Acquisition method for actual pixel size of charged particle beam scanning imaging equipment

Abstract

The invention discloses a method for acquiring the actual pixel size of charged particle beam scanning imaging equipment. The method comprises the following steps of: planning a working area for measuring the actual pixel size on a wafer; dividing the amplification factors corresponding to all the to-be-detected pixels into a first amplification factor and a higher second amplification factor; when the actual pixel size of the first magnification is obtained, a template meeting a preset condition is extracted, the template comprises one or more sub-templates, and the sub-templates comprise sub-images and / or independent feature points; when the actual pixel size of the second magnification is obtained, extracting an image template or a curve template based on a curve; the method comprises the following steps of: moving a wafer to acquire a target image, performing matching in the target image by using a template, acquiring target displacement between the position of the template and a matching position, and acquiring the actual pixel size according to the actual displacement of the wafer and the target displacement. The invention can improve the success rate, the measurement accuracy and the time cost.

Description

technical field [0001] The invention relates to the technical field of semiconductor devices, in particular to a method for obtaining actual pixel size of a charged particle beam scanning imaging device. Background technique [0002] In the front-end manufacturing process of semiconductor large-scale integrated circuits, charged particle beam scanning imaging equipment is often used, for example, scanning electron microscopy (Scanning Electron Microscopy, SEM) equipment is used to detect defects or re-inspect wafers. The SEM equipment can be an initial inspection equipment (E-Beam Inspection, EBI) or a re-inspection equipment (E-Beam Review, EBR), which are electron beam scanning imaging systems in principle. Take the EBR equipment as an example, but it does not exclude the EBI equipment in some applications. A quick and easy method is needed to obtain the actual pixel size of the equipment for wafer scanning imaging, which is used to estimate the actual physical size of the...

AI Technical Summary

Problems solved by technology

[0022] In method 1, it is necessary to use a special wafer or sample prepared in the prior art to obtain the actual pixel size, and the time cost is high (taking up precious machine time of the equipment on the IC production line), and the wafer or sample to the electron optical column There is a large difference between the working distance of the customer's wafer and the working distance of the electron optical lens barrel, resulting in errors, resulting in poor accuracy of the actual pixel size obtained and inconvenient to use;

[0023] In method 2, template matching methods that do not distinguish images of different magnifications in the prior art use general template matching algorithms, for example, use cross-correlation algorithm (NCC) or use general feature-based template matching, mainly edge features The algorithm is only barely applicable to the image of the charged particle beam scanning imaging device in LM, and basically cannot be used in HM

This is due to the fact that the features in the image become sparse (or the features are insufficient in a certain direction) during HM and the image signal-to-noise ratio is worse due to the physical mechanism of imaging itself, and the local differences in the image increase, so that the template matching in the prior art is basically Unusable, poor reliability, poor accuracy

In addition, the measurement of the prior art only uses a single match (because the template is worried that the template will run out of the target image when the wafer is moved), at this time, the moving distance of the mechanical motion platform is limited, and the relative error of the measurement is relatively large, especially when using the existing The relative error of the template matching algorithm is larger

[0024] Therefore, for the existing technology, it is poor in terms of accuracy, reliability, time cost

Images