Method and system for non-destructive inspection of cell etch redeposition
By capturing and processing the grayscale images of wafers, the grayscale levels of etched redeposits are automatically evaluated, which solves the problem of yield reduction caused by dry etching, and realizes efficient and accurate etched redeposit inspection, improving semiconductor manufacturing yield and throughput.
Patent Information
- Application Number
- CN202110250974.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-03-08
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2041-03-08
AI Technical Summary
Prior Art In the semiconductor manufacturing process, sidewall re-deposits caused by dry etching are difficult to accurately evaluate, resulting in a reduction in yield, and traditional manual inspection methods are uncontrollable and destructive, affecting throughput and yield.
By capturing the grayscale image of the wafer, identifying units and subtracting pixels of non-interest areas, etching residues based on the grayscale score of the remaining pixels, and using image processing devices to automatically evaluate and process, achieving non-destructive, fast and high sensitivity etching and sediment inspection.
Improves the evaluation accuracy and yield of etched redeposits, reduces manufacturing costs, increases throughput, and improves the reliability and efficiency of evaluation through automated processing.
Smart Images

Figure CN113192856B_ABST
Abstract
Description
Technical Field
[0001] Embodiments of the present invention are directed to methods and systems for non-destructive inspection of cell etch redeposition. Background Art
[0002] During the manufacture of integrated circuits (ICs), a multi-step sequence of semiconductor manufacturing processes is performed to gradually form electronic circuits on a semiconductor substrate. One such semiconductor manufacturing process is dry etching. Dry etching removes material by ion bombardment and is typically highly anisotropic. Summary of the invention
[0003] According to one aspect of an embodiment of the present invention, a method for non-destructive inspection of unit etching redeposition is provided, comprising: capturing a grayscale image of a plurality of units on a wafer, wherein the grayscale image provides a top view and is captured when etching is completed to form a unit; identifying the unit in the grayscale image; subtracting an area of the grayscale image corresponding to the identified unit from the grayscale image; scoring an amount of etching residue on the sidewalls of the unit and in the grooves between the units based on the grayscale levels of the remaining pixels at the remaining portion of the grayscale image; and processing the wafer based on the score of the scoring.
[0004] According to another aspect of an embodiment of the present invention, a method for non-destructive inspection of unit etching redeposition is provided, comprising: capturing a grayscale image of multiple units on a wafer, wherein the grayscale image provides a top view and is captured after etching to form units; performing image processing on the grayscale image, the image processing comprising: identifying units in the grayscale image; determining region of interest (ROI) pixels based on the identification, wherein the ROI pixels include pixels between the identified units rather than at the identified units; determining an average grayscale level between the ROI pixels; and determining a ratio of ROI pixels whose grayscale levels exceed a threshold to the total number of ROI pixels, wherein the threshold is greater than the average grayscale level; and processing the wafer based on the ratio.
[0005] According to another aspect of an embodiment of the present invention, a system for non-destructive inspection of unit etching redeposition is provided, comprising: an etching process tool configured to etch a multilayer film on top of a wafer to form a plurality of units from the multilayer film; an imaging device configured to capture a grayscale image of the unit, wherein the grayscale image provides a top view of the unit; and an image processing device configured to: identify the unit in the grayscale image; subtract non-region of interest (non-ROI) pixels from the grayscale image, wherein the non-ROI pixels include pixels at the identified unit; and generate a score for a certain amount of etching residue on the sidewalls of the unit and in the grooves between the units based on the grayscale levels of the remaining pixels at the grayscale image; and a processing controller configured to process the wafer based on the score. BRIEF DESCRIPTION OF THE DRAWINGS
[0006] Aspects of the present invention will be best understood from the following detailed description when read in conjunction with the accompanying drawings. It should be noted that, in accordance with standard practice in the industry, the various components are not drawn to scale. In fact, the dimensions of the various components may be arbitrarily increased or decreased for clarity of discussion.
[0007] Figure 1 A schematic flow chart illustrating some embodiments of a method for non-destructive inspection of cell etch redeposition.
[0008] Figure 2A and Figure 2B Show Figure 1 Schematic flow chart of some optional embodiments of the method.
[0009] Figure 3 , Figure 4 , Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C , Fig.10 and Fig.11 Shown in more detail Figure 1 Some embodiments of the method.
[0010] Fig.12 Show Figure 3 , Figure 4 , Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C , Fig.10 and Fig.11 A block diagram of some embodiments of the method.
[0011] Fig.13 Show Figure 4 Some optional embodiments of grayscale images.
[0012] Fig.14 Show Figure 5A Some optional embodiments of the mask image.
[0013] FIG. 15A to FIG. 15C Shown separately 9A to 9C Some optional embodiments of a mild ROI image, a severe ROI image, and a mild / severe ROI image.
[0014] Fig.16 Show Fig.10 Some optional embodiments of the illustrations, wherein scores are determined individually for IC dies.
[0015] Fig.17A and Fig. 17B , Fig.18 , Fig.19 , Fig. 20A , Fig. 20B and Figures 21 to 28 Some embodiments of methods for forming integrated circuits (ICs) and performing non-destructive inspection for cell etch redeposition are shown.
[0016] Fig.29 Show Fig.17A and Fig. 17B , Fig.18 , Fig.19 , Fig. 20A , Fig. 20B and Figures 21 to 28 A block diagram of some embodiments of the method.
[0017] Fig.30 shows the use of parallel processing of wafers and grayscale images of wafers Fig.29 Block diagram of some optional embodiments of the method.
[0018] Fig.31 Schematic diagrams showing some embodiments of systems for forming ICs and performing non-destructive inspection for cell etch redeposition. DETAILED DESCRIPTION
[0019] The following disclosure provides many different embodiments or examples for implementing different features of the present invention. Specific examples of components and arrangements are described below to simplify the present invention. Of course, these are merely examples and are not intended to limit the present invention. For example, in the following description, forming a first component above or on a second component may include an embodiment in which the first component and the second component are directly contacted, and may also include an embodiment in which an additional component may be formed between the first component and the second component so that the first component and the second component may not be in direct contact. In addition, the present invention may repeat reference numerals and / or characters in various examples. This repetition is for the purpose of simplicity and clarity, and does not itself indicate the relationship between the various embodiments and / or configurations discussed.
[0020] Furthermore, for ease of description, spatially relative terms such as "below," "beneath," "lower," "above," "upper," etc. may be used herein to describe the relationship of one element or component to another (or additional) elements or components as shown in the figures. The spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. The device may be otherwise oriented (rotated 90 degrees or at other orientations), and the spatially relative descriptors used herein should likewise be interpreted accordingly.
[0021] A method for forming a magnetoresistive random access memory (MRAM) may include: depositing a multilayer film over a wafer; forming a hard mask over the multilayer film; and performing dry etching on the multilayer film to form an array of magnetic tunnel junctions (MTJs). The challenge of this method is that dry etching uses ion bombardment to etch and is therefore somewhat uncontrollable. The etched material may be redeposited on the sidewalls of the MTJ and may cause a large amount of leakage between the fixed element and the free element of the MTJ. Due to high leakage, the redeposited material may significantly reduce the yield.
