JP2008249611A - Manufacturing method for semiconductor device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a manufacturing method for a semiconductor device capable of accurately determining the existence of defectives. <P>SOLUTION: A solder film is formed at first on an electrode. Then, the solder film is irradiated diagonally with a light, and a reflected light thereof is measured to obtain an original image. Then, the laminance of the original image is differentiated along the long-side direction to obtain a differential data. Then, the differential data are integrated along a short-side direction to find an integrated value. Then, the defective is determined, when a difference between the maximum value and the minimum value of the integrated value is a prescribed value or larger; and when the width between the position of the integrated value becoming a maximum and the position of the integrated value becoming a minimum is equal to a prescribed width or smaller. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a method for manufacturing a semiconductor device in which a solder film is formed on an electrode and whether or not the solder film is defective is determined.

  Flip chip bonding is used when connecting a semiconductor chip and a wiring board. In flip chip bonding, a solder film is formed in advance on an electrode of a wiring board, and external terminals of the semiconductor chip are connected to the electrode of the wiring board through this solder film (see, for example, Patent Document 1).

  48 and 49 are images of the solder film formed on the electrode taken from above. 48, when there is a DC (Dewet Center) defect where there is no solder in the middle of the electrode or a DE (Dewet Edge) defect where there is no solder at the end of the electrode as shown in FIG. There is.

JP 2006-324642 A

  Conventionally, the presence or absence of DC failure or DE failure has been determined as follows. First, light is applied obliquely to the solder film and the reflected light is measured to obtain an original image. Then, the presence or absence of a defect is determined by looking at the original image binarized depending on whether the luminance is above or below the threshold. However, due to the narrowing of the pitch of the electrodes and the like, the conventional method has a problem that there are many false reports for overlooking a defective product and determining that a non-defective product is defective.

  Moreover, a defect that the solder amount of the solder film is insufficient may occur. However, conventionally, since it was determined whether or not the amount of solder was insufficient by measuring the height with a laser displacement meter, there was a problem that it took a long time to measure.

  In addition, there may be a defect in which the gap between the leads becomes too narrow due to the presence of foreign matter. However, it is difficult to identify the foreign matter because the difference in luminance from the surface of the insulating film of the wiring board is smaller than the DC failure or the DE failure.

  In addition, black defects may be generated in which the solder corrodes and becomes black due to a flux residue or the like. FIG. 50 is a laser microscope image showing a solder film in which a black defect has occurred. Conventionally, the presence or absence of a black defect has been determined by observing an original image obtained by irradiating light on the solder film obliquely and measuring the reflected light. However, in the original image obtained by the conventional method, as shown in FIG. 51, the inclined surface of the solder film appears white, but the plane appears black. Therefore, it was difficult to detect minor black defects.

  The present invention has been made to solve the above-described problems, and a first object thereof is to obtain a semiconductor device manufacturing method capable of accurately determining the presence or absence of defects.

  A second object of the present invention is to obtain a method of manufacturing a semiconductor device that can quickly determine whether or not the solder film has a short amount of solder.

  A third object of the present invention is to obtain a method of manufacturing a semiconductor device that can determine the presence or absence of a defect in which the gap between leads becomes too narrow due to the presence of foreign matter.

  A fourth object of the present invention is to obtain a method of manufacturing a semiconductor device that can detect even a slight black defect.

  In the method of manufacturing a semiconductor device according to one embodiment of the present invention, first, a solder film is formed on an electrode. Next, light is applied obliquely to the solder film and the reflected light is measured to obtain an original image. Next, differential data is obtained by differentiating the luminance of the original image in the long side direction. Next, the differential data is integrated in the short side direction to obtain an integrated value. Next, when the difference between the maximum value and the minimum value of the integrated values is equal to or greater than a predetermined value and the width between the position where the integrated value is maximum and the minimum position is equal to or less than the predetermined width, it is determined as defective.

  According to this embodiment, the presence / absence of a defect can be accurately determined.

