Method and apparatus for inspection of the appearance of bumps

• US 5,761,337 A
• Filed: 08/05/1996
• Issued: 06/02/1998
• Est. Priority Date: 05/13/1993
• Status: Expired due to Fees

• Current Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Original Assignee: Sharp Kabushiki Kaisha (Hon Hai Precision Industry Co., Ltd.)
• Inventors: Shirouchi, Yuichi, Nishimura, Hidetoshi
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): KELLEY, CHRISTOPHER S
• Application Number:
  US08/692,153
• Time in Patent Office: 666 Days
• Field of Search: 382/149, 382/150, 382/151, 382/108, 348/109, 348/126, 356/237, 356/374, 356/376
• US Class Current: 382/150
• CPC Class Codes: G01N 21/95684 Patterns showing highly reflecting parts, e.g. metallic elements
• CPC Class Codes: G01R 31/303 of integrated circuits G01R31/305 - G01R31/315 take precedence
• CPC Class Codes: H01L 24/11 Manufacturing methods for bumps on insulating substrates H01L21/4853
• CPC Class Codes: H01L 24/13 of an individual bump connector

Abstract

A semiconductor chip with bumps to be inspected is placed directly below a CCD, which captures an image of the bumps under dark field illumination. A reference window is set around each bump in the image. The reference window is enlarged or reduced or divided into subwindows such that optimal windows are provided for extraction of characteristic values of various defects. Characteristic values of particular defects are extracted from the image data within the corresponding optimal windows. The characteristic values for the bumps are statistically processed to obtain an evaluation value having a permissible range. The characteristic value of each bump is compared with the evaluation value and when the comparison indicates that the characteristic value is outside of the range, the bump is decided as a defective bump.

Claims

1. A bump appearance inspection method for inspecting appearance of bumps formed at respective pad electrodes on a semiconductor chip, comprising the steps of:

(a) illuminating a surface of the semiconductor chip with the bumps at a given angle to the surface of the semiconductor chip from all sides of the semiconductor chip;

(b) capturing an image of the surface of the semiconductor chip with the bumps using light reflected by the surface;

(c) storing image data of the captured image in an image memory;

(d) defining for each bump a reference window along an outer edge of the bump in the image;

(e) recognizing for each bump based on the image data a higher brightness, ring-shaped bump region at a periphery of the bump and a lower brightness bump region at a center of the bump enclosed by the higher brightness, ring-shaped bump region;

(f) defining for each bump a new window within the reference window on the basis of a boundary between the higher brightness bump region and the lower brightness bump region;

(g) extracting the characteristic value of each bump from the image data within the new window;

(h) obtaining an evaluation value having a permissible range for the new window based on an average value and a dispersion value determined using of the characteristic values of the bumps within the respective new windows;

(i) comparing the characteristic value of each bump within the new window with the evaluation value for the new window; and

(j) considering the bump to be a defective bump when the comparison result indicates that the characteristic value of the bump is outside of the range of the evaluation value.

2. The bump appearance inspection method according to claim 1, wherein the step (f) comprises the substeps of:

defining a lower brightness window which defines an inner edge of the ring-shape of the higher brightness bump region and which surrounds the center region of the bump; and

dividing the lower brightness window into a plurality of subwindows each of which serves as the new window

wherein the characteristic values in the image are extracted from the subwindows, and wherein the evaluation value for each new window is calculated from the characteristic values extracted from the image data within the subwindows located at same positions in the respective lower brightness windows for the bumps such that effects of flaws or surface roughness of the lower brightness region of each bump are permitted within a certain range.

3. The bump appearance inspection method according to claim 2, further comprising the steps of:

selecting, as candidates for a defective bump, bumps considered to be defective bumps at the step (j);

defining the subwindows to include a first subwindow formed by masking a center portion of the lower brightness window and a second subwindow occupying the center portion of the lower brightness window;

thresholding the image data within the first and second subwindows into binary image data;

extracting a first characteristic value from the binary image data within the first subwindow and a second characteristic value from the binary image data within the second window for each defective bump candidate;

calculating a characteristic value of each defective bump candidate attributable to a defect in the recess portions of the bumps, using the extracted first and second characteristic values;

obtaining an evaluation through statistical processing of the characteristic values of the defective bump candidates, the evaluation value having a permissible range;

comparing the characteristic value of each defective bump candidate with the evaluation value; and

determining that the defective bump candidate is a defective bump when its characteristic value is outside of the permissible range of the evaluation value,

wherein a defect in the recess portion of the bump is detected without detecting as a defect an effect of surface roughness or a shape of the bump within a certain range.