[0022] In order to mitigate the reduction in yield due to sidewall redeposition, a quantitative inspection may be performed to determine the amount of redeposition and determine whether to reprocess the wafer. According to the first inspection method, an individual manually inspects one or several MTJs from top to bottom using a scanning electron microscope (SEM). However, the redeposited material occupies a smaller area when viewed from top to bottom, and the contrast between the redeposited material and the adjacent material may be low. As a result, it may be difficult for an individual to reliably assess the sidewall redeposition, and the sensitivity of the first inspection process may be low. In addition, since the inspection is performed manually, the throughput may be low and the number of MTJs evaluated may be low. According to the second inspection method, an individual manually prepares and inspects a cross section of one or several MTJs using a transmission electron microscope (TEM). However, preparing the cross section is destructive and therefore results in wasting expensive wafers. In addition, since the preparation and inspection are performed manually, the throughput may be low and the number of MTJs evaluated may be low. In addition, due to the preparation, the inspection is performed non-in-situ and adds complexity to the workflow.
[0023] Various embodiments of the present invention are directed to a method for non-destructive inspection of cell etching and redeposition. In some embodiments of the method, a grayscale image of a plurality of cells on a wafer is captured. The grayscale image provides a top view of the cell, and in some embodiments, is captured in situ after etching to form the cell. In some embodiments, a SEM is used to capture the grayscale image, and / or the cell is an MTJ. The cell is identified in the grayscale image to determine a non-region of interest (non-ROI) pixel corresponding to the cell. An image processing device is provided to subtract non-ROI pixels from the grayscale image to determine ROI pixels. ROI pixels are remaining pixels and correspond to materials on the sidewalls of the cell and in the grooves between the cells. The amount of etching redeposition on the sidewalls and in the grooves is then scored by the image processing device based on the grayscale level of the ROI pixels. In addition, the wafer is processed based on the score. For example, if the score indicates a threshold amount of etching and redeposition, the cell may be reprocessed (e.g., removed and re-formed). Otherwise, the wafer may proceed to the next processing step for forming an integrated circuit (IC) including the cell.
[0024] Because the images are captured in situ, the method can be performed with little impact on existing workflows and with less impact on preparation work. Because the images are captured from top to bottom (e.g., using a SEM), the images can capture large areas and a large number of cells. In addition, the method can be performed non-destructively. Because most of the method is performed by an image processing device, the etching redeposition on a large number of cells can be quickly and reliably evaluated. Because the method focuses on the gray level of a single pixel, the method can achieve high sensitivity. Because the method simplifies the amount of etching redeposition into a score, the amount of problematic etching redeposition can be easily identified.
[0025] For some embodiments, the cells may be in a periodic pattern. For example, when the cells are MTJs in an array, the cells may be in a periodic pattern. In embodiments where the cells are in a periodic pattern, positioning within a grayscale image may be easily achieved. After locating a single cell in the grayscale image, the location of the remaining portion of the cell may be calculated using the location of the single cell and the periodic pattern.
[0026] refer to Figure 1 , a schematic flow chart 100 of some embodiments of a method for non-destructive inspection of cell etch redeposition is provided. The method is performed after etching, thereby forming a plurality of cells 102 (represented by black circles) on a wafer 104. The etching may be performed, for example, by dry etching using ion bombardment, so that the etched material has a high propensity to redeposit on the sidewalls of the cells 102. The cells 102 are spread across a plurality of IC dies 106 of the wafer 104, and may, for example, be or include an MTJ, a gate stack of a logic device, or some other suitable cell structure.
[0027] At 108, a grayscale image 110 (only partially shown) of the cell 102 at a portion of the wafer 104 is captured. Note that different hashes are used to represent different gray levels. The grayscale image 110 provides a top view of the cell 102 and may be captured, for example, by a SEM, a inspection SEM (RSEM), or some other suitable imaging device. Because the grayscale image 110 provides a top view of the cell 102, the grayscale image 110 may be captured non-destructively, and with little or no preparation of the wafer 104. This, in turn, may reduce manufacturing costs and / or increase yield. Furthermore, because the grayscale image 110 is captured from the top down, the grayscale image 110 may capture a large area and a large number of cells.
[0028] In some embodiments, the grayscale image 110 is captured in-situ in a process chamber used for etching to form the unit 102. For example, etching and image capture can be performed in a common process chamber, and the wafer 104 can remain in the common process chamber from the beginning of etching to the end of image capture. In at least embodiments where the grayscale image 110 is captured in-situ, the method can be performed with little impact on existing workflows and less impact on preparation work. As a result, throughput may be higher. In an optional embodiment, the grayscale image 110 is captured ex-situ, and therefore captured outside of the process chamber used for etching.
[0029] After capturing the grayscale image 110, image processing is performed on the grayscale image 110 at 112 to assess the amount of etch redeposition on the sidewalls of the cell 102. Depending on the etched material, the etch redeposition may be conductive and may include, for example, titanium, ruthenium, tantalum, some other suitable material, or any combination of the foregoing. Because the etch redeposition may be conductive, the etch redeposition may increase leakage current. For example, in the case where the cell 102 is an MTJ, the etch redeposition may increase leakage current from the fixed layer to the free layer. The increased leakage current may in turn reduce the yield. In some embodiments, the image processing is fully automated and performed by an image processing device.
[0030] At 114, a non-ROI 116 is identified and subtracted from the grayscale image 110 to identify a ROI 118. For clarity, this is schematically illustrated by a mask image 120. The non-ROI 116 (shown in black in the mask image 120) includes the region of the grayscale image 110 corresponding to the cell 102, and also includes a peripheral region of the grayscale image 110. The peripheral region extends in a closed path along the periphery of the grayscale image 110, and may, for example, have a square ring shape or some other suitable shape. In an alternative embodiment, the non-ROI 116 is limited to the region of the grayscale image corresponding to the cell, and therefore does not include the peripheral region. After the non-ROI 116 is subtracted from the grayscale image 110, the ROI 118 (shown in white in the mask image 120) corresponds to the remaining portion of the grayscale image 110. In other words, the ROI 118 corresponds to the region of the grayscale image 110 between the cells 102.
[0031] Identifying the non-ROI 116 includes identifying the cells 102 in the grayscale image 110. In some embodiments, the cells 102 are randomly or pseudo-randomly arranged. In at least some embodiments where the cells 102 are randomly or pseudo-randomly arranged, each of the cells 102 is individually identified manually or automatically. With respect to manual identification, a person may, for example, use a human interface device (HID) (e.g., a mouse) and a graphical user interface (GUI) displayed on a display device to draw a circle or other closed shape around the cells 102. With respect to automatic identification, an image processing device may, for example, automatically identify the cells 102 by computer vision. In other embodiments, the cells 102 are arranged in a periodic pattern. For example, the cells 102 may be in multiple rows and columns to define an array. In at least some of these embodiments, a single or multiple cells 102 but less than all of the cells 102 are manually or automatically identified by computer vision, and the remaining portion of the cells 102 is identified by calculating the positions of the remaining cells based on the positions of the single cells and the periodic pattern. When more than one of the cells 102 is identified manually or automatically by computer vision, the positions of the remaining cells may be more accurately calculated.