  Hereinafter, a semiconductor device manufacturing method according to an embodiment of the present invention will be described in detail with reference to the drawings.

  First, as shown in FIG. 1, a Cu electrode 12 (electrode) is formed on a wiring board 11. The Cu electrodes 12 are arranged at a pitch of 40 μm, and the planar shape has a long side and a short side.

  Next, as shown in FIG. 2, a solder paste 13 in which solder particles and a liquid organic material are mixed is printed on the Cu electrode 12. Here, lead-free solder is used as the solder particles. The lead-free solder is a solder that does not contain lead or contains only lead with a low environmental load (less than 1 wt%). Here, as the lead-free solder, Sn containing 1 to 3% of Ag is used.

  Next, as shown in FIG. 3, reflow is performed to form a solder film 14 on the Cu electrode 12. At this time, however, solder balls 15 are also formed on the solder film 14. Next, as shown in FIG. 4, the solder paste 13 is removed by washing. Then, as shown in FIG. 5, reflow is performed again to eliminate the solder balls 15. Thereafter, whether or not the solder film 14 is defective is determined. This will be described in detail later.

  Next, as shown in FIG. 6, NCP (Non Conductive Paste) 17 is applied using a dispenser 16 so as to cover the Cu electrode 12. Then, as shown in FIG. 7, the semiconductor chip 18 is temporarily mounted on the wiring board 11 with the mounting surface of the semiconductor chip 18 facing down. At this time, the Au stud bump 19 (external terminal) formed on the mounting surface of the semiconductor chip 18 is brought into contact with the solder film 14 formed on the Cu electrode 12 of the wiring board 11.

  Next, as shown in FIG. 8, the semiconductor chip 18 is pressed from above through a Teflon (registered trademark) sheet 22 using a tool 21. At this time, the semiconductor chip 18 is heated to about 150 ° C. by the tool 21. As a result, the NCP 17 is cured and the solder film 14 is melted to solder the Au stud bump 19 of the semiconductor chip 18 to the Cu electrode 12 of the wiring board 11. Through the above steps, a semiconductor device in which the semiconductor chip 18 is flip-chip bonded onto the wiring substrate 11 is manufactured as shown in FIG.

  Next, the process of determining whether or not the solder film 14 is defective will be described in detail with reference to the flowchart of FIG.

  FIG. 11 is a diagram showing an optical system for illuminating and photographing the solder film. This optical system includes a CCD area camera 31, an XYZ table 32, a glass calibration substrate 33, X and Y axis galvanometer mirrors 34 and 35, X and Y axis line sensors 36 and 37, a lens 38, The objective lens 23, the coaxial illumination means 24, and the ring illumination means 25 are provided. The CCD area camera 31 is a photographing means for photographing a measurement target. The XYZ table 32 moves the measurement target in the X-axis, Y-axis, and Z-axis directions, respectively.

  The illumination system will be described in further detail. FIG. 12 is a diagram showing an optical system for illuminating the solder film. The objective lens 23 is a light input means for inputting reflected light. The coaxial illumination means 24 is made of, for example, a blue LED, and illuminates the solder film 14 coaxially with the objective lens 23. The ring illumination means 25 is configured in a ring shape with a circular center located on the solder film 14, and illuminates the solder film 14 at three angles of 13 °, 21 °, and 32 °. When light is applied obliquely to the solder film 14 by the ring illumination means 25, reflected light from the slope of the solder film 14 is obtained as shown in FIG. When the solder film 14 is illuminated by the coaxial illumination means 24, reflected light from the plane of the solder film 14 is obtained as shown in FIG.

  First, the ring illumination means 25 is used to apply light obliquely to the solder film 14 and measure the reflected light to obtain an original image as shown in FIG. In this original image, the slope of the solder film 14 appears white, and the plane of the solder film 14 appears black. In addition, the horizontal direction of drawing is the short side direction of Cu electrode, and the vertical direction of drawing is the long side direction of Cu electrode.

