors (four in this embodiment) # 1 to # 4 detect light scattered on the substrate W at a predetermined scattering angle θ1. Thereby, a dark-field image on the substrate W is obtained. These four detectors # 1 to # 4 are arranged obliquely above the substrate W (scattering angle θ1) at equal intervals of 45 degrees so as to surround the substrate W. Each of the detectors # 1 to # 4 converts the detected optical signal into an electric signal and generates luminance value data represented by 256 gradations. Each of the detectors # 1 to # 4 is electrically connected to the data storage unit 32 of the data analysis unit 30.
[0018]
In addition, the detection unit 20 has a detector # 5 that detects light reflected almost vertically by the substrate W. A bright field image on the substrate W is acquired by the detector # 5. Further, the detection unit 20 has a detector # 6 that detects light scattered at a scattering angle θ2 larger than the scattering angle θ1 of the scattered light detected by the detectors # 1 to # 4. With this detector # 6, a dark-field image on the substrate W at the scattering angle θ2 is obtained. Note that the reflected light from the substrate W and the scattered light at the scattering angle θ2 are transmitted to the detectors # 5 and # 6 by the desired optical system 24, respectively. Note that each of the detectors # 5 and # 6 is also electrically connected to the data storage unit 32 of the data analysis unit 30.
[0019]
As shown in FIG. 1B, the data analysis unit 30 includes a data storage unit 32, a calculation unit 34, and a display unit 36 for displaying a calculation result. The luminance value data detected by each of the detectors # 1 to # 4 and the detectors # 5 and # 6 of the detection unit 20 are input to the data storage unit 32 and stored. Then, based on the data stored in the data storage unit 32, various calculations are performed in the calculation unit 34. The result of the calculation is displayed on the display unit 36.
[0020]
A board inspection method using the above board inspection apparatus 10 will be described.
[0021]
(1st Embodiment)
The substrate inspection method according to the first embodiment will be described with reference to the flowchart in FIG.
[0022]
First, a laser beam is emitted from the light source 22 to scan the inspection region I and the comparison region R on the substrate W, and the scattered light scattered on the substrate W is detected by the four detectors # 1 to # 4. Two inspection luminance value groups and four comparison luminance value groups are acquired in advance. In the present embodiment, a case is considered in which a plurality of chips having the same pattern are formed on a substrate W, as shown in FIG. The area occupied by the chip to be inspected is defined as an inspection area I, and the area occupied by a chip adjacent thereto is defined as a comparison area R.
[0023]
When obtaining four inspection brightness value groups and four comparison brightness value groups, as shown by arrows in FIG. 3B, the chip to be inspected and the chip to be compared are scanned with a laser beam to obtain scattered light. At a predetermined sampling cycle by the four detectors # 1 to # 4.
[0024]
FIG. 4 is a diagram for explaining a test luminance value group. As shown in FIGS. 4A to 4D, an inspection luminance value g is acquired for each detector in association with each inspection position on the inspection area I, whereby one detection luminance value group is obtained. Is configured. For example, as shown in FIG. 4A, the detector # 1 associates the inspection luminance value g with the inspection position i on the inspection area I._I1(I) is acquired and one detection luminance value group g_I1(1) -g_I1(N) is configured. In the present embodiment, since the scattered light is detected by the four detectors # 1 to # 4, four inspection luminance value groups Δg_I1(1) -g_I1(N)｝, ｛g_I2(1) -g_I2(N)｝, ｛g_I3(1) -g_I3(N)｝, ｛g_I4(1) -g_I4(N) is acquired and stored in the data storage unit 32.
[0025]
FIG. 5 is a diagram for explaining a comparative brightness value group. As shown in FIGS. 5A to 5D, a comparison luminance value g is acquired for each detector in association with each comparison position on the comparison region R, thereby obtaining one comparison luminance value group. Is configured. For example, as shown in FIG. 5A, the detector # 1 associates the comparison luminance value g with the comparison position i on the comparison area R._R1(I) is acquired, and one comparative brightness value group g_R1(1) -g_R1(N) is configured. In the present embodiment, since the scattered light is detected by the four detectors # 1 to # 4, the four comparative brightness value groups ｛g_R1(1) -g_R1(N)｝, ｛g_R2(1) -g_R2(N)｝, ｛g_R3(1) -g_R3(N)｝, ｛g_R4(1) -g_R4(N) is acquired and stored in the data storage unit 32.