4. The bump appearance inspection method according to claim 3, further comprising the steps of:

calculating an average value of the obtained characteristic values of the defective bump candidates;

comparing the calculated average value with a predetermined evaluation value; and

deciding, when the average value exceeds the evaluation value, that a plurality of bumps are defective,

wherein an abnormal peripheral edge upheaval effecting a plurality of bumps due to a current density abnormality during bump formation is detected.

5. The bump appearance inspection method according to claim 2, further comprising the steps of:

selecting, as candidates for a defective bump, bumps considered to be defective bumps at the step (j);

thresholding the image data of each defective bump candidate into binary image data and extracting the lower brightness region from the binary image data to obtain an outline of the lower brightness region;

calculating a characteristic value indicative of a shape of the lower brightness region based on the outline obtained;

comparing the calculated characteristic value with a predetermined evaluation value; and

deciding, when the characteristic value exceeds the evaluation value, that the defective bump candidate is a defective bump,

wherein a defective recess in a peripheral portion of the bump is detected.

6. The bump appearance inspection method according to claim 3, further comprising the steps of:

thresholding the image data of each defective bump candidate into binary image data and extracting the lower brightness region from the binary image data to obtain an outline of the lower brightness region;

calculating a characteristic value indicative of a shape of the lower brightness region based on the outline obtained;

comparing the calculated characteristic value with a predetermined evaluation value; and

deciding, when the characteristic value exceeds the evaluation value, that the defective bump candidate is a defective bump,

wherein a defective recess in a peripheral portion of the bump is detected.

7. The bump appearance inspection method according to claim 4, further comprising the steps of:

selecting, as candidates for a defective bump, bumps considered to be bumps at the step (j);

thresholding the image data of each defective bump candidate into binary image data and extracting the lower brightness region from the binary image data to obtain an outline of the lower brightness region;

calculating a characteristic value indicative of a shape of the lower brightness region based on the outline obtained;

comparing the calculated characteristic value with a predetermined evaluation value; and

deciding, when the characteristic value exceeds the evaluation value, that the defective bump candidate is a defective bump,

wherein a defective recess in a peripheral portion of the bump is detected.

8. The bump appearance inspection method of claim 2, further comprising the steps of:

selecting, as candidates for a defective bump, bumps considered to be bumps at the step (j);

enlarging the reference window for each candidate at a given magnification;

thresholding the image data of each defective bump candidate within the enlarged reference window into binary image data;

extracting the higher brightness region corresponding to the peripheral portion of the defective bump candidate from the binary image data to obtain an outline of the higher brightness region;

calculating a characteristic value indicative of a shape of the higher brightness region based on the outline obtained;

comparing the calculated characteristic value with a predetermined evaluation value; and

deciding, when the characteristic value exceeds the evaluation value, that the defective bump candidate is a defective bump,

wherein an abnormal shape of the bump is detected.

9. The bump appearance inspection method of claim 3, further comprising the steps of:

selecting, as candidates for a defective bump, bumps considered to be defective bumps at the step (j);

enlarging the reference window for each defective bump candidate at a given magnification;

thresholding the image data of each defective bump candidate within the enlarged reference window into binary image data;

extracting the higher brightness region corresponding to the peripheral portion of the defective bump candidate from the binary image data to obtain an outline of the higher brightness region;

calculating a characteristic value indicative of a shape of the higher brightness region based on the outline obtained;

comparing the calculated characteristic value with a predetermined evaluation value; and

deciding, when the characteristic value exceeds the evaluation value, that the defective bump candidate is a defective bump,

wherein an abnormal shape of the bump is detected.

10. The bump appearance inspection method of claim 4, further comprising the steps of:

selecting, as candidates for a defective bump, bumps considered to be defective bumps at the step (j);

enlarging the reference window for each candidate at a given magnification;

thresholding the image data of each candidate within the enlarged reference window into binary image data;

extracting the higher brightness region corresponding to the peripheral portion of the candidate from the binary image data to obtain an outline of the higher brightness region;

calculating a characteristic value indicative of a shape of the higher brightness region based on the outline obtained;

comparing the calculated characteristic value with a predetermined evaluation value; and

deciding, when the characteristic value exceeds the evaluation value, that the defective bump candidate is a defective bump,

wherein an abnormal shape of the bump is detected.