  Next, differential data as shown in FIG. 16 is obtained by differentiating the luminance of the original image in the long side direction. Then, the differential data is integrated in the short side direction to obtain an integrated value as shown in FIG. When the difference between the maximum value and the minimum value of the integrated values is equal to or greater than a predetermined value, and the width between the position where the integrated value is maximum and the minimum position is equal to or less than the predetermined width, it is determined as defective (step S1). .

  Next, when it is determined that the threshold value is greatly deviated in step S1, it is determined as defective and the determination is terminated. On the other hand, if it is determined in step S1 that there is a defect just below the threshold for determining pass / fail, the next detection is performed (step S2).

  As shown in FIG. 18, when there is a DC failure, the black region whose luminance is lower than the threshold is divided into two or more on the Cu electrode. Therefore, when the black area is divided into two or more, it is determined as defective, and when the black area is one or less, it is determined as good (step S3). If the change from the white area to the black area is not clear, the presence or absence of DC failure cannot be determined by the inspection in step S1, but it can be determined by the inspection in step S3.

  If it is determined in step S1 that it is good, or if it is determined in S3 that there are one or less black areas, the presence or absence of a DE failure is inspected as follows. First, as in step S1, the luminance of the original image shown in FIG. 19 is differentiated in the long side direction to obtain differential data, and the luminance of this differential data is integrated in the short side direction and integrated as shown in FIG. Find the value. On the Cu electrode, when the maximum value of the integrated value is not less than a predetermined value, it is determined as defective (step S4).

  When the Cu electrode is thin, the ratio of the slope of the solder film increases. Therefore, when obtaining the differential data, it is preferable to differentiate the luminance of the original image in the long side direction only for the middle part of the Cu electrode in the short side direction as shown in FIG. Thereby, even when the Cu electrode is thin, it is possible to capture the feature of DE failure.

  If it is determined to be defective in step S4, the luminance of the original image as shown in FIG. 19 is integrated in the short side direction to obtain an integrated value as shown in FIG. If there is a portion on the Cu electrode where the integrated value is 60% or less of the maximum value, it is determined as defective (step S5).

  If it is determined as good in step S4 or S5, an approximate curve of the outer periphery of the solder film in the original image is obtained by the least square method as shown in FIG. 23 (step S6). If the approximate curve cannot be obtained, the determination value is tightened and the same inspection as in step S4 is performed (step S7). On the other hand, when the approximate curve is obtained, the original image is binarized depending on whether the luminance is above or below the threshold value, and when the protruding area from the approximate curve of the white region is equal to or larger than a predetermined value, it is determined as defective (step S8). Thereby, the presence or absence of foreign matter on the Cu electrode can be inspected.

  Next, a binary inspection is performed for the presence or absence of a DE failure. FIG. 24 is a laser microscope image of a solder film having a DE defect, FIG. 25 is an original image obtained by measuring the reflected light by shining light on the solder film having a DE defect, and FIG. This is an image obtained by binarizing the original image depending on whether the luminance is above or below the threshold. In the binarized image, if the width between the edge of the long side direction of the white region and the center of the electrode is equal to or smaller than a predetermined width, it is determined as defective (step S9).

  If it is determined to be defective in step S9, a multi-value inspection is performed for the presence or absence of a DE defect. That is, edge enhancement processing is performed on the original image as shown in FIG. In this processed image, if the width between the edge of the long side direction of the white region and the center of the electrode is equal to or smaller than a predetermined width, it is determined as defective (step S10). By performing the edge emphasis process in this way, the influence of wrinkles on the surface of the solder film can be eliminated.

  If it is determined that it is defective in step S10, an approximate curve of the outer periphery of the solder film is obtained in the original image by the least square method as shown in FIG. 28 (step S11). When the approximate curve is obtained, if the width in the short side direction between the approximate curves is equal to or smaller than the predetermined width, it is determined as defective (step S12). Thereby, it is possible to inspect for the presence or absence of dewetting with almost no solder on the Cu electrode.