[0032] At 122, a gray level distribution 124 is determined for the pixels of the grayscale image 110 in the ROI 118. Such pixels may also be referred to as ROI pixels. In some embodiments, the ROI pixels are 8-bit pixels, so that the gray level varies from 0 to 255. In other embodiments, the ROI pixels have some other suitable number of bits. The gray level distribution 124 includes a pixel count for each gray level or a quantile of the gray level. In some embodiments, the gray level distribution 124 is automatically determined by an image processing device.
[0033] At 126, the ROI pixels are classified into a heavy category, a light category, and a normal category in grayscale. In optional embodiments, more or fewer categories are applicable. The heavy category is defined by ROI pixels corresponding to the possibility of etching redeposition, and is shown by a heavy ROI image 128. The heavy ROI image 128 is black and white. In addition, the heavy ROI pixels are shown in white, and the remaining portion in the grayscale image 110 is shown in black. The light category is defined by ROI pixels corresponding to the possibility of etching redeposition moderate, and is shown by a light ROI image 130. The light ROI image 130 is black and white. In addition, the light ROI pixels are shown in white, and the remaining portion in the grayscale image 110 is shown in black. The normal category is defined by the remaining portion of the ROI pixels, which have a low probability of corresponding to etching redeposition. In some embodiments, the classification is automatically performed by the image processing device.
[0034] It is understood that the probability of a pixel corresponding to etching redeposition is proportional to the gray level. Thus, ROI pixels having a gray level less than a first threshold are assigned to the normal category, and ROI pixels having a gray level greater than a second threshold are assigned to the severe category. In addition, ROI pixels between the first threshold and the second threshold are assigned to the mild category.
[0035] Although the method focuses on a single grayscale image, in practice the method will be repeated for multiple grayscale images. For example, one or more grayscale images may be captured per IC die 106. Furthermore, although the image capture conditions are ideally the same when capturing multiple grayscale images, this may not always be the case. Thus, normalization is performed while classifying the ROI pixels using the average grayscale level of the ROI pixels. In particular, the first and second thresholds are the sum of the average grayscale levels with corresponding offsets. For example, the first threshold may be the sum of the average grayscale levels with a first offset, and the second threshold may be the sum of the average grayscale levels with a second offset greater than the first offset.
[0036] Defining the first and second thresholds as described above has the effect of using the relative brightness differences of the ROI pixels to classify the ROI pixels without having to directly calculate the relative brightness differences of the ROI pixels. In particular, the relative brightness difference of a given ROI pixel is the difference between the average gray level and the gray level of the given ROI pixel. Since the classification is performed by mathematically comparing the gray level of the ROI pixel with the first threshold and the second threshold, and since the first threshold and the second threshold are the sum of the average gray level and the corresponding offset, the average gray level can be subtracted from each side of the mathematical comparison without changing the effect of the mathematical comparison. In addition, the mathematical comparison can be rewritten as a mathematical comparison of the offset and the relative brightness difference. Therefore, the offset can be regarded as a threshold for the classification of the ROI pixel in the relative brightness difference domain.
[0037] At 132, a score 134 is determined based on the number of ROI pixels in different categories. The greater the ratio of heavy ROI pixels to the total number of ROI pixels, the higher the score. Additionally, in some embodiments, the greater the ratio of light ROI pixels to the total number of ROI pixels, the higher the score. The higher the score, the more etch redeposition on the sidewalls of the cell 102. In some embodiments, the scoring is performed automatically by an image processing device. In some embodiments, the score 134 is a percentage of heavy ROI pixels. In other embodiments, the score 134 is a percentage of heavy and light ROI pixels.
[0038] At 136, the wafer 104 is subjected to processing based on the score 134. For example, prior to the method, the wafer 104 may be subjected to a series of processing steps to form an IC at each of the IC dies 106. Based on the score 134, the wafer 104 may be subjected to the next processing step in the series, or may otherwise be subjected to reprocessing. Reprocessing may, for example, include removing and reforming a unit. Reforming may, for example, include repeating the processing steps in the series.
[0039] Processing at 136 includes evaluating the score 134 to determine whether the wafer 104 should be subjected to reprocessing. Note that the score 134 can be one of many other parameters evaluated. To the extent that reprocessing is deemed appropriate, the wafer 104 may be subjected to reprocessing in its entirety. For example, all cells 102 on the wafer 104 may be reprocessed. Optionally, one or more selected portions of the wafer 104 (e.g., a portion of the wafer 104 corresponding to the grayscale image 110) may be reprocessed. To the extent that reprocessing is deemed inappropriate, the wafer 104 may proceed to the next processing step. In some embodiments, if the score 134 is greater than a threshold value (e.g., the amount of etch redeposition is high), the portion of the wafer 104 corresponding to the grayscale image 108 is reprocessed.
[0040] The processing at 136 may be performed manually or automatically. For example, the score 134 may be displayed on a display device. A person may then evaluate the score 134 and appropriately adjust the processing of the wafer 104 based on the score 134. As another example, the process control system may automatically compare the score 134 to a threshold value and may use a conveyor system to automatically route the wafer 104.
[0041] refer to Figure 2A and Figure 2B ,supply Figure 1 Schematic flowcharts 200A, 200B of some optional embodiments of the method.
[0042] exist Figure 2A In FIG. 1 , wafer 104 undergoes processing at 202 that runs in parallel with the image processing at 112. For example, a protective layer may be deposited over cell 102. Figure 1 Instead, processing of wafer 104 is serial with the image processing and is therefore paused until the imaging processing is complete. Processing at 202 may continue according to a series of processing steps that are independent of the method and used to form ICs at each of IC dies 106. Furthermore, once the image processing at 112 is complete, unit 102 may be reprocessed based on score 134 if deemed appropriate. Otherwise, wafer 104 may continue through a series of processing steps.
[0043] exist Figure 2B1 , a plurality of grayscale images 110 of the cell 102 are captured at a plurality of different portions of the wafer 104. Furthermore, each grayscale image 110 is individually subjected to image processing at 112. In some embodiments, the plurality of grayscale images of the cell are captured per IC die 106. By performing image processing on the grayscale images 110 individually, a plurality of scores 134 are generated at 136 for the grayscale images 110 and used to process the wafer 104.
[0044] At 136, the scores 134 are evaluated to select which (if any) parts of the wafer 104 should be subjected to reprocessing. In some embodiments, the scores 134 are evaluated separately, and for each score 134 that exceeds a threshold, the corresponding part of the wafer 104 is selected for reprocessing. In an optional embodiment, the scores 134 are grouped and each group is evaluated separately. For example, the scores 134 can be grouped by IC die 106 or adjacent IC die 106 in pairs or groups. For each group, a comprehensive score is calculated, and if the comprehensive score exceeds a threshold, the corresponding part of the wafer 104 is selected for reprocessing. The comprehensive score of the group may be, for example, the mean, median, maximum, minimum, or standard deviation of the group score. To the extent that reprocessing is considered inappropriate, the wafer 104 may proceed according to a series of processing steps that are independent of the method and used to form an IC at each of the IC die 106. To the extent that reprocessing is considered appropriate, the wafer 104 may be reprocessed as a whole. Optionally, only one or more selected parts of the wafer 104 are reprocessed.