11. The bump appearance inspection method of claim 5, further comprising the steps of:

selecting, as candidates for a defective bump, bumps considered to be defective bumps at the step (j);

enlarging the reference window for each defective bump candidate at a given magnification;

thresholding the image data of each defective bump candidate within the enlarged reference window into binary image data;

extracting the higher brightness region corresponding to the peripheral portion of the candidate from the binary image data to obtain an outline of the higher brightness region;

calculating a characteristic value indicative of a shape of the higher brightness region based on the outline obtained;

comparing the calculated characteristic value with a predetermined evaluation value; and

deciding, when the characteristic value exceeds the evaluation value, that the defective bump candidate is a defective bump,

wherein an abnormal shape of the bump is detected.

12. The bump appearance inspection method of claim 6, further comprising the steps of:

selecting, as candidates for a defective bump, bumps considered to be defective bumps at the step (j);

enlarging the reference window for each defective bump candidate at a given magnification;

thresholding the image data of each candidate within the enlarged reference window into binary image data;

extracting the higher brightness region corresponding to the peripheral portion of the candidate from the binary image data to obtain an outline of the higher brightness region;

calculating a characteristic value indicative of a shape of the higher brightness region based on the outline obtained;

comparing the calculated characteristic value with a predetermined evaluation value; and

deciding, when the characteristic value exceeds the evaluation value, that the defective bump candidate is a defective bump,

wherein an abnormal shape of the bump is detected.

13. The bump appearance inspection method of claim 7, further comprising the steps of:

selecting, as candidates for a defective bump, bumps considered to be defective bumps at the step (j);

enlarging the reference window for each defective bump candidate at a given magnification;

thresholding the image data of each defective bump candidate within the enlarged reference window into binary image data;

extracting the higher brightness region corresponding to the peripheral portion of the candidate from the binary image data to obtain an outline of the higher brightness region;

calculating a characteristic value indicative of a shape of the higher brightness region based on the outline obtained;

comparing the calculated characteristic value with a predetermined evaluation value; and

deciding, when the characteristic value exceeds the evaluation value, that the defective bump candidate is a defective bump,

wherein an abnormal shape of the bump is detected.

14. The bump appearance inspection method according to claim 1, further comprising the steps of:

thresholding image data of each bump within the reference window into binary image data;

determining a width of each bump based on the binary image data;

calculating a height of the bump based on the width of the bump;

calculating an evaluation value with a permissible range through application of a statistical processing to the heights of the bumps;

comparing the height of each bump with the evaluation value; and

determining that the bump is defective with respect to its height when a value of the height is outside of the range of the evaluation value.

15. A bump appearance inspection apparatus for detecting defective bumps, comprising:

a support for supporting a semiconductor chip with bumps formed at its respective pad electrode sites;

image capture means for capturing an image of a surface of the semiconductor chip with the bumps using light reflected by the surface;

a drive for moving the support such that the bumps of the semiconductor chip on the support are positioned directly below the image capture means;

illumination for illuminating the surface of the semiconductor chip with the bumps at a given angle to the surface of the semiconductor chip from its entire periphery;

a memory for storing image data of the image captured by the capture means;

data process circuitry for performing the following tasks;

defining a reference window along an outer edge of each bump in the image recognizing for each bump based on the image data a higher brightness, ring-shaped bump region at a periphery of the bump and a lower brightness bump region at a center of the bump enclosed by the higher brightness, ring-shaped bump region, wherein the higher brightness, ring-shaped region in the image corresponds to a peripheral, protruded portion of the bump and the low brightness bump region in the image corresponds to a recess portion of the bump;

defining a new window within the reference window according to a surface state of the bump so that characteristic values of the bumps attributable to a defect in the recess portion can be accurately extracted;

extracting the characteristic values of the bumps attributable to the defect from the image data within the new window;

calculating an obtaining value having a permissible range based on an average value and a dispersion value calculated using the extracted characteristic values of the bumps; and

comparing the characteristic value of each bump with the calculated evaluation value and discriminating based on the comparison result whether the bump is defective or not and deciding that the bump is a defective bump when its characteristic value is outside of the permissible range.