  Next, a binary area inspection is performed for the presence or absence of a DE failure. That is, as shown in FIG. 29, when the area of a white area whose luminance is equal to or greater than a threshold value within a predetermined measurement area of the original image is equal to or less than a predetermined value, it is determined as defective (step S13). Thereby, it is possible to detect a defect with a small solder slope.

  Only the solder film determined to be defective in step S13 is measured with a laser displacement meter as shown in FIG. 30 (step S14). As a result, the number of laser height measurements with a long measurement time can be reduced to increase production efficiency.

  Laser height measurement is performed at points A, B, and C as shown in FIGS. Point A is a region on the Cu lead 26 drawn from the Cu electrode 12, point B is a region on the Cu lead 26 drawn from the Cu electrode 12 in the direction opposite to the point A, and point C is the Cu electrode. This is a 16 μm square region including 12 centers. At each point, the height of a plurality of points is measured in the long side direction and the short side direction, and the highest point is taken as the measured value of each point.

The lower height of the point A and point B and _{H AB,} the height of the point C as _{H C,} satisfies the relationship _{5 <H C -H AB +2 <} 20 (μm), determined as good. That is, if the difference between the height of the solder film 14 on the Cu electrode 12 and the height of the Cu lead 26 drawn from the Cu electrode 12 is 3 μm or less or 18 μm or more, it is determined as defective.

  Here, the reason why 2 μm is added is that the solder is placed on the points A and B about 2 μm. The reason why the difference between the point A and the point B is obtained on the basis of the lower one of the points A and B is that the Cu lead 26 may be inclined.

  Next, in the original image subjected to the following first and second image processing, a narrow gap inspection is performed in which a defect is determined when the distance between the closest positions of areas having luminance equal to or higher than a threshold is equal to or less than a predetermined length. This is performed (step S15). Accordingly, as shown in FIG. 33, it is possible to determine whether there is a defect in which the gap is too narrow due to the presence of the foreign matter 27 or the like.

  As the first image processing, as shown in FIG. 34, processing for increasing the luminance of the portion corresponding to the Cu lead 26 drawn from the Cu electrode 12 to a threshold value or more (filled in white) is performed on the original image. Since only a thin solder is placed on the Cu lead 26, it appears as if black and white are mixed in the original image. Therefore, it is possible to prevent erroneous determination in the narrow gap inspection by painting this portion white.

  However, as shown in FIG. 36, the first to third processed images obtained by performing the second image processing in the following three directions are combined, and the combined image is binarized to perform a narrow gap inspection. First, the second processing is performed with the processing direction as a first direction parallel to the short side direction (the right direction in the drawing) to obtain a first processed image. Thereby, the left side of a solder film and a foreign material can be made to stand out. Next, an image processing step is performed with the processing direction as a second direction (downward in the drawing) parallel to the long side direction to obtain a second processed image. Thereby, the upper side of a foreign material can be made to stand out. Next, the third processing image is obtained by performing the image processing step with the processing direction as the third direction (upward direction in the drawing) opposite to the second direction. Thereby, the lower side of a foreign material can be made to stand out.

  Here, if the original image is simply binarized, the whole image expands, and an accurate narrow gap inspection cannot be performed. On the other hand, since a foreign material can be made conspicuous by combining the first to third processed images, an accurate narrow gap inspection can be performed.

  Next, an original image of the solder film is taken by the following method, a determination area is determined in the original image, and it is determined to be defective when the area of the area whose luminance is below the threshold in the determination area is a predetermined value or more. (Step S16). Thereby, the presence or absence of a black defect can be determined.

  FIG. 37 is a diagram showing an optical system for illuminating and photographing the solder film in the black defect inspection. This optical system has not only the objective lens 23, the coaxial illumination means 24, and the ring illumination means 25, but also a shadowless ring illumination means 28 as in the optical system of FIG. The shadowless ring illumination means 28 has a ring-shaped light guide material with a circular center located on the solder film 14. Light input to the light guide from a light source such as a light emitting diode is diffused in various directions inside the light guide. That is, the shadowless ring illumination means 28 is a diffusion uniform irradiation type shadowless illumination in which a shadow is not easily generated. The shadowless ring illumination means 28 is used to illuminate the solder film 14 from a plurality of directions as shown in FIG.

