Search

CN-122029564-A - Method, system and storage medium for processing back image of cut wafer

CN122029564ACN 122029564 ACN122029564 ACN 122029564ACN-122029564-A

Abstract

The invention provides a processing method of a back image of a cut wafer, a processing system of the back image of the cut wafer and a computer readable storage medium. The processing method of the back image of the cut wafer comprises the following steps of obtaining the back image of the cut wafer, and integrating gray values of a plurality of pixel points in the back image along the X direction and the Y direction respectively to detect a plurality of scribing lanes extending along the X direction and the Y direction in the back image respectively. And calculating the coordinates of the intersection points of the first scribing lanes extending along the X direction and the second scribing lanes extending along the Y direction respectively, and generating a grain distribution diagram on the back surface of the cut wafer according to the coordinates. The invention can be used for generating the grain distribution diagram of the back surface of the cut wafer, so as to detect the defects of the back surface of the corresponding grain, thereby improving the detection accuracy.

Inventors

  • XUE XIAOWEI
  • GE LIANG

Assignees

  • 睿励科学仪器（上海）有限公司

Dates

Publication Date
20260512
Application Date
20240830

Claims (14)

  1. The processing method of the back image of the cut wafer is characterized by comprising the following steps: Acquiring a back image of the cut wafer; Integrating gray values of a plurality of pixels in the back image along an X direction and a Y direction respectively to detect a plurality of scribe lanes extending along the X direction and the Y direction in the back image respectively, wherein the X direction is perpendicular to the Y direction, and And respectively calculating the coordinates of the intersection points of the first scribing lanes extending along the X direction and the second scribing lanes extending along the Y direction, and generating a grain distribution diagram of the back surface of the cut wafer according to the coordinates.
  2. The method of processing of claim 1, wherein the step of acquiring a backside image of the diced wafer comprises: collecting a first image of the back surface of the cut wafer under a preset first illumination brightness; Parsing the first image to determine noise effects therein; Responding to the noise influence being larger than a preset influence threshold value, increasing illumination brightness, re-collecting a second image of the back surface of the cut wafer, analyzing the second image until the noise influence is smaller than or equal to the influence threshold value, and And acquiring back images of at least one cut wafer in the same batch under the current illumination brightness in response to the noise influence being smaller than or equal to the influence threshold.
  3. The method of processing of claim 1, wherein the step of acquiring a backside image of the diced wafer comprises: Scanning the back surface of the cut wafer column by column along a preset first direction to obtain a plurality of line scan images of the back surface of the cut wafer, and And splicing the plurality of line scan images along a preset second direction to obtain a back image of the cut wafer, wherein the second direction is perpendicular to the first direction.
  4. The method of processing of claim 1, wherein the step of integrating the gray values of the plurality of pixels in the back image in the X-direction and the Y-direction to detect the plurality of scribe lanes extending in the X-direction and the Y-direction in the back image, respectively, comprises: Determining a plurality of first peaks of gray scale integration values of the plurality of pixel points along the X direction and a plurality of second peaks of gray scale integration values of the plurality of pixel points along the Y direction respectively by a peak detection method; Determining a first coordinate trace of a plurality of first scribe lanes extending in the X-direction according to a first position of the plurality of first peaks, and And determining a second coordinate track of a plurality of second scribing lanes extending along the Y direction according to the second positions of the plurality of second peaks.
  5. The processing method according to claim 4, wherein before integrating the gradation values of the plurality of pixel points in the back image in the X-direction and the Y-direction, respectively, the processing method further comprises the steps of: Preliminarily calibrating a first angle interval [ theta start_X ,θ end_X ] where the X direction is located, and defining a first search step theta step_X ; within the first angle interval [ theta start_X ,θ end_X ], respectively calculating gray scale integration arrays RadonResult i at a plurality of angles [ theta ] i according to the first search step [ theta ] step_X , wherein [ theta ] i ＝θ start_X +i×θ step_X , the gray scale integration arrays RadonResult i comprise gray scale integration values of a plurality of rows of pixel points along a third direction of the angle [ theta ] i , and And determining a third direction corresponding to an angle theta i ' of the most obvious peak characteristic in the gray integral array RadonResult i under each angle theta i as the X direction.