[0045] In some embodiments, multiple grayscale images of the unit are captured per IC die 106 so that each IC die 106 has multiple scores. For each IC die 106, the scores of the IC die are combined into a composite score, and the composite score is compared to a threshold value. For example, the composite score of the IC die can be the average, median, maximum, minimum, or standard deviation of the scores of the IC die. If any one of the IC dies 106 or a threshold number of IC dies 106 has a composite score that exceeds a threshold, the wafer 104 may be subjected to reprocessing in which the unit is removed and reformed. For example, all IC dies 106 may be subjected to reprocessing. As another example, only those IC dies 106 that have a composite score that exceeds a threshold may be subjected to reprocessing.
[0046] although Figure 2B The schematic flow chart 200B is shown without Figure 2A 202 of the process, but an alternative embodiment of the schematic flowchart 200B may include Figure 2A 202 of the processing. Thus, in Figure 2B In an alternative embodiment, wafer 104 may undergo processing at 202 that runs in parallel with the image processing at 112 .
[0047] refer to Figure 3 , Figure 4 , Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C , Fig.10 and Fig.11 , showing in more detail Figure 1 Some embodiments of the method.
[0048] like Figure 3 As shown in FIG. 300 , a wafer 104 is provided upon completion of etching to form cells (not shown) on the wafer 104. The wafer 104 has a plurality of IC dies 106 having the same layout and having separate cell regions 302. The cell regions 302 accommodate the cells and are divided into cell sub-regions 304 (demarcated by dashed lines). The cell sub-regions 304 may, for example, have the same size as the field of view of a SEM or some other suitable imaging device used thereafter to capture grayscale images.
[0049] like Figure 4 As shown, capture Figure 3 FIG. 1 is a grayscale image 110 of a plurality of cells 102 at a cell sub-region 304 of FIG. 1 . Note that different hashes are used to represent different gray levels. The grayscale image 110 provides a top view of the cell 102 and may be captured, for example, by a SEM, RSEM, or some other suitable imaging device.
[0050] like Figure 5A and Figure 5B As shown, non-ROI 116 is identified and subtracted from grayscale image 110 to identify ROI 118. Figure 5A This is shown by the mask image 120, in which the non-ROI 116 is shown in black, and in which the ROI 118 is shown in white. Figure 5B 1, this is illustrated by overlaying black areas of the mask image 120 (corresponding to non-ROIs 116) on the grayscale image 110. The non-ROIs 116 include areas of the grayscale image 110 that correspond to the cells 102, and also include peripheral areas of the grayscale image 110. After subtracting the non-ROIs 116, the ROI 118 corresponds to the remaining portion of the grayscale image 110, and therefore corresponds to areas of the grayscale image 110 between the cells 102.
[0051] like Figure 6 As shown, a grayscale distribution 124 is determined for pixels of the grayscale image 110 in the ROI 118 (e.g., ROI pixels). The ROI pixels correspond to pixels that remain in the grayscale image 110 after subtracting pixels of the non-ROI 116 (e.g., non-ROI pixels). Figure 5AThe mask image 120 is covered by Figure 5B The gray level distribution 124 includes the pixel count of each gray level or the quantile of the gray level.
[0052] like Figure 7 As shown, the average gray level Gavg is determined from the ROI pixels. As will be discussed in more detail below, the average gray level Gavg is used as a reference for normalization across multiple grayscale images.
[0053] like Figure 8 As shown, ROI pixels are divided into normal category, severe category and mild category according to gray level. In optional embodiments, more or fewer categories are applicable. The normal category is defined by ROI pixels corresponding to the possibility of etching redeposition with low probability. Such ROI pixels are identified as pixels with gray levels less than the first threshold value T1. The severe category is defined by ROI pixels corresponding to the possibility of etching redeposition with high probability. Such ROI pixels are identified as ROI pixels with gray levels greater than the second threshold value T2. The mild category is defined by ROI pixels corresponding to the possibility of etching redeposition with moderate probability. Such ROI pixels are identified as ROI pixels with gray levels between the first threshold value T1 and the second threshold value T2.
[0054] The first threshold T1 and the second threshold T2 are functions of the average gray level Gavg and the corresponding offsets O1, O2 to allow normalization. For example, the first threshold T1 may be the sum of the average gray levels Gavg with the first offset O1, and the second threshold T2 may be the sum of the average gray levels Gavg with the second offset O2 greater than the first offset O1. Defining the first threshold T1 and the second threshold T2 as described above has the effect of using the relative brightness difference of the ROI pixels to classify the ROI pixels without having to directly calculate the relative brightness difference of the ROI pixels.
[0055] The relative brightness difference of a given ROI pixel is the difference between the average gray level Gavg and the gray level of the given ROI pixel. Since the classification is performed by mathematically comparing the gray level of the ROI pixel with the first threshold T1 and the second threshold T2, and since the first threshold T1 and the second threshold T2 are the sum of the average gray level Gavg and the corresponding offsets O1, O2, the average gray level Gavg can be subtracted from each side of the mathematical comparison without changing the effect of the mathematical comparison. In addition, the mathematical comparison can be rewritten as a mathematical comparison of the offsets O1, O2 and the relative brightness difference. Therefore, the offsets O1, O2 can be regarded as relative brightness difference thresholds for classifying the ROI pixels, and the first threshold T1 and the second threshold T2 can be regarded as gray level thresholds for classifying the ROI pixels.
[0056] like 9A to 9CAs shown, the heavy and light ROI pixels are shown in white. Fig. 9A , the light ROI image 130 shows the ROI pixels classified as light in white, while the rest of the pixels in the grayscale image 110 (e.g., see Figure 4 ) is black. Fig. 9B In FIG. 1 , the heavy ROI image 128 shows the ROI pixels classified as heavy in white, while the rest of the pixels in the grayscale image 110 are black. Fig. 9C , the light / heavy ROI image 902 shows the ROI pixels classified as light in white, while the remainder of the pixels in the grayscale image 110 are black.
[0057] like Fig.10 1000, which is a grayscale image 110 (e.g., see Figure 4 ) corresponds to the unit sub-area 304. The score 134 is determined based on the number of ROI pixels in different categories. The greater the ratio of heavy ROI pixels to the total number of ROI pixels, the higher the score. In addition, in some embodiments, the greater the ratio of light ROI pixels to the total number of ROI pixels, the higher the score. In some embodiments, the score 134 is a percentage of heavy ROI pixels (e.g., the number of heavy ROI pixels divided by the total number of ROI pixels multiplied by 100). In other embodiments, the score 134 is a percentage of heavy and light ROI pixels. In some embodiments, the score 134 takes into account the number of heavy and light ROI pixels, but heavy ROI pixels have a greater weight than light ROI pixels. For example, the weight of heavy ROI pixels in the score 134 may be 1.5 to 10 times that of light ROI pixels. In some of these embodiments, the score 134 may be equal to the ratio of heavy ROI pixels to the total number of ROI pixels multiplied by a weighting factor (e.g., 1.5-10) plus the ratio of light ROI pixels to the total number of ROI pixels.