  First, the coaxial illumination unit 24, the ring illumination unit 25, and the shadowless ring illumination unit 28 are used to apply light to the solder film 14 from above and obliquely to measure the reflected light, thereby obtaining an original image as shown in FIG. In this way, the entire solder film 14 can be uniformly illuminated by irradiating the solder film 14 not only obliquely but also from above. Thereby, even a minor black defect can be detected. In particular, by using the shadowless ring illumination means 28, it is possible to irradiate the solder film 14 with light from multiple directions, so that black defects can be easily detected. If the solder film 14 is illuminated from a plurality of directions using direct illumination, the illumination approaches the center of the objective lens 23 and the field of view is narrowed. On the other hand, if the shadowless ring illumination means 28 is used, a wide visual field can be secured.

  Here, as shown in FIG. 40, the reflected light from the peripheral region of the work 29 is blocked by the shadowless ring illumination means 28 and is not incident on the objective lens 23. For this reason, as shown in FIG. 41, the luminance in the peripheral area of the original image is lower than that in the central area, and it is not possible to determine the black defect as it is. Therefore, as shown in FIG. 42, defect determination is performed after increasing the brightness of the peripheral area of the original image. As a result, it is possible to determine the black defect not only in the central region of the image but also in the peripheral region, so that the determination time can be shortened.

  Next, in the original image taken as described above, as shown in FIG. 43, an approximate curve (dotted line) of the outer periphery of the solder film is obtained by the method of least squares. A region surrounded by a range from the center of the Cu electrode to ½ of the long side and the approximate curve is determined as a black defect determination region. However, since the least square method includes an outer dark portion inside the approximate curve, it is preferable to set a region (solid line) on the inner side (solid line) by a predetermined width (for example, 3 μm) from the approximate curve as the determination region.

  Further, when it is determined that the black defect is defective in the black defect determination process, the defect determination process is performed again while enhancing the contrast of the original image. Thereby, it is possible to prevent the surface roughness of the solder film 14 from being erroneously determined as a black defect. The contrast of the original image is enhanced as follows.

  First, FIG. 44 shows a photographed original image, a luminance frequency distribution graph thereof, and a luminance cumulative frequency distribution graph. As shown in FIG. 45, gradation equalization is performed on the original image. Next, γ correction is performed as shown in FIG. Finally, the final image is obtained by improving the contrast as shown in FIG.