  6. The processing method according to claim 5, wherein before integrating the gradation values of the plurality of pixel points in the back image in the X-direction and the Y-direction, respectively, the processing method further comprises the steps of: Preliminarily calibrating a second angle interval [ theta start_Y ,θ end_Y ] where the Y direction is located, and defining a second search step theta step_Y ; Within the second angle interval [ theta start_Y ,θ end_Y ], respectively calculating gray scale integration arrays RadonResult j at a plurality of angles [ theta ] j according to the second search step [ theta ] step_Y , wherein [ theta ] j ＝θ start_Y +j×θ step_Y , the gray scale integration arrays RadonResult j comprise gray scale integration values of a plurality of columns of pixel points along the fourth direction of the angle [ theta ] j , and And determining a fourth direction corresponding to an angle theta i 'of the gray scale integration array RadonResult j with the angle theta j , wherein the angle theta i ' is the most obvious of peak characteristics. The larger the value RadonResultAbsDiffSum, the more pronounced the peak characteristic of the gray integral array RadonResult at the corresponding angle.
  7. The processing method of claim 5, wherein determining a first coordinate trace of a plurality of first scribe lanes extending in the X-direction based on a first location of the plurality of first peaks comprises: And respectively determining a first linear equation of the plurality of first dicing lanes in the back image of the cut wafer according to the angle theta i 'and the intercept of the plurality of first peaks in the X direction in the corresponding gray scale integration array RadonResult i '.
  8. The processing method of claim 6, wherein determining a second coordinate trace of a plurality of second scribe lanes extending in the Y-direction based on a second location of the plurality of second peaks comprises: And respectively determining a second linear equation of the plurality of second scribe lanes in the back image of the cut wafer according to the angle theta j 'and the intercept of the plurality of second peaks in the Y direction in the corresponding gray scale integration array RadonResult j '.
  9. The method of processing of claim 4, wherein the step of calculating coordinates of intersections of the first plurality of streets extending in the X-direction and the second plurality of streets extending in the Y-direction, respectively, and generating a die distribution map of the backside of the diced wafer therefrom comprises: Determining intersection coordinates of a plurality of intersections of the plurality of first scribe lanes and the plurality of second scribe lanes according to the first coordinate tracks of the plurality of first scribe lanes and the second coordinate tracks of the plurality of second scribe lanes; Determining the coordinate position, width and height of a corresponding die on the back of the cut wafer according to the coordinates of the intersection points p i,j 、p i,j+1 、p i+1,j 、p i+1,j+1 of every four adjacent intersection points along the X direction and the Y direction, and And generating a grain distribution diagram of the back surface of the cut wafer according to the coordinate position, the width and the height of each grain on the back surface of the cut wafer.
  10. The method of processing of claim 10, wherein generating a die map of the backside of the diced wafer based on the coordinate position, width and height of each die on the backside of the diced wafer comprises: Determining the sum of pixel values of a plurality of pixel points in the wafer according to the coordinate position, the width and the height of each grain on the back surface of the wafer after cutting; Screening out data of coordinate positions, widths and heights of the grains in response to the sum of pixel values of any one of the grains being less than a preset threshold value, and And generating a grain distribution diagram of the back surface of the cut wafer according to the coordinate positions, the widths and the heights of a plurality of grains of which the sum of pixel values is larger than or equal to the preset threshold value on the back surface of the cut wafer.
  11. The process of claim 1, further comprising the step of: determining the coordinate position of at least one die on the back side of the cut wafer according to the die distribution diagram, and And detecting defects on the back surface of the corresponding crystal grain on the cut wafer according to the coordinate positions.
  12. A system for processing a backside image of a diced wafer, comprising: A memory having stored thereon computer instructions, and A processor coupled to the memory and configured to execute computer instructions stored on the memory to implement the method of processing a backside image of a diced wafer as claimed in any of claims 1-12.
  13. The processing system of claim 13, further comprising: Linear array camera, and The motion platform supports three-degree-of-freedom motion which is horizontally and vertically translated and axially rotated in a carrying plane and is used for moving a plurality of detection areas on the back surface of the cut wafer to an image acquisition range of the linear array camera row by row so as to scan the back surface of the cut wafer row by row to obtain a plurality of line scanning images on the back surface of the cut wafer and/or rotating the cut wafer to calibrate the X direction and/or the Y direction.
  14. A computer readable storage medium having stored thereon computer instructions which, when executed by a processor, implement the method of processing a backside image of a diced wafer according to any of claims 1 to 12.