[0026]
Then, a defect inspection is performed based on the data obtained in advance and stored in the data storage unit 32 as described above.
[0027]
First, as shown in FIG._R ^*(I) is generated (step S10). In the generation of the comparative luminance composite value, the comparative luminance value g at the i-th comparative position i from each of the four comparative luminance value groups._R1(I) -g_R4(I) is extracted. This comparison position i corresponds to the inspection position i in the inspection area I. Then, the four comparative luminance values are combined to form one comparative luminance composite value g._R ^*Generate (i). Here, “synthesis” refers to generating a single luminance value by reflecting the variation in the comparative luminance value at the same comparison position acquired by the n detectors. For example, it means obtaining an average value of n comparison luminance values, or obtaining an nth root by multiplying the n comparison luminance values. Further, a luminance value difference between the average value or the nth root obtained in this way and the luminance value under another dark field illumination acquired by the detector # 6 may be calculated. Here, a case will be described in which an average value of four comparison luminance values is obtained. Then, for all the comparison positions 1 to n in the comparison region R, the comparison luminance composite value g_R ^*(1) -g_R ^*(N) is generated.
[0028]
Next, as shown in FIG. 2, a test luminance composite value is generated (step S12). In the generation of the inspection luminance composite value, as shown in FIG. 4, the inspection luminance value g at the i-th inspection position i from the four inspection luminance value groups._I1(I) -g_I4(I) is extracted. This inspection position i corresponds to the comparison position i in the comparison region R. Then, these four test luminance values are combined to form one test luminance composite value g._I ^*Generate (i). Here, “synthesis” refers to generating one luminance value by reflecting the degree of variation of the inspection luminance value at the same inspection position acquired by the n detectors. For example, this means obtaining an average value of n inspection luminance values, or obtaining an nth root by multiplying the n inspection luminance values. Further, a luminance value difference between the average value or the nth root obtained in this way and the luminance value under another dark field illumination acquired by the detector # 6 may be calculated. Here, a case where an average value of four inspection luminance values is obtained will be described. Then, for all inspection positions 1 to n in the inspection area I, the inspection luminance composite value g_I ^*(1) -g_I ^*(N) is generated.
[0029]
Note that the above-described “combining” processing will be visually described with reference to FIGS. 6 and 7. FIGS. 6A to 6D show an example of an inspection luminance value group (or a comparison luminance value group) detected by each of the detectors # 1 to # 4. As shown in FIG. 6, in each graph, the bottom surface S corresponds to the inspection region I (comparison region R), and the luminance value is plotted on the vertical axis, so that the luminance value for each inspection position i (comparison position i) is three-dimensional. It is drawn in.
[0030]
As shown in FIG. 6, since the scattered light from the defect has low directionality, there is not much difference between the luminance values detected by the detectors # 1 to # 4. On the other hand, since the scattered light from the grains has high directionality, the scattered light is detected by the detectors # 1 and # 3 arranged opposite to each other, but hardly detected by the detectors # 2 and # 4. However, there is a large difference between the luminance values of the detectors # 1 to # 4. Therefore, the inspection luminance value g detected by the four detectors # 1 to # 4 based on the scattered light from the inspection position i (comparison position i)_I1(I) -g_I4(I) (Comparative luminance value g_R1(I) -g_R4(I) averages one test luminance composite value g_I ^*(I) (Comparison luminance composite value g_R ^*By generating (i)), as shown in FIG. 7, the inspection luminance composite value (comparative luminance composite value) based on the scattered light from the defect can be kept high, but based on the scattered light from the grain. The test luminance composite value (comparative luminance composite value) can be reduced, and as a result, noise components can be reduced.
[0031]
Next, the test luminance composite value g thus generated_I ^*(1) -g_I ^*(N) and comparative luminance composite value g_R ^*(1) -g_R ^*From (n), the corresponding test luminance composite values g_I ^*(I) and luminance composite value g for comparison_R ^*(I) is extracted and the extracted test luminance composite value g_I ^*(I) and comparative luminance composite value g_R ^*Based on (i), the presence or absence of a defect at each inspection position i in the inspection area I is determined (step S14).