It is sectional drawing for demonstrating the manufacturing method of the semiconductor device which concerns on embodiment of this invention. It is sectional drawing for demonstrating the manufacturing method of the semiconductor device which concerns on embodiment of this invention. It is sectional drawing for demonstrating the manufacturing method of the semiconductor device which concerns on embodiment of this invention. It is sectional drawing for demonstrating the manufacturing method of the semiconductor device which concerns on embodiment of this invention. It is sectional drawing for demonstrating the manufacturing method of the semiconductor device which concerns on embodiment of this invention. It is sectional drawing for demonstrating the manufacturing method of the semiconductor device which concerns on embodiment of this invention. It is sectional drawing for demonstrating the manufacturing method of the semiconductor device which concerns on embodiment of this invention. It is sectional drawing for demonstrating the manufacturing method of the semiconductor device which concerns on embodiment of this invention. It is sectional drawing for demonstrating the manufacturing method of the semiconductor device which concerns on embodiment of this invention. It is a flowchart which shows the defect determination method which concerns on embodiment of this invention. It is a figure which shows the optical system for illuminating and image | photographing a solder film. It is a figure which shows the optical system for illuminating a solder film. It is sectional drawing which shows a mode that a solder film is illuminated by a ring illumination means. It is sectional drawing which shows a mode that a solder film is illuminated by a coaxial illumination means. It is an original image of a solder film having a DC defect. This is differential data obtained by differentiating the luminance of the original image in the long side direction. It is a figure which shows the integrated value which integrated the differential data in the short side direction. It is an original image of a solder film having a DC defect. It is an original image of a solder film having a DE defect. It is a figure which shows the integrated value which integrated the differential data in the short side direction. It is a figure for demonstrating differentiating a brightness | luminance only about the center part of Cu electrode. It is a figure which shows the integrated value which integrated the brightness | luminance of the original image of FIG. 19 in the short side direction. It is the original image which drew the approximated curve of the outer peripheral part of a solder film. It is a laser microscope image of the solder film which has DE defect. It is the original image obtained by irradiating light to the solder film which has DE defect from diagonally, and measuring reflected light. It is the image which binarized the original image of FIG. It is a figure for demonstrating the multi-value test | inspection of DE failure. It is a figure for demonstrating the inspection of a dewet. It is a figure for demonstrating the binary area test | inspection of DE failure. It is a figure which shows the result of the height measurement by a laser displacement meter. It is sectional drawing which shows the position which performs a laser height measurement. It is a top view which shows the position which performs a laser height measurement. It is a top view which shows the gap between leads. It is a figure for demonstrating 1st image processing. It is a figure for demonstrating 2nd image processing. It is a figure for demonstrating the synthesis | combination of a 1st-3rd process image. It is a figure which shows the optical system for illuminating and image | photographing a solder film in a black defect test | inspection. It is sectional drawing which shows a mode that a solder film is illuminated by a shadowless ring illumination means. It is the original image obtained by applying light to the solder film from above and obliquely and measuring the reflected light. It is an expanded sectional view for demonstrating the path | route of the reflected light at the time of using a shadowless ring illumination means. It is the original image which image | photographed the solder film formed on the some Cu wiring. It is the image which raised the brightness | luminance of the peripheral area | region of the original image of FIG. This is an original image in which a black defect determination area is determined. It is the image | photographed original image, the frequency distribution graph of the brightness | luminance, and the cumulative frequency distribution graph of a brightness | luminance. It is the image which performed the gradation equalization, the frequency distribution graph of the brightness | luminance, and the cumulative frequency distribution graph of the brightness | luminance. It is the image which performed (gamma) correction | amendment, the frequency distribution graph of the brightness | luminance, and the cumulative frequency distribution graph of the brightness | luminance. It is the image which improved contrast, the frequency distribution graph of the brightness | luminance, and the cumulative frequency distribution graph of the brightness | luminance. It is the image which image | photographed the solder film which has DC defect. It is the image which image | photographed the solder film which has DE defect. It is a laser microscope image which shows the solder film which black defect generate | occur | produced. It is the original image obtained by the conventional method.

Explanation of symbols

11 Wiring board 12 Cu electrode (electrode)
14 Solder film 24 Coaxial illumination means 25 Ring illumination means 26 Cu lead (lead)
28 Shadowless ring illumination means

Claims (17)