Description

Method, system and storage medium for processing back image of cut wafer Technical Field The present invention relates to the field of semiconductor technology, and in particular, to a method for processing a backside image of a diced wafer, a system for processing a backside image of a diced wafer, and a computer readable storage medium. Background In the semiconductor field, in order to detect a defect of each die in a wafer, an image of each die needs to be extracted, and then various methods are used to detect the defect. In order to extract a complete image of each die, the position of each die in the wafer needs to be obtained. In a typical wafer, a wafer manufacturer may provide a Die Map (Die Map). After the wafer is aligned, the position of each die can be obtained through the die distribution diagram. However, after dicing the wafer (Frame), scribe lines (scribe lines) will exist between each die. Dicing streets can increase die-to-die distance, thereby resulting in die actual distribution that differs from the die profile provided by the manufacturer. Therefore, the grain profile cannot be used any more to obtain the actual position of each grain. In addition, in both CN117132603B, CN116153818A, CN116612113a and CN116542945A, a template matching method is used to generate the grain distribution map. Firstly, selecting a complete grain area as a matching template, then searching each grain in the searching area successively, recording the grain coordinates, finally finding all grains and generating a grain distribution diagram. In CN113538586a, a plurality of selected grain images and coordinates thereof are used, fourier transformation is used to obtain periodic information of grains, then fourier transformation is performed in a detection area, the periodic information is used to obtain a feature matrix describing the grains, and finally, position information of all grains is obtained through calculation and a grain distribution map is generated. However, the above method is mainly suitable for generating a front grain distribution diagram of a cut wafer, and since the back surface of the cut wafer does not have a stable and fixed pattern, template matching or periodic distribution searching cannot be applied, so that the method cannot be normally used for the back surface of the cut wafer. In order to overcome the above-mentioned drawbacks of the prior art, there is a need in the art for an improved method for processing a backside image of a diced wafer, which is used for generating a die distribution map of the backside of the diced wafer, so as to detect defects on the backside of the corresponding die, so as to improve the accuracy of the detection. Disclosure of Invention The following presents a simplified summary of one or more aspects in order to provide a basic understanding of such aspects. This summary is not an extensive overview of all contemplated aspects, and is intended to neither identify key or critical elements of all aspects nor delineate the scope of any or all aspects. Its sole purpose is to present some concepts of one or more aspects in a simplified form as a prelude to the more detailed description that is presented later. In order to overcome the above-mentioned drawbacks of the prior art, the present invention provides a method for processing a backside image of a diced wafer, a system for processing a backside image of a diced wafer, and a computer readable storage medium, which can be used for generating a distribution diagram of dies on the backside of a diced wafer by detecting a plurality of scribe lanes in an X direction and a Y direction of the diced wafer, respectively, so as to detect defects on the backside of corresponding dies, thereby improving the accuracy of the detection. The method for processing the back image of the cut wafer comprises the steps of obtaining the back image of the cut wafer, and integrating gray values of a plurality of pixel points in the back image along an X direction and a Y direction respectively to detect a plurality of scribing lanes extending along the X direction and the Y direction in the back image. And calculating the coordinates of the intersection points of the first scribing lanes extending along the X direction and the second scribing lanes extending along the Y direction respectively, and generating a grain distribution diagram on the back surface of the cut wafer according to the coordinates. Further, in some embodiments of the present invention, the step of acquiring a backside image of the diced wafer includes acquiring a first image of the backside of the diced wafer at a preset first illumination intensity, parsing the first image to determine noise effects therein, responsive to the noise effects being greater than a preset impact threshold, increasing illumination intensity, re-acquiring a second image of the backside of the diced wafer and parsing the second image until the noise effects therein are less than