Method for characterizing defects on semiconductor wafers

• US 5,808,735 A
• Filed: 11/26/1996
• Issued: 09/15/1998
• Est. Priority Date: 06/17/1993
• Status: Expired due to Term

• Current Assignee: Ultrapointe Corp. (Lumentum Holdings Inc.)
• Original Assignee: Ultrapointe Corp. (Lumentum Holdings Inc.)
• Inventors: Han, Ke, Srinivasan, Lakshman, Lee, Ken K., Worster, Bruce W.
• Primary Examiner(s): Font, Frank G.
• Assistant Examiner(s): Ratliff, Reginald A.
• Application Number:
  US08/756,420
• Time in Patent Office: 658 Days
• Field of Search: 356/237, 356/369, 250/559.42, 250/559.48
• US Class Current: 356/237.2
• 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 21/95607 using a comparative method
• CPC Class Codes: G01R 31/311 of integrated circuits G01R31/31728 takes precedence
• CPC Class Codes: G02B 21/006 focusing arrangements; selection of the plane to be imaged

Abstract

10A method is described for detecting and characterizing defects on a test surface of a semiconductor wafer. A three-dimensional image of the test surface is aligned and compared with a three-dimensional image of a defect-free reference surface. Intensity differences between corresponding pixels in the two images that exceed a predefined threshold value are deemed defect pixels. According to the method, the pixels of the reference image are grouped according to their respective z values (elevation) to identify different physical layers of the reference surface. Because different surface layers can have different image properties, such as reflectance and image texture, the groups of pixels are analyzed separately to determine an optimal threshold value for each of the groups, and therefore for each layer of the reference surface.

Claims

1. A method comprising:

representing a three-dimensional surface, using a Cartesian coordinate system having x, y, and z axes, as a plurality of points on the surface, wherein each point is defined by an intensity value, a unique x-y coordinate, and a z coordinate specified by a z value;

collecting a first group of the points having similar z values, the first group representing a first layer of the surface;

collecting a second group of the points, each point of the second group having a z value similar to the z values of other points of the second group and dissimilar to the z values of the points of the first group, the second group representing a second layer of the surface;

determining a first error threshold for the intensity values of the first group of points; and

determining a second error threshold for the intensity values of the second group of points.

2. The method of claim 1, further comprising:

representing a second three-dimensional surface as a second plurality of points on the second surface, wherein the second plurality of points is represented by a set of intensity values and a set of x-y coordinates; and

aligning the x-y coordinates of the first-mentioned plurality of points with the x-y coordinates of the second plurality of points such that each of the first plurality of points corresponds with one of the second plurality of points;

determining, for each of the points in the first group, whether the intensity value of the point differs from the intensity value of the corresponding one of the second plurality of points by an amount greater than the first error threshold; and

determining, for each of the points in the second group, whether the intensity value of the point differs from the intensity value of the corresponding one of the second plurality of points by an amount greater than the second error threshold.

3. The method of claim 2, wherein points in the first group that have an intensity value that differs from the intensity value of the corresponding one of the second plurality of points by an amount greater than the first error threshold are identified as difference pixels, and

wherein points in the second group that have intensity values that differ from the corresponding one of the second plurality of points by an amount greater than the second error threshold are also identified as difference pixels.

4. The method of claim 3, further comprising grouping neighboring difference pixels within a difference boundary, the difference boundary defining an defect area that includes a potential defect on at least one of the first and second surfaces.

5. The method of claim 4, further comprising extracting defect parameters from the defect area.

6. The method of claim 5, wherein the defect parameters include at least one of the size, shape, and reflectivity of the defect area.

7. The method of claim 4, wherein defining the defect area includes dilating the neighboring difference pixels.

8. The method of claim 7, wherein dilating the neighboring difference pixels includes forming a ring of radius R around each of the neighboring difference pixels.

9. The method of claim 7, wherein defining the defect area includes eroding the dilated neighboring difference pixels.

10. A method of locating defects on a test surface, wherein the test surface is contained within a test volume represented by a Cartesian coordinate system having x, y, and z axes describing a set of unique x-y-z coordinates, the method comprising the steps of:

scanning the test surface in the test volume with a focused beam so that the focal point of the focused beam coincides, in turn, with each unique x-y-z coordinate within the test volume;

determining, for each column of points specified by a unique x-y coordinate in the test volume, a maximum reflected intensity value of the focused beam;

determining, for each column of points specified by a unique x-y coordinate in the test volume, a Z value corresponding to the maximum reflected intensity value of the focused beam;

storing all the Z values to form an array of test data representing a three-dimensional image of the test surface; and

comparing the array of test data with an array of reference data to identify elevational differences between the array of test data and the array of reference data.

11. The method of claim 10, further comprising comparing the differences between the array of test data and the array of reference data with defect data indicative of different defect types.

12. The method of claim 10, wherein the elevational differences are described as an average Z-difference of a defect area on the test surface.

13. The method of claim 10, wherein the elevational differences are described as a fraction of a first number of points imaged on the test surface that extend above a selected physical layer of the test surface over a second number of points imaged on the test surface that extend below the selected physical layer of the test surface.

14. A defect characterization system comprising:

a microscope configured to obtain a three-dimensional image of a surface, the image including a plurality of pixels each of which is expressed an intensity value and x, y, and z coordinates of a Cartesian coordinate system;

means for collecting a first group of the pixels having similar z coordinates, the first group representing a first physical layer of the surface;

means for collecting a second group of the pixels, each pixel of the second group having a z coordinate similar to the z coordinates of other pixels of the second group and dissimilar to the z coordinates of the pixels of the first group, the second group representing a second physical layer of the surface;

means for determining a first error threshold for the intensity values of the first group of pixels; and

means for determining a second error threshold for the intensity values of the second group of pixels.