Preparing a wiring board having a solder film on an electrode having a long side and a short side;
Applying an oblique light to the solder film and measuring the reflected light to obtain an original image;
Differentiating the luminance of the original image in the long side direction to obtain differential data;
A step of integrating the differential data in the short side direction to obtain an integrated value;
A first inspection for determining a defect when a difference between the maximum value and the minimum value of the integrated values is equal to or greater than a predetermined value, and a width between a position where the integrated value is maximum and a minimum position is equal to or less than a predetermined width. A method of manufacturing a semiconductor device.   When it is determined as defective in the first inspection process, it is determined as defective when the original image is divided into two or more regions whose luminance is lower than the threshold on the electrode, and is determined as good when the number is one or less. The method of manufacturing a semiconductor device according to claim 1, further comprising a second inspection step for determining. Preparing a wiring board having a solder film on an electrode having a long side and a short side;
Applying an oblique light to the solder film and measuring the reflected light to obtain an original image;
Differentiating the luminance of the original image in the long side direction to obtain differential data;
Integrating the luminance of the differential data in the short side direction to obtain an integrated value;
And a step of determining a failure when the maximum value of the integrated values is equal to or greater than a predetermined value.   4. The method of manufacturing a semiconductor device according to claim 1, wherein, in the step of obtaining the differential data, the luminance is differentiated only with respect to a middle portion of the electrode. 5. Preparing a wiring board having a solder film on the electrode;
Applying an oblique light to the solder film and measuring the reflected light to obtain an original image;
A method for manufacturing a semiconductor device, comprising: a first inspection step for determining a defect when an area of a region having a luminance equal to or higher than a threshold in the original image is equal to or less than a predetermined value.   6. The method of manufacturing a semiconductor device according to claim 5, further comprising a second inspection step of measuring a height of the solder film determined to be defective in the first inspection step with a laser displacement meter.   In the second inspection step, if the difference between the height of the solder film on the electrode and the height of the lead drawn out from the electrode is out of a predetermined value range, it is determined as defective. Item 7. A method for manufacturing a semiconductor device according to Item 6. Preparing a wiring board having a solder film on the electrode;
Applying an oblique light to the solder film and measuring the reflected light to obtain an original image;
For the original image, performing a process of raising the luminance of the portion corresponding to the lead extracted from the electrode to a threshold value or more;
A method of manufacturing a semiconductor device, comprising: determining a defect when an interval between regions having a luminance equal to or higher than a threshold in a processed image is equal to or less than a predetermined length. Preparing a wiring board having a solder film on the electrode;
Applying an oblique light to the solder film and measuring the reflected light to obtain an original image;
For each pixel of the original image, the luminance of the predetermined number of pixels in the predetermined processing direction from the pixel is integrated, and the predetermined number of pixels from the pixel in the direction opposite to the processing direction An image processing step for performing a process of setting a value obtained by subtracting the sum of luminance to the luminance of the pixel;
A method for manufacturing a semiconductor device, comprising: an inspection step of determining a defect when an interval between regions having luminance equal to or higher than a threshold value is equal to or less than a predetermined length in a processed image. Performing the image processing step with the processing direction as a first direction parallel to the short side direction to obtain a first processed image;
Performing the image processing step with the processing direction as a second direction parallel to the long side direction to obtain a second processed image;
Performing the image processing step with the processing direction as a third direction opposite to the second direction to obtain a third processed image;
The method for manufacturing a semiconductor device according to claim 9, wherein the first processed image, the second processed image, and the third processed image are combined, and the inspection process is performed on the combined image. Preparing a wiring board having a solder film on the electrode;
Applying the light from above and obliquely to the solder film to measure reflected light to obtain an original image;
A semiconductor device manufacturing method, comprising: determining a determination region in the original image; and determining a failure when an area of a region having a luminance of a threshold value or less in the determination region is equal to or greater than a predetermined value. Method.   In the step of obtaining the original image, a light input means for inputting the reflected light, a coaxial illumination means for illuminating the solder film coaxially with the light input means, and a ring shape in which a circular center is located on the solder film 12. The method of manufacturing a semiconductor device according to claim 11, wherein the ring illumination means is used.   13. The method of manufacturing a semiconductor device according to claim 12, wherein in the step of obtaining the original image, a shadowless ring illuminating unit that diffuses light from a light source and illuminates the solder film from a plurality of directions is further used. .   The method of manufacturing a semiconductor device according to claim 13, wherein the defect determination step is performed after increasing the luminance of a peripheral area of the original image. A step of obtaining an approximate curve of the outer periphery of the solder film in the original image by a least square method;
The method for manufacturing a semiconductor device according to claim 11, wherein a region surrounded by the approximate curve is set as the determination region.   16. The method of manufacturing a semiconductor device according to claim 15, wherein an area inside the predetermined curve by a predetermined width is set as the determination area.   17. The semiconductor device manufacturing according to claim 11, wherein when the defect is determined to be defective in the defect determination step, the defect determination step is performed again with emphasis on the contrast of the original image. Method.

Images