Automated wafer defect inspection system and a process of performing such inspection

• US 6,324,298 B1
• Filed: 07/13/1999
• Issued: 11/27/2001
• Est. Priority Date: 07/15/1998
• Status: Expired due to Term

• Current Assignee: Rudolph Technologies Incorporated (Onto Innovation Inc.)
• Original Assignee: August Technology Corporation (Onto Innovation Inc.)
• Inventors: OʼDell, Jeffrey, Verburgt, Thomas, Harless, Mark
• Primary Examiner(s): Boudreau, Leo
• Assistant Examiner(s): Werner, Brian P.
• Application Number:
  US09/352,564
• Time in Patent Office: 868 Days
• Field of Search: 382/145, 382/148, 382/149, 382/217, 382/218, 382/228, 348/87, 348/126, 356/237.4, 356/237.5, 250/559.34, 250/559.4, 250/559.41, 250/559.45, 250/559.46
• US Class Current: 382/149
• CPC Class Codes: G01N 2021/8838 Stroboscopic illumination; synchronised illumination
• CPC Class Codes: G01N 21/8803 Visual inspection measuring projectors G01B9/08
• CPC Class Codes: G01N 21/8806 Specially adapted optical and illumination features
• CPC Class Codes: G01N 21/9501 Semiconductor wafers manufacturing processes per se of semiconductor devices implementing a measuring step H01L22/10
• CPC Class Codes: G01N 2201/06113 Coherent sources; lasers
• CPC Class Codes: G01N 2201/0696 Pulsed
• CPC Class Codes: G01N 2201/0697 Pulsed lasers
• CPC Class Codes: G01N 2201/12 Circuits of general importance; Signal processing
• CPC Class Codes: G06T 2207/30148 Semiconductor; IC; Wafer
• CPC Class Codes: G06T 7/0004 Industrial image inspection
• CPC Class Codes: H01L 21/67242 Apparatus for monitoring, sorting or marking testing or measuring during manufacture H01L22/00, marks per se H01L23/544; testing individual semiconductor devices G01R31/26
• CPC Class Codes: H01L 21/67271 Sorting devices
• CPC Class Codes: H01L 21/67288 Monitoring of warpage, curvature, damage, defects or the like
• CPC Class Codes: H01L 22/12 for structural parameters, e.g. thickness, line width, refractive index, temperature, warp, bond strength, defects, optical inspection, electrical measurement of structural dimensions, metallurgic measurement of diffusions electrical measurement of diffusions H01L22/14
• CPC Class Codes: H01L 2924/00 Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
• CPC Class Codes: H01L 2924/0002 Not covered by any one of groups H01L24/00, H01L24/00 and H01L2224/00
• CPC Class Codes: H04N 23/56 provided with illuminating means

Abstract

An automated defect inspection system has been invented and is used on patterned wafers, whole wafers, broken wafers, partial wafers, sawn wafers such as on film frames, JEDEC trays, Auer boats, die in gel or waffle packs, MCMs, etc. and is specifically intended and designed for second optical wafer inspection for such defects as metalization defects (such as scratches, voids, corrosion, and bridging), diffusion defects, passivation layer defects, scribing defects, glassivation defects, chips and cracks from sawing, solder bump defects, and bond pad area defects.

Claims

1. An automated method of inspecting a semiconductor wafer in any form including whole patterned wafers, sawn wafers, broken wafers, and wafers of any kind on film frames, dies, die in gel paks, die in waffle paks, multi-chip modules often called MCMs, JEDEC trays, Auer boats, and other wafer and die package configurations for defects, the method comprising:

training a model as to parameters of a good wafer via optical viewing of multiple known good wafers, the training including aligning and viewing the multiple known good wafers, taking a gray scale measurement of each pixel in a grid of pixels on each of the multiple known good wafers, calculating the mean and standard deviation of all of the gray scale measurements for each pixel location on the multiple known good wafers, and then determining an upper and lower limit for each pixel location in between which a gray scale measurement on an equivalent pixel on an unknown quality wafer is deemed good while a gray scale measurement on an equivalent pixel on an unknown quality wafer outside of the upper and lower limits is deemed of at least questionable integrity and so noted, the upper limit being calculated as the maximum of one of the mean plus the product of the standard deviation multiplied by a first sensitivity, or the mean plus a first minimum contrast, and the lower limit being calculated as the minimum of one of the mean minus the product of the standard deviation multiplied by a second sensitivity, or the mean plus a second minimum contrast; and

inspecting unknown quality wafers using the model, where such inspecting involves calculating the gray scale measurement for each pixel on the unknown quality wafer and comparing the gray scale measurement of each pixel with the upper and lower limit for that given pixel location to determine if the gray scale measurement for that pixel is between the upper and lower limits and thus deemed good or is outside of a range between the upper and lower limits and thus deemed of at least questionable integrity.