[0058] like Fig.11 As shown in FIG. 1100 , wafer 104 is subjected to processing based on score 134. Specifically, a determination 1102 is made based on score 134 as to whether to reprocess wafer 104. Determination 1102 may also be based on other suitable parameters. In some embodiments, if score 134 exceeds a threshold, reprocessing is deemed appropriate. To the extent that reprocessing is deemed appropriate, wafer 104 may be subjected to reprocessing 1104. Otherwise, wafer 104 may continue to be subjected to a series of processing steps that began before the method of forming IC die 106. In other words, the next processing step 1106 in the series of processing steps may be performed.
[0059] Although the reference method describes Figure 3 , Figure 4 , Figure 5A, Figure 5B , Figures 6 to 8 , 9A to 9C , Fig.10 and Fig.11 , but it should be understood that Figure 3 , Figure 4 , Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C , Fig.10 and Fig.11 The structure shown is not limited to the method but can be separated from it individually. Figure 3 , Figure 4 , Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C , Fig.10 and Fig.11 Although described as a series of actions, it should be appreciated that the order of the actions may be changed in other embodiments. Figure 3 , Figure 4 , Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C , Fig.10 and Fig.11 A particular set of actions is shown and described, but some of the actions shown and described may be omitted in other embodiments. Additionally, actions not shown and / or described may be included in other embodiments.
[0060] In addition, it should be understood that Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C and Fig.10 It can be performed automatically or semi-automatically by an electronic processing device. For a semi-automatic embodiment, Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C and Fig.10 For example, Figure 5A and Figure 5B As described, manual input may be provided to identify cells 102 in grayscale image 110, and Figures 6 to 8 , 9A to 9C and Fig.10 The remainder of the described image processing may be fully automatic. For an automatic embodiment, Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C and Fig.10 No human input may be provided when an action is being taken. This may allow, for example, high throughput.
[0061] refer to Fig.12 ,supply Figure 3 , Figure 4 , Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C , Fig.10 and Fig.11 1200 is a block diagram of some embodiments of the method of . In some embodiments, the method is performed using artificial intelligence (AI) for optimizing and enhancing performance.
[0062] At 1202, a grayscale image of a plurality of cells at a portion of a wafer is captured, wherein the grayscale image provides a top view and is captured when etching is completed to form the cells. For example, see Figure 3 and Figure 4 .
[0063] At 1204, a non-ROI is identified in the grayscale image, wherein the non-ROI includes a region of the grayscale image corresponding to the cell. Figure 5A and Figure 5B .
[0064] At 1206, the non-ROI is subtracted from the grayscale image to determine the ROI, wherein the ROI corresponds to the remaining portion of the grayscale image. Figure 5A and Figure 5B .
[0065] At 1208, the grayscale distribution of the pixels of the grayscale image in the ROI is determined. Figure 6 .
[0066] At 1210, an average gray level is determined for the ROI pixels. For example, see Figure 7 .
[0067] At 1212, the ROI pixels are classified into a heavy category, a light category, and a normal category by relative brightness difference, where the relative brightness difference of a given ROI pixel is the gray level of the given ROI pixel minus the average gray level. Figure 8 and 9A to 9C .
[0068] At 1214, a score is determined based on the number of heavy ROI pixels, where the score is proportional to the ratio of heavy ROI pixels to the total number of ROI pixels. Fig.10 .
[0069] At 1216, the wafer is processed based on the score. For example, see Fig.11 .
[0070] Although this article will Fig.12The block diagram 1200 is shown and described as a series of actions or events, but it should be understood that the illustrated order of such actions or events should not be interpreted in a limiting sense. For example, certain actions may occur in a different order and / or simultaneously with other actions or events in addition to those actions or events shown and / or described herein. In addition, all of the illustrated actions may not be required to implement one or more aspects or embodiments described herein, and one or more of the actions depicted herein may be performed in one or more separate actions and / or stages.
[0071] refer to Fig.13 , Fig.14 and FIG. 15A to FIG. 15C , some optional embodiments of the images described above are provided. Fig.13 In, provide Figure 4 Some alternative embodiments of the grayscale image 110 of FIG. 1 , wherein the ellipse is removed and more details are shown. Fig.14 In, provide Figure 5A Some optional embodiments of the mask image 120, wherein the ellipse is removed. FIG. 15A to FIG. 15C In, respectively provide 9A to 9C Some optional embodiments of the light ROI image 130, the heavy ROI image 128, and the light / heavy ROI image 902 in FIG. 1 , wherein the ellipse is removed.
[0072] refer to Fig.16 ,supply Fig.10 1000 , where each of the unit sub-regions 304 has a single score 134. Note that due to space limitations, the individual scores 134 for the unit sub-regions 304 are only shown for one of the IC dies 106. For each sub-region, by performing Figure 4 , Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C , Fig.10 The actions described and / or Fig.12 Actions 1202 through 1214 in FIG. 12 are used to determine score 134 separately for each of unit sub-areas 304 .
[0073] Scores 134 are combined by IC die 106 into score 134 die. For example, the die score for a given IC die may be the mean, median, maximum, minimum, or standard deviation of the scores 134 for the corresponding unit sub-regions 304. Die score 134 die may then be used to determine whether to reprocess wafer 104. For example, if a threshold number of die score 134 die exceed a threshold, wafer 104 may be completely reprocessed. As another example, if any one of the die score 134 die exceeds a threshold, only the corresponding IC die may be reprocessed. If the aforementioned reprocessing criteria are not met, wafer 104 may proceed according to a series of processing steps to form an IC at each of IC die 106.
[0074] refer to Fig.17A and Fig. 17B , Fig.18 , Fig.19 , Fig. 20A , Fig. 20B and Figures 21 to 28 , provide some embodiments of methods for forming ICs and performing non-destructive inspection of cell etch redeposition. Figure 1 , Figure 2A , Figure 2B and Fig.12 Non-destructive inspection is performed by any of the methods in .
[0075] like Fig.17A and Fig. 17B As shown in diagrams 1700A and 1700B, a plurality of IC dies 106 are partially formed on a wafer 104 . Fig.17A A top layout diagram 1700A is provided, and Fig. 17B Provide along Fig.17A FIG. 17 is a cross-sectional view of a portion of line A in FIG. 1700B. The IC die 106 has a single cell region 302. At the cell region 302, a plurality of bottom electrode vias (BEVAs) 1702b cover the wafer 104 on top of the interconnect structure 1704. The BEVAs 1702b are conductive and include corresponding plugs 1706 and corresponding liners 1708 that cap the underside of the plugs 1706. The liner 1708 is configured to prevent material from outdiffusion from the plugs 1706.
[0076] The interconnect structure 1704 includes a plurality of conductive lines 1710 and a plurality of vias 1702. The vias 1702 include BEVAs 1702b and are alternately stacked with the conductive lines 1710 to define a conductive path from the BEVAs 1702b to an access transistor 1712 (only partially shown) located below the interconnect structure 1704 on the wafer 104. The conductive lines 1710 and the vias 1702 are located in an interlayer dielectric (ILD) layer 1714, an intermetal dielectric (IMD) layer 1716, and an etch stop layer (ESL) 1718 stacked above the wafer 104.
