Search

CN-116659382-B - Microstructure machining error calculation method based on image processing

CN116659382BCN 116659382 BCN116659382 BCN 116659382BCN-116659382-B

Abstract

The invention discloses a microstructure processing error calculating method based on image processing, micro-nano machining errors include depth errors, line width errors and overlay errors. The three-dimensional contour data is obtained by utilizing an image processing method and is truncated along the step tangent plane direction, the depth error and the line width error are calculated, then the projection is carried out on the three-dimensional contour data in the overlooking direction, a step edge line segment graph can be obtained, and the overlay error is calculated. The invention provides a microstructure processing error calculation method based on image processing in order to improve the precision requirement of the existing processing error.

Inventors

  • LUO QIAN
  • GAO GUOHAN
  • DU JUNFENG

Assignees

  • 中国科学院光电技术研究所

Dates

Publication Date
20260512
Application Date
20230601

Claims (5)

  1. 1. The microstructure processing error calculation method based on image processing comprises depth error, line width error and overlay error, and is characterized by comprising the following steps: step one, auxiliary detection mark points are made on a detected microstructure; Step two, detecting the step contours of the graph areas near all the marking points by using a white light interferometer to obtain three-dimensional contour data of corresponding positions; Cutting the three-dimensional profile data in the second step into a plurality of cut lines along the direction of the step tangential plane, and averaging the depth value and the line width value of the steps on the plurality of cut lines to obtain the actually measured depth value and line width value of each point position; Step four, comparing the actually measured depth value and line width value on each point position with the design value, calculating the depth error and line width error on each point position, and calculating the average value of the depth error and line width error of all point positions to obtain the depth error and line width error of the microstructure; Step five, projecting the three-dimensional profile data in the overlooking direction of the step two to obtain a step edge line segment diagram, fitting the line segment diagram, taking a plurality of intercept lines along the vertical direction of the edge line segment to obtain the intercept of each intercept line in each period, obtaining the line width value of steps in each period of the point position by the average intercept of the plurality of intercept lines, wherein the step number is 2 n , and the measuring range of the interferometer is a plurality of periods; Step six, calculating the difference of line width values between the current step and the previous step in each period, averaging the difference to obtain an overlay error in the period, averaging the overlay errors of a plurality of periods to obtain an overlay error of the point location, and averaging the overlay errors of all the point locations to obtain the overlay error of the microstructure.
  2. 2. The method of claim 1, wherein the marking point bitmap set in the first step is dense or sparse according to the size and detection accuracy of the microstructure to be measured.
  3. 3. The method of claim 1, wherein more than 3 stubs are taken in each of the third and fifth steps.
  4. 4. The method of claim 1, wherein the third step comprises the step of determining the average line position of the whole number of the cross-section lines.
  5. 5. The method for calculating the processing error of the microstructure based on image processing of claim 1, wherein n is 4, 8 and 16.

Description

Microstructure machining error calculation method based on image processing Technical Field The invention belongs to the field of micro-nano machining errors, and particularly relates to a micro-structure machining error calculation method based on image processing. Background With the development of optical technology, optical, mechanical and electrical integration is a trend, and an optical system is required to be integrated, arrayed and miniaturized, so that the diffraction optical element is more and more widely applied. The traditional diffraction optical element is manufactured by adopting a plurality of mask alignment processing technologies, and in theory, the diffraction efficiency of the 4-step diffraction optical element can reach 81 percent, and the diffraction efficiency of the 8-step diffraction optical element can reach 95 percent. But various processing errors such as depth errors, line width errors, overlay errors, etc. are generated during the processing of the microstructures. The actual phase structure of the device deviates from the design distribution, so that the diffraction efficiency of the device is greatly reduced, the diffraction optical element is designed and manufactured better, the requirement on the processing technology is required to be put forward according to the design requirement or the effect of the diffraction optical element is estimated in advance according to the existing technology level, and therefore, the analysis of the influence of the processing error on the diffraction optical element is significant. In the etching process, the etching rate calibration is inaccurate, the etching rate is reduced due to material deposition in the etching process, the etching depth deviates from the design value, taking a period of a four-step diffraction element as an example, and when only the etching depth error exists, a constant phase deviation is introduced. Line width errors-when either the development time is too long or the overexposure results in a line width greater than the design value, whereas, when either the underexposure or the development time is too short, the processed line width value is less than the design value. Overlay error-ideally, the complete overlay between the current layer pattern and the previous layer pattern of the integrated circuit, i.e., the relative displacement is zero. However, there is always a certain registration error in the actual process, because the mask of the current layer pattern cannot be perfectly aligned with the mask of the previous layer pattern during the alignment and exposure steps. The overlay error refers to the relative offset between the current layer pattern and the previous layer pattern caused by the non-ideal overlay of the mask. In the prior art, an optical microscope is adopted to record vernier data, an overlay error result is obtained by reading the vernier data, then a white light interferometer is adopted to detect the step profile of a pattern area near all mark points, and etching depth data of corresponding positions is obtained to obtain depth and line width errors. The invention can obtain depth, line width and overlay error simultaneously by carrying out one-time detection of the white light interferometer and carrying out program calculation. And a plurality of sampling points are arranged, coarse errors can be removed, an average value is obtained, a detection area covers the effective range of the whole detected mirror, and the reliability is higher. Disclosure of Invention The invention aims to solve the problem of calculation of the introduced machining error in the micro-nano machining process. The invention provides a microstructure processing error calculation method based on image processing in order to improve the precision requirement of the existing processing error. The technical scheme adopted by the invention is that the micro-structure processing error calculation method based on image processing, wherein the micro-structure processing error comprises depth error, line width error and overlay error, and is characterized by comprising the following steps: step one, auxiliary detection mark points are made on a detected microstructure; Step two, detecting the step contours of the graph areas near all the marking points by using a white light interferometer to obtain three-dimensional contour data of corresponding positions; Cutting the three-dimensional profile data in the second step into a plurality of cut lines along the step tangential direction, and averaging the depth values and the line width values on the plurality of cut lines to obtain the actually measured depth values and line width values of each point position; Step four, comparing the actually measured depth value and line width value on each point position with a design value, calculating the depth error and line width error on each point position, and calculating the average value of the depth error and line width err