2. The automated inspection method of claim 1 wherein the training a model step includes aligning and viewing via a camera the multiple known good wafers.

3. The automated inspection method of claim 1 wherein the training model step further includes taking the gray scale measurement of each pixel on a 0 to 255 gray scale.

4. The automated inspection method of claim 1 further comprising training a system as to inspection parameters which are values that indicate how many standard deviations away from the mean are acceptable in defining the upper and lower limits.

5. The automated inspection method of claim 1 further comprising training a system as to alignment features and parameters of wafers.

6. The automated inspection method of claim 1 further comprising training a system as to at least one of the following:

wafer geometry, wafer size, die geometry on a wafer, die size on a wafer, number of rows of die on a wafer, and number of columns of die on a wafer.

7. The automated inspection method of claim 1 further comprising training a system as to inspection parameters.

8. The automated inspection method of claim 1 further comprising creating an inspection recipe that is used during inspecting to perform at least one of the following:

defining how substrates are selected from storage receptacles, defining how dies on a wafer are selected for defect inspection, and defining how an inspection map is imported or exported, if needed.

9. The method of claim 1 further comprising defect reviewing following the completion of inspecting whereby only die failing to meet the good wafer model are reviewed.

10. An automated method of inspecting a die on a substrate such as a wafer in any form including whole patterned wafers, sawn wafers, broken wafers, and wafers of any kind on film frames, dies, die in gel paks, die in waffle paks, multi-chip modules often called MCMs, JEDEC trays, Auer boats, and other wafer and die package configurations for defects, the method comprising:

training a model as to parameters of a good die via optical viewing of multiple known good die, the training including aligning and viewing the multiple known good die, taking a gray scale measurement of each pixel in a grid of pixels on each of the multiple known good die, calculating the mean and standard deviation of all of the gray scale measurements for each pixel location on the multiple known good die, and then determining an upper and lower limit for each pixel location in between which a gray scale measurement on an equivalent pixel on an unknown quality die is deemed good while a gray scale measurement on an equivalent pixel on an unknown quality die outside of the upper and lower limits is deemed of at least questionable integrity and so noted, the upper limit being calculated as the maximum of one of the mean plus the product of the standard deviation multiplied by a first sensitivity, or the mean plus a first minimum contrast, and the lower limit being calculated as the minimum of one of the mean minus the product of the standard deviation multiplied by a second sensitivity, or the mean plus a second minimum contrast; and

inspecting unknown quality die using the model, where such inspecting involves calculating the gray scale measurement for each pixel on the unknown quality die and comparing the gray scale measurement of each pixel with the upper and lower limit for that given pixel location to determine if the gray scale measurement for that pixel is between the upper and lower limits and thus deemed good or is outside of a range between the upper and lower limits and thus deemed of at least questionable integrity.

11. The automated inspection method of claim 10 wherein the aligning and viewing is performed via a camera.

12. The automated inspection method of claim 10 wherein the training model step further includes taking the gray scale measurement of each pixel on a 0 to 255 gray scale.

13. The automated inspection method of claim 10 further comprising training a system as to inspection parameters which are values that indicate how many standard deviations away from the mean are acceptable in defining the upper and lower limits.