[0077] Access transistor 1712 is defined in part by wafer 104 and includes corresponding source / drain regions 1720 in wafer 104. In addition, although not visible, access transistor 1712 includes a corresponding gate stack that interfaces with source / drain regions 1720. Access transistor 1712 may be, for example, a metal oxide semiconductor field effect transistor (MOSFET) or some other suitable type of transistor. Wafer 104 may be, for example, or include a bulk wafer of single crystal silicon, a silicon on insulator (SOI) wafer, or some other suitable type of semiconductor wafer.
[0078] In some embodiments, access transistors 1712 are separated from each other by trench isolation structures 1722. Trench isolation structures 1722 include dielectric materials and may be, for example, shallow trench isolation (STI) structures, deep trench isolation (DTI) structures, or some other suitable trench isolation structure.
[0079] like Fig.18 As shown in the cross-sectional view 1800 of FIG. 1 , a bottom electrode layer 1802, an MTJ film 1804, and a top electrode layer 1806 are deposited above the interconnect structure 1704. Note that for compactness, only the upper portion of the interconnect structure 1704 is shown hereinafter. The remaining portion of the interconnect structure 1704 and other structures below the interconnect structure 1704 are shown in FIG. Fig. 17B As shown. The MTJ film 1804 includes a fixed layer 1808, a barrier layer 1810 covering the fixed layer 1808, and a free layer 1812 covering the barrier layer 1810. The fixed layer 1808 and the free layer 1812 are ferromagnetic. In addition, the fixed layer 1808 has a fixed magnetization, and the free layer 1812 has a freely changeable magnetization.
[0080] like Fig.19 As shown in the cross-sectional view 1900 of FIG. 1 , at locations for forming MTJ cells thereafter, hard masks 1902 are formed over the top electrode layer 1806, respectively. The hard masks 1902 may be or include, for example, silicon nitride and / or some other suitable dielectric.
[0081] like Fig. 20A and Fig. 20B2000A and 2000B, the bottom electrode layer 1802 and the top electrode layer 1806 (see, for example, Fig.19 ) and MTJ film 1804 (see, for example, Fig.19 ) performs etching to form a plurality of cells 102 on BEVA 1702b, respectively, separate from BEVA 1702b. FIG. 2000A shows a cross-sectional view 2000B of a cell 102, and FIG. 2000B shows a view 2000B of an etching process tool performing the etching.
[0082] Focus on Fig. 20A , the cell 102 includes a single bottom electrode 2002, a single MTJ 2004 respectively covering the bottom electrode 2002, and a single top electrode 2006 respectively covering the MTJ 2004. The MTJ 2004 includes a single fixed element 2008, a single barrier element 2010 respectively covering the fixed element 2008, and a single free element 2012 respectively covering the barrier element 2010.
[0083] During operation of any of the cells 102, the corresponding barrier element 2010 selectively allows electrons to quantum mechanically tunnel through the barrier element 2010. When the magnetizations of the corresponding fixed element 2008 and the free element 2012 are antiparallel, quantum mechanical tunneling may be blocked. As such, the cell may have a high resistance and may be in a first data state. When the magnetizations of the fixed element 2008 and the free element 2012 are parallel, quantum mechanical tunneling may be allowed. As such, the cell may have a low resistance and may be in a second data state.
[0084] The etching is performed by ion beam etching (IBE). However, the etching can alternatively be performed by some other suitable type of dry etching or some other suitable type of etching using ion bombardment. IBE relies on ion bombardment to etch. Specifically, kinetic energy is transferred from the ions to the layer being etched to interrupt the material of the layer. Due to the transfer of kinetic energy, the etched material tends to "fly away" and redeposit elsewhere in some uncontrollable manner. Depending on the density of this etch redeposit 2014, the etch redeposit 2014 can increase leakage current and reduce yield. For example, when the etch redeposit 2014 is located on the sidewall of the unit 102, the etch redeposit 2014 can create a conductive bridge that increases the leakage current from the fixed element 2008 to the free element 2012. The etch redeposit may, for example, be or include tantalum, ruthenium, some other suitable conductive material, or any combination of the foregoing.
[0085] Focus on Fig. 20B , the wafer table 2016 is located in the process chamber 2018 and supports the wafer 104 (see, for example, Fig.17A and Fig. 17B ), and thus supports Fig.19 structure. In addition, the wafer stage 2016 is configured to rotate the wafer 104 around an axis extending into and out of the page. An exhaust pump 2020 is located at an exhaust port 2022 of the process chamber 2018 along the bottom of the process chamber 2018. An ion beam source 2024 is along the top of the process chamber 2018 and generates an ion beam 2026 using a processing gas from a gas delivery system 2028. An imaging device 2030 is further located in the process chamber 2018. The imaging device 2030 may be, for example, a SEM, an RSEM, or some other suitable type of imaging device. In an optional embodiment, the imaging device 2030 is outside the process chamber 2018.
[0086] If Fig.21 As shown in FIG. 2100 , the imaging device 2030 captures multiple grayscale images 110 of the unit 102 (e.g., see Fig. 20A ). The capture is performed in-situ in the process chamber 2018. The so-called in-situ means that the wafer 104 is captured from the Fig. 20A and Fig. 20B The start of the etching described to the end of the image capture remains in the process chamber 2018. In an alternative embodiment, the capture is performed outside of the process chamber 2018. The grayscale image 110 corresponds to the cell 102 at different portions of the wafer 104. For example, as shown in FIG. Figure 2B and Fig.16 As described, each IC die 106 may have multiple grayscale images, and the grayscale images of each IC die 106 may correspond to different portions (e.g., cell regions and / or sub-regions) of the IC die. In some embodiments, the tilt of the wafer 104 is changed prior to capture to better focus the field of view of the imaging device 2030 on the cell 102.
[0087] If Fig. 22 As shown in FIG. 2200 , the grayscale image 110 is processed to determine the individual die scores 134 of the IC die 106. The die scores 134 are proportional to the etch residue, and thus, a higher score results in a higher leakage current and a lower yield. To determine the die scores 134, the die is processed according to Figure 1 The image processing at 112 and / or according to Fig.12 The grayscale image 110 is processed separately to determine a single score (e.g., Figure 1 134 in). To the extent that an IC die is associated with a single grayscale image, the score of the single grayscale image is the die score of the IC die. To the extent that an IC die is associated with multiple grayscale images, the scores of the grayscale images are combined into a die score for the IC die. The scores may be combined using a mean function, a median function, a minimum function, a maximum function, a standard deviation function, or some other suitable function.
[0088] Also by Fig. 22 2200 shows that the die scores 134 of the IC dies 106 are evaluated to determine which, if any, of the IC wafer 106 should be reprocessed. Specifically, the die scores 134 of the IC dies 106 are compared to a threshold, and any IC die whose score exceeds the threshold is marked for reprocessing. For example, the threshold may be 10, such that both IC dies 106 are marked for reprocessing. Note that the box with a bold line highlights the marked IC dies. In an optional embodiment, another suitable process is performed to mark which, if any, of the IC dies 106 are to be reprocessed. To the extent that one or more IC dies 106 are marked for reprocessing, only the one or more IC dies 106 marked for reprocessing are reprocessed. In an optional embodiment, the wafer 104 may be completely reprocessed.