[0032]
The process in step S14 will be described with reference to FIG. In FIG. 8, the horizontal axis represents the luminance composite value g of the comparison region R._R ^*(I), and the vertical axis represents the luminance composite value g of the inspection area I._I ^*(I). Since the plurality of chips provided on the substrate W have the same pattern, the i-th inspection position i in the inspection region I and the i-th comparison position i in the comparison region R corresponding thereto are different. , Inspection luminance composite value g_I ^*(I) and comparative luminance composite value g_R ^*(I) is basically the same value. Therefore, the corresponding test luminance composite values g_I ^*(I) and comparative luminance composite value g_R ^*When a point is plotted in FIG. 8 based on (i), the point is basically represented by an equivalent line L_BCome on.
[0033]
However, when there is a defect at the inspection position i, the inspection luminance composite value g_I ^*(I) is a comparative luminance composite value g for a comparative position i having no defect._R ^*The value is much larger than (i), and the points are plotted in FIG.₁As shown in FIG._BDivergence from. Therefore, the point plotted as the point D1 is the threshold line J and the equivalent line L_BIf it is not between them, this is detected as a defect.
[0034]
In the present embodiment, as described with reference to FIG. 7, since the influence of noise such as grains is reduced, the luminance composite value g for inspection based on the scattered light from the grains and the like is used._I ^*(I) and comparative luminance composite value g_R ^*When a point is plotted in FIG. 8 based on (i), the point N is equivalent to the equivalent line L_BDeviation from is small.
[0035]
On the other hand, according to the conventional method, since the inspection luminance value and the comparison luminance value are compared for each of the detectors # 1 to # 4, the points are plotted as in FIG. Thus, the point N is equivalent to the equivalent line L_BThe deviation from is large. Therefore, when the defect detection level is lowered (when the threshold line J is raised), the number of undetected defects increases, and the detection accuracy decreases. On the other hand, when the detection level of the defect is increased (when the threshold line J is decreased), many grains are detected as defects, and the accuracy of defect detection is reduced.
[0036]
On the other hand, in the present embodiment, the inspection luminance generated by combining a plurality of inspection luminance values without comparing the inspection luminance value with the comparison luminance value for each of the detectors # 1 to # 4. Composite value g_I ^*A comparative luminance composite value g generated by composing (i) with a plurality of comparative luminance values_R ^*By comparing with (i), the possibility that noise such as grain is erroneously detected as a defect is reduced, and conversely, the possibility that a defect to be detected becomes undetected is reduced. It is possible to improve the detection accuracy.
[0037]
In FIG. 8, the threshold line J can be set according to a predetermined condition, and the equivalent line L_BThe offset amount and the inclination thereof can be arbitrarily set.
[0038]
Further, in the above description, it has been described that a defect is detected with high accuracy when there is a defect in the inspection area I. However, as shown in FIG._BA threshold line K for the comparison area R is set below and the equivalent line L_BBy detecting a point that is not in the area between the threshold line K and the threshold line K, the defect D in the comparison area R is detected.₂Can also be accurately detected at the same time.
[0039]
By repeating the above process for all the chips on the substrate W, for example, in order from the upper left chip of the substrate W shown in FIG.
[0040]
(2nd Embodiment)
Next, a substrate inspection method according to the second embodiment will be described with reference to the flowchart in FIG.
[0041]
The substrate inspection method according to the present embodiment includes a selection method for selecting an optimal inspection method for detecting a defect according to the state of noise on the substrate W. In this embodiment, a case is considered in which a plurality of chips having the same pattern are formed on a substrate W as shown in FIG.
[0042]
In this substrate inspection method, first, an inspection method for inspecting the substrate W is selected in steps S20 to S30 as shown in FIG. Then, in step S32, the substrate W is inspected by the selected inspection method.
[0043]
In the selection of the inspection method, in step S20, first, laser light is emitted from the light source 22 to scan the inspection region I and the comparison region R on the substrate W, and scattered light is detected by the four detectors # 1 to # 4. . As a result, four inspection brightness value groups and four comparison brightness value groups as shown in FIGS. 4 and 5 are obtained (a brightness value group obtaining step).
[0044]
Next, in step S22, based on the acquired four inspection luminance value groups and the four comparison luminance value groups, at each of the positions corresponding to the inspection region I and the comparison region R for each of the detectors # 1 to # 4. A difference between the luminance value for inspection and the luminance value for comparison is obtained to generate a luminance value difference group. For example, referring to FIG. 4A and FIG. 5A, the inspection luminance value g at the corresponding position for the detector # 1 will be described._I1(I) and luminance value g for comparison_R1By extracting (i) and taking these differences, a luminance value difference group for detector # 1 is obtained. As a result, the pattern on the chip is canceled, and a data group having only noise information is obtained. The calculation of the brightness value difference group is performed for all the detectors # 1 to # 4, and four brightness value difference groups are obtained. Next, in step S24, a standard deviation σ of the luminance value difference is calculated based on the generated four luminance value difference groups (standard deviation calculation step).
[0045]
In the above description, four detector brightness value groups and four comparison brightness value groups are obtained by the four detectors # 1 to # 4, and four brightness value difference groups are calculated based on these. Although the standard deviation σ of the luminance value difference was obtained, the present invention is not limited to this, and one inspection luminance value group and one comparison luminance value group are obtained by at least one detector, and one luminance value group is obtained based on these. The difference group may be calculated to determine the standard deviation σ of the luminance value difference.
[0046]
Next, in step S26, the calculated standard deviation σ is set to a predetermined reference value σ₀It is determined whether it is greater than. The standard deviation σ is equal to a predetermined reference value σ₀If it is larger, the process proceeds to step S28, and the first inspection method is selected. On the other hand, the standard deviation σ is a predetermined reference value σ₀In the following cases, the process proceeds to step S30, and the second inspection method is selected. As described above, one inspection method is selected from a plurality of (in this case, two) inspection methods for detecting a defect on the substrate W according to the standard deviation σ (selection step). Here, a predetermined reference value σ₀Can be specified by a parameter file, but in normal operation the desired value is automatically applied. The selection of the inspection method is usually performed at the time of set-up when inspecting a predetermined lot.
[0047]
Then, in step S32, the substrate W is inspected by the first inspection method selected in step S28 or the second inspection method selected in step S30. Here, the first inspection method is the substrate inspection method described in the first embodiment. In this inspection method, as described above, when scattering from noise such as grains is large, the influence can be reduced and a defect can be detected with high sensitivity. Therefore, when the standard deviation σ of the luminance value difference between the inspection area I and the comparison area R is large, that is, when the influence from noise such as grain is large, the influence can be reduced and the defect can be detected with high sensitivity. Inspection by the inspection method 1 can improve the accuracy of defect detection.
[0048]
On the other hand, the second inspection method is a conventional substrate inspection method described in the related art. This method may cause a decrease in defect detection accuracy when the influence from noise such as grain is large, but is still an effective inspection method when the influence from noise such as grain is small. For example, when the scattered light from the defect is weak and the signal level is low, or when the scattered light has directionality such as a scratch after CMP (chemical mechanical polishing), the inspection is performed by the first inspection method. Then, necessary information may be weakened by the synthesis, and the accuracy of defect detection may be reduced. Therefore, when the standard deviation σ is small and the influence from noise such as grains is small, the accuracy of detecting a defect can be improved by inspecting the substrate W by the second inspection method.
[0049]
Here, the second conventional inspection method will be described. As in the first inspection method, laser light is emitted from the light source 22 to scan the inspection region I and the comparison region R on the substrate W, Are detected by the four detectors # 1 to # 4 to obtain four inspection luminance value groups and four comparison luminance value groups in advance.
[0050]
Next, a luminance value for inspection and a luminance value for comparison at a corresponding position i are extracted from the luminance value group for inspection and the luminance value group for comparison for each of the detectors # 1 to # 4, and a luminance value difference is obtained. For example, referring to FIGS. 4A and 5A, the inspection luminance value g at the corresponding position will be described._I1(I) and luminance value g for comparison_R1(I) is extracted, and a luminance value difference is obtained. The calculation of the luminance value difference is performed for all the detectors # 1 to # 4. Then, the presence or absence of a defect at the position i is determined based on the magnitude of the luminance value difference and the number of detectors from which the scattered light has been detected. By performing this operation for all the inspection positions i in the inspection area I, the chip to be inspected is inspected. Further, by repeating this series of processes for all chips on the substrate W, for example, in order from the upper left chip of the substrate W shown in FIG. Is
[0051]
Note that the present invention is not limited to the above-described embodiment, and various changes can be made.