14. An automated method of inspecting a substrate such as a wafer in any form including bare wafers, whole wafers, patterned wafers, sawn wafers, and wafers of any kind on film frames, dies, die in gel paks, die in waffle paks, multi-chip modules often called MCMs, JEDEC trays, Auer boats, and other water and die package configurations for defects, the method comprising:

determining an upper and a lower limit for each pixel on a substrate divided into a grid of pixels, the upper and lower limits for each pixel being calculated as a relationship of a plurality of visually perceived contrasts for that pixel on a plurality of substrates known as good quality substrates and specifically as the upper limit is a maximum of one of a mean of the visually perceived contrast for a given pixel plus a product of a standard deviation of the visually perceived contrast for the given pixel multiplied by a first sensitivity, or the mean for the given pixel plus a first minimum contrast for the given pixel, and the lower limit is a minimum of one of the mean of the visually perceived contrast for a given pixel plus a product of the standard deviation of the visually perceived contrast for the given pixel multiplied by a second sensitivity, or the mean for the given pixel plus a second minimum contrast for the given pixel;

comparing a visually perceived contrast, for each pixel from an unknown quality substrate with the upper and lower limit for that pixel; and

categorizing each compared pixel as either good if Its visually perceived contrast is between the upper and lower limit, or bad if its visually perceived contrast is outside of the upper and lower limit.

15. An automated method of inspecting a substrate such as a wafer in any form including bare wafers, whole wafers, patterned wafers, sawn wafers, and wafers of any kind on film frames, dies, die in gel paks, die in waffle paks, multi-chip modules often called MCMs, JEDEC trays, Auer boats, and other wafer and die package configurations for defects, the method comprising:

determining an upper and a lower limit for each pixel on a substrate divided into a grid of pixels, the upper and lower limits for each pixel being calculated as a relationship of a gray scale measurement for that pixel on a plurality of substrates known as good quality substrates, and specifically as the upper limit is a maximum of one of a mean of the gray scale measurement for a given pixel plus a product of a standard deviation of the gray scale measurement for the given pixel multiplied by a first sensitivity, or the mean for the given pixel plus a first minimum contrast for the given pixel, and the lower limit is a minimum of one of the mean of the gray scale measurement for a given pixel plus a product of the standard deviation of the gray. scale measurement for the given pixel multiplied by a second sensitivity, or the mean for the given pixel plus a second minimum contrast for the given pixel;

comparing a gray scale measurement for each pixel from an unknown quality substrate with the upper and lower limit for that pixel; and

categorizing each compared pixel as either good if its gray scale measurement is between the upper and lower limit, or bad if its gray scale measurement is outside of the upper and lower limit.

16. An automated method of inspecting a substrate such as a wafer in any form including bare wafers, whole wafers, patterned wafers, sawn wafers, and wafers of any kind on film frames, dies, die in gel paks, die in waffle paks, multi-chip modules often called MCMs, JEDEC trays, Auer boats, and other wafer and die package configurations for defects, the method comprising:

determining an upper and a lower limit for each pixel on a substrate divided into a grid of pixels, the upper and lower limits for each pixel being calculated as a relationship of a color measurement for that pixel on a plurality of substrates known as good quality substrates, and specifically as the upper limit is a maximum of one of a mean of the color measurement for a given pixel plus a product of a standard deviation of the color measurement for the given pixel multiplied by a first sensitivity, or the mean for the given pixel plus a first minimum contrast for the given pixel, and the lower limit is a minimum of one of the mean of the color measurement for a given pixel plus a product of the standard deviation of the color measurement for the given pixel multiplied by a second sensitivity, or the mean for the given pixel plus a second minimum contrast for the given pixel;

comparing a color measurement for each pixel from an unknown quality substrate with the upper and lower limit for that pixel; and

categorizing each compared pixel as either good if its color measurement is between the upper and lower limit, or bad if its color measurement is outside of the upper and lower limit.

17. An automated method of inspecting a substrate such as a wafer in any form including bare wafers, whole wafers, patterned wafers, sawn wafers, and wafers of any kind on film frames, dies, die in gel paks, die in waffle paks, multi-chip modules often called MCMs, JEDEC trays, Auer boats, and other wafer and die package configurations for defects, the method comprising:

determining a gray scale measurement for each pixel in a plurality of pixels on a known good quality substrate;

repeating the determining a gray scale measurement step for a plurality of known good quality substrates;

determining an upper and a lower limit from the gray scale measurements for each pixel, the upper limit being determined by calculating a maximum of one of a mean of the gray scale measurement for a given pixel plus a product of a standard deviation of the gray scale measurement for the given pixel multiplied by a first sensitivity, or the mean for the given pixel plus a first minimum contrast for the given pixel, and the lower limit being determined by calculating a minimum of one of the mean of the gray scale measurement for a given pixel plus a product of the standard deviation of the gray scale measurement for the given pixel multiplied by a second sensitivity, or the mean for the given pixel plus a second minimum contrast for the given pixel;

comparing each pixel from an unknown quality substrate with the upper and lower limit for that given pixel location; and

categorizing each compared pixel as either good if its gray scale measurement is between the upper and lower limit, or bad if its gray scale measurement is outside of the upper and lower limit.