[0089] like Figure 23 to Figure 25 As shown in the cross-sectional views 2300 to 2500, Fig. 22 The IC dies 106 marked for reprocessing are reprocessed, and in some embodiments, the entire wafer 104 is reprocessed. Fig.23 , unit 102 and BEVA 1702b are removed. The removal may be performed, for example, by chemical mechanical polishing (CMP), etching, some other suitable removal process, or any combination of the foregoing. Fig.24 In addition, the bottom electrode layer 1802, the MTJ film 1804, the top electrode layer 1806 and the hard mask 1902 are as described above. Fig.18 and Fig.19 described and reformed. Fig.25 In the embodiment, the bottom electrode layer 1802 and the top electrode layer 1806 are etched (for example, see Fig.24 ) and MTJ film 1804 (see, for example, Fig.24 ) to form unit 102.
[0090] like Fig.26 As shown in the diagram, repeat about Fig.21 and Fig. 22 The actions described above are to determine which IC dies 106 (if any) to reprocess. Fig.21 As described, an imaging device 2030 is provided to capture a plurality of grayscale images 110 of a cell 102. Then, as described with respect to Fig. 22 As described, grayscale image 110 is processed to determine a single die score 134 for IC die 106. Fig. 22 As described, the die scores 134 are evaluated to determine which IC dies 106, if any, to reprocess. To the extent that reprocessing is appropriate, the steps described above are repeated with respect to Figure 23 to Figure 26 To the extent that reworking is not appropriate, the IC continues to completion. For example, interconnect structure 1704 may extend over cell 102.
[0091] If Fig. 27 As shown in the cross-sectional view 2700 of FIG. 27 , a protective layer 2702 is deposited over the cell 102. In some embodiments, Fig.21 The grayscale image 110 is subjected to Fig. 22 The protective layer 2702 is deposited while the treatment is performed at Figure 23 to Figure 25 The depicted reprocessing includes removing protective layer 2702 and redepositing protective layer 2702. In an alternative embodiment, protective layer 2702 is deposited only after reprocessing of wafer 104 is complete. Protective layer 2702 may be or include, for example, silicon nitride and / or some other suitable dielectric.
[0092] If Fig.28 As shown in cross-sectional view 2800 of FIG. 1 , interconnect structure 1704 is completed over cell 102. This includes forming additional conductive lines 1710 and additional vias 1702 stacked over cell 102 and electrically coupled to cell 102 in additional IMD layers 1716.
[0093] Although reference method description 17A and Fig. 17B , Fig.18 , Fig.19 , Fig. 20A , Fig. 20B and Figures 21 to 28 , but it should be understood that 17A and Fig. 17B , Fig.18 , Fig.19 , Fig. 20A , Fig. 20B and Figures 21 to 28 The structure shown is not limited to this method, but can be separated from this method individually. Fig. 17B , Fig.18 , Fig.19 , Fig. 20A , Fig. 20B and Figures 21 to 28 Although described as a series of actions, it should be understood that the order of the actions may be changed in other embodiments. Fig. 17B , Fig.18 , Fig.19 , Fig. 20A , Fig. 20B and Figures 21 to 28 A particular set of actions is shown and described, but some of the actions shown and described may be omitted in other embodiments. Additionally, actions not shown and / or described may be included in other embodiments.
[0094] refer to Fig.29 , providing 17A and Fig. 17B, Fig.18 , Fig.19 , Fig. 20A , Fig. 20B and Figures 21 to 28 Block diagram 2900 of some embodiments of the method.
[0095] At 2902, an etch is performed on a wafer having a plurality of partially formed integrated circuit (IC) dies, wherein the etching forms a single region of a cell at the IC die. See, for example, 17A and Fig. 17B , Fig.18 , Fig.19 , Fig. 20A and Fig. 20B .
[0096] At 2904, a grayscale image of the wafer is captured, wherein the grayscale image provides a top view of different portions of the wafer. Fig.21 .
[0097] At 2906, the grayscale image is processed to determine a single die score for the IC die, wherein the die score for the IC die is proportional to a ratio of region of interest (ROI) pixels having light and / or heavy gray levels to a total number of ROI pixels, and wherein ROI pixels are pixels located in a region of the IC die and corresponding to pixels of a discrete unit. For example, see Fig. 22 and Fig.26 .
[0098] At 2908, a determination is made as to whether to reprocess the wafer based on the score. Fig. 22 and Fig.26 .
[0099] At 2910, the wafer is reprocessed in response to the score satisfying the reprocessing criteria. For example, see Figure 23 to Figure 25 .
[0100] At 2912, processing of the wafer continues according to a series of processing steps to form an IC at each of the IC dies in response to the score failing the rework criteria. For example, a next processing step in the series of processing steps may be performed to form an MRAM device on the wafer 104. For example, see Fig. 27 and Fig.28 .
[0101] Although this article will Fig.29The block diagram 2900 is shown and described as a series of actions or events, but it should be understood that the illustrated order of such actions or events should not be interpreted in a limiting sense. For example, certain actions may occur in a different order and / or simultaneously with other actions or events in addition to those shown and / or described herein. In addition, all of the illustrated actions may not be required to implement one or more aspects or embodiments described herein, and one or more of the actions depicted herein may be performed in one or more separate actions and / or stages.
[0102] refer to Fig.30 ,supply Fig.29 Block diagram 3000 of some optional embodiments of a method of wherein the method employs parallel processing of a wafer and a grayscale image. In particular, processing of the wafer continues at 2912' while the grayscale image is processed at 2906. Further, processing of the wafer continues according to a series of processing steps to form an IC at each of the IC dies. In response to satisfying a rework criterion, processing according to the series of processing steps is stopped at 2910' and rework is performed.
[0103] Although this article will Fig.30 The block diagram 3000 is shown and described as a series of actions or events, but it should be understood that the illustrated order of such actions or events should not be interpreted in a limiting sense. For example, certain actions may occur in a different order and / or simultaneously with other actions or events in addition to those actions or events shown and / or described herein. In addition, all of the illustrated actions may not be required to implement one or more aspects or embodiments described herein, and one or more of the actions depicted herein may be performed in one or more separate actions and / or stages.
[0104] refer to Fig.31 , a schematic diagram 3100 of some embodiments of a system for forming an IC and performing non-destructive inspection of cell etch redeposition is provided. The system may, for example, be configured to perform any of the above methods, including (but not limited to) Figure 1 , Figure 2A , Figure 2B , Fig.12 , Fig.29 and Fig.30 The method in .
[0105] The etch processing tool 3102 is configured to perform etching on the wafer 104 to form a plurality of cells (not separately shown) distributed across the plurality of IC dies 106 of the wafer 104. The cells may be, for example, MTJ cells, logic cells, or some other suitable type of cells. Fig. 20A Provide non-limiting examples of units, and Fig. 20B A non-limiting example of an etch processing tool 3102 is provided. In addition, Fig.19, Fig. 20A and Fig. 20B Non-limiting examples of processes for performing etching are provided.