[0052]
For example, the number of detectors of the board inspection apparatus 10 is not limited to four, and it is sufficient that at least three or more detectors are provided. That is, the number of test luminance values (comparative luminance values) for calculating the test luminance composite values (comparative luminance composite values) is not limited to four, and may be at least three or more. Further, the calculation for the combination is not limited to the calculation of the average value or the nth root.
[0053]
Further, in the above-described second embodiment, an inspection method is selected from two types of methods according to the standard deviation σ, but an inspection method may be selected from three or more types of methods according to the value.
[0054]
【The invention's effect】
According to the present invention, there is provided a substrate inspection method and a method for selecting a substrate inspection method capable of improving the accuracy of defect detection.
[Brief description of the drawings]
FIG. 1 is a diagram schematically illustrating a configuration of a substrate inspection apparatus for executing a substrate inspection method according to an embodiment. FIG. 1A shows the configuration of the detection unit, and FIG. 1B shows the configuration of the data analysis unit.
FIG. 2 is a flowchart illustrating a board inspection method according to the first embodiment.
FIG. 3A is a diagram showing a substrate to be inspected, and shows a state in which a plurality of chips having the same pattern are formed. FIG. 3B is a diagram illustrating a state in which the inspection region and the comparison region are scanned by the laser light.
FIGS. 4A to 4D are diagrams illustrating a test luminance value group of a test area detected and acquired by each detector.
FIGS. 5A to 5D are diagrams illustrating a comparative brightness value group of a comparison area detected and acquired by each detector.
6 (a) to 6 (d) plot an example of an inspection luminance value group (comparison luminance value group) detected and obtained by each detector for each inspection position (comparison position). It is a graph.
FIGS. 7A to 7D show a test luminance composite value (comparative luminance composite value) obtained by synthesizing a plurality of test luminance values (comparative luminance values) in the cases shown in FIGS. And a graph plotted for each inspection position (comparison position).
FIG. 8 is a diagram for explaining a method of determining the presence or absence of a defect based on the combined luminance for inspection and the combined luminance for comparison at the corresponding positions by the board inspection method according to the first embodiment. .
FIG. 9 is a diagram for explaining a method of determining the presence or absence of a defect based on an inspection luminance value and a comparison luminance value at corresponding positions by a conventional substrate inspection method.
FIG. 10 is a flowchart illustrating a board inspection method according to a second embodiment.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Board inspection apparatus, 20 ... Detection part, 22 ... Light source, 30 ... Data analysis part, 32 ... Data storage part, 34 ... Operation part, 36 ... Display part, I ... Inspection area, R ... Comparison area, ..., # 1 to # 4: detector, W: semiconductor substrate.

Claims (3)

A plurality of inspection luminance values and a plurality of comparison values obtained by detecting scattered light from an inspection position on a substrate and a corresponding comparison position on the substrate under dark field illumination by a plurality of detectors arranged in different directions. A board inspection method for detecting a defect based on a luminance value,
A first combining step of combining the plurality of test luminance values to generate one test luminance composite value;
A second combining step of combining the plurality of comparative brightness values to generate one comparative brightness combined value;
A defect determination step of determining the presence or absence of a defect at the inspection position based on the inspection luminance composite value and the comparison luminance composite value;
A substrate inspection method, comprising: In the first synthesizing step, the plurality of inspection luminance values are averaged to generate the inspection luminance composite value, and in the second synthesizing step, the plurality of comparison luminance values are averaged to perform the comparison. The board inspection method according to claim 1, wherein a combined luminance value is generated. A luminance value group acquiring step of detecting scattered light from the inspection region and the comparison region on the substrate under dark field illumination to acquire a luminance value group for inspection and a luminance value group for comparison,
Based on the inspection luminance value group and the comparison luminance value group, a difference between the inspection luminance value and the comparison luminance value at a position corresponding to each of the inspection region and the comparison region is calculated to obtain a luminance value difference group. Generating a standard deviation calculating step of calculating a standard deviation of the luminance value difference based on the luminance value difference group,
A selecting step of selecting one inspection method from a plurality of types of inspection methods for detecting a defect on the substrate according to the standard deviation,
A method for selecting a substrate inspection method, comprising:

Images