18. An automated system for inspecting a substrate such as a wafer in any form including whole patterned wafers, sawn wafers, broken wafers, and wafers of any kind on film frames, dies, die in gel paks, die in waffle paks, multi-chip modules often called MCMs, JEDEC trays, Auer boats, and other wafer and die package configurations for defects, the system comprising:

a wafer test plate;

means for providing a wafer to the test plate;

a wafer alignment device for aligning each and every wafer provided to the test plate at the same x, y, z, and θ

location;

a visual inspection device for visual inputting of a plurality of known good quality wafers during training and for visual Inspection of other unknown quality wafers during inspection,

a microprocessor having processing and memory capabilities for developing a model of a good quality wafer by taking a gray scale measurement for each said pixel on a plurality of known good quality wafers and determining an upper and a lower limit for each pixel from the plurality of known good quality wafers, the upper limit being a maximum of one of a mean of the gray scale measurement for a given pixel plus a product of a standard deviation of the gray scale measurement for the given pixel multiplied by a first sensitivity, or the mean for the given pixel plus a first minimum contrast for the given pixel, and the lower limit being a minimum of one of the mean of the gray scale measurement for a given pixel plus a product of the standard deviation of the gray scale measurement for the given pixel multiplied by a second sensitivity, or the mean for the given pixel plus a second minimum contrast for the given pixel, and the microprocessor further comparing other unknown quality wafers in comparison to the model by determining a gray scale measurement for each pixel thereon and comparing it to the upper and lower limits.

19. The automated system of claim 18 further comprising a focusing mechanism for focusing an image of at least a portion of the wafer as seen by the camera.

20. The automated system of claim 18 where the visual inspection device is a camera.

21. The automated system of claim 18 where the visual inspection device is a CCD camera.

22. The automated system of claim 18 further comprising a parameter input device for inputting parameters and other constraints or information including sensitivity parameters, geometries, die size, die shape, die pitch, number of rows, and number of columns.

23. The automated system of claim 18 further comprising a display for displaying the view being seen by the visual inspection device at present or at any previous saved period.

24. The automated system of claim 18 further including a marking head for placing a mark on or adjacent a pixel on a wafer that is outside of the model.

25. The automated system of claim 18 further including a frame on which the wafer test plate, the wafer alignment device, the visual inspection device, and the computer system are attached, and a hood for enclosing an inspection area over the test plate.

26. An automated system for inspecting a substrate such as a wafer in any form including bare wafers, whole wafers, patterned wafers, sawn wafers, and wafers of any kind on film frames, dies, die in gel paks, die in waffle paks, multi-chip modules often called MCMs, JEDEC trays, Auer boats, and other wafer and die package configurations for defects, the system comprising:

a wafer test plate;

means for providing a wafer to the test plate;

a wafer alignment device for aligning each and every wafer provided to the test plate at the same x, y, z, and θ

location;

a visual inspection device for visual inputting of a plurality of known good quality wafers during training and for visual inspection of other unknown quality wafers during inspection,

a microprocessor having processing and memory capabilities for developing a model of a good quality wafer by taking a visually perceived contrast for each said pixel on a plurality of known good quality wafers and determining an upper and a lower limit for each pixel from the plurality of known good quality wafers, the upper limit being a maximum of one of a mean of the visually perceived contrast for a given pixel plus a product of a standard deviation of the visually perceived contrast for the given pixel multiplied by a first sensitivity, or the mean for the given pixel plus a first minimum contrast for the given pixel, and the lower limit being a minimum of one of the mean of the visually perceived contrast for a given pixel plus a product of the standard deviation of the visually perceived contrast for the given pixel multiplied by a second sensitivity, or the mean for the given pixel plus a second minimum contrast for the given pixel, and the microprocessor further comparing other unknown quality wafers in comparison to the model by determining a visually perceived contrast for each pixel thereon and comparing it to the upper and lower limits.