[0106] The imaging device 2030 is associated with the etch processing tool 3102 and is configured to capture a grayscale image 110 of the cell while the wafer 104 is still in a process chamber of the etch processing tool 3102 when etching is completed. Fig.21 Non-limiting examples of imaging device 2030 and etch processing tool 3102 are provided, respectively, and non-limiting examples of capturing grayscale image 110 are also provided. Grayscale image 110 provides a top view of cells at different parts of the wafer. Imaging device 2030 can be, for example, a SEM or some other suitable imaging device.
[0107] The image processing device 3106 is configured to automatically or semi-automatically process the grayscale image 110 alone to determine the score 134 of the grayscale image 110. For example, Figure 1 , Figure 2A , Figure 2B and Fig.12 Any description and / or according to Fig.12 In some embodiments, the processing for each of the grayscale images 110 is performed according to Figure 1 The processing of a single grayscale image is performed by the imaging processing at 112. A non-limiting example of processing for a single grayscale image is as follows: Figure 3 , Figure 4 , Figure 5A , Figure 5B , Figures 6 to 8 , 9A to 9C and Fig.10 The image processing device 3106 is electronic and may be, for example, a computer, an application specific integrated circuit (ASIC), a microcontroller, or some other suitable type of electronic device.
[0108] In some embodiments, the image processing device 3106 includes an electronic processor 3108 and an electronic memory 3110. The electronic processor 3108 retrieves processor-executable instructions from the electronic memory 3110 for performing the processing of the grayscale image 110. In addition, the electronic processor 3108 executes the retrieved processor-executable instructions to perform the processing of the grayscale image 110. In some embodiments, the image processing device 3106 also includes or is associated with a display device 3112 and an HID 3114. The display device 3112 may, for example, be configured to display the score 134, and / or may, for example, be configured to display a GUI for interacting with the processing of the grayscale image 110. The HID 3114 may, for example, be configured to allow an individual to interact with the processing of the grayscale image 110 via the GUI. For example, when performing Figure 1At the action at 114, the individual may use the HID 3114 to identify the unit within the grayscale image.
[0109] The process controller 3116 is configured to evaluate the scores 134 to determine how to process the wafer 104. In particular, the process controller 3116 is configured to determine whether the wafer 104 is to be reprocessed or whether the wafer 104 should continue to be processed according to a series of processing steps to form an IC at each of the IC dies 106. For example, the determination may be made by grouping the scores 134 and / or comparing the scores 134 to one or more thresholds. In some embodiments, if a threshold number of the scores 134 exceeds a threshold, reprocessing may be in order. Otherwise, the wafer 104 may continue without reprocessing. For example, as described with respect to Figure 1 , Figure 2A and Figure 2B Any of the actions in action 136 described above, performing an evaluation, such as with respect to Fig.11 Figure 17 Fig. 22 and Fig.26 Any of the descriptions of Fig.12 The evaluation may be performed as described in action 1216 of , or according to any other suitable process. To the extent that reprocessing is sequential, the process controller 3116 is configured to control the conveyor system 3118 to convey the wafer 104 to one or more reprocessing process tools 3120 for reprocessing. Otherwise, the process controller 3116 is configured to control the conveyor system 3118 to convey the wafer 104 to a deposition processing tool 3122 or some other suitable processing tool to continue processing the wafer according to a series of processing steps to form an IC at each of the IC dies 106.
[0110] In view of the foregoing, some embodiments of the present invention provide a method comprising: capturing a grayscale image of a plurality of cells on a wafer, wherein the grayscale image provides a top view and is captured when etching is completed to form the cells; identifying the cells in the grayscale image; subtracting a region of the grayscale image corresponding to the identified cells from the grayscale image; scoring an amount of etch residue on the sidewalls of the cells and in the grooves between the cells based on the grayscale levels of the remaining pixels at the remaining portion of the grayscale image; and processing the wafer based on a score of the scoring. In some embodiments, the cells are MTJs. In some embodiments, the cells are gate stacks of logic devices. In some embodiments, the method further comprises: prior to the scoring, subtracting a peripheral region of the grayscale image extending in a closed path along the periphery of the grayscale image. In some embodiments, the method further comprises: determining an average grayscale level of the remaining pixels; and determining the score as a percentage of the remaining pixels whose grayscale levels exceed a threshold, wherein the threshold is the average grayscale level plus a non-zero offset. In some embodiments, the method further comprises: performing the etching in a process chamber, wherein the capturing is performed in the process chamber, and wherein the wafer is continuously in the process chamber from the start of the etching to the end of the capturing. In some embodiments, the capturing is performed by SEM. In some embodiments, the processing comprises: re-forming the plurality of cells in response to the score exceeding a threshold. In some embodiments, the processing comprises: depositing a protective layer covering the cells in response to the threshold exceeding the score.
[0111] In some embodiments, the present invention provides another method, comprising: capturing a grayscale image of a plurality of cells on a wafer, wherein the grayscale image provides a top view and is captured after etching to form the cells; performing image processing on the grayscale image, the image processing comprising: identifying the cells in the grayscale image; determining region of interest (ROI) pixels based on the identification, wherein the ROI pixels include pixels between the identified cells but not at the identified cells; determining an average grayscale level between the ROI pixels; and determining a ratio of ROI pixels whose grayscale exceeds a threshold to a total number of ROI pixels, wherein the threshold is greater than the average grayscale level; and processing the wafer based on the ratio. In some embodiments, the method further comprises: forming an IC on the wafer according to a series of processing steps, wherein the series includes the etching, and wherein the next processing step in the series is performed in parallel with the image processing. In some embodiments, the method further comprises: forming an IC on the wafer according to a series of processing steps, wherein the series includes the etching, and wherein the next processing step in the series is performed in series with the image processing. In some embodiments, the cells are arranged in an array comprising a plurality of rows and a plurality of columns. In some embodiments, the cells have a periodic pattern. In some embodiments, the method further comprises: providing the wafer, wherein the wafer comprises a plurality of IC dies covered by a multilayer stack; and performing etching on the multilayer stack to form the plurality of cells, wherein the cells are at each of the IC dies. In some embodiments, the method further comprises: capturing a plurality of grayscale images corresponding to different subsets of the cells, wherein the plurality of grayscale images comprises the grayscale image and is captured after the etching, and wherein the image processing is performed separately on each of the grayscale images.
[0112] In some embodiments, the present invention also provides a system, comprising: an etch process tool configured to etch a multilayer film on top of a wafer to form a plurality of cells from the multilayer film; an imaging device configured to capture a grayscale image of the cell, wherein the grayscale image provides a top view of the cell; and an image processing device configured to: identify the cell in the grayscale image; subtract non-ROI pixels from the grayscale image, wherein the non-ROI pixels include pixels at the identified cell; and generate a score for a certain amount of etch residue on the sidewall of the cell and in the groove between the cells based on the grayscale level of the remaining pixels at the grayscale image; and a process controller configured to process the wafer based on the score. In some embodiments, the etch process tool comprises a process chamber, and wherein the imaging device is configured to capture the grayscale image while the wafer is in the process chamber. In some embodiments, the imaging device is a SEM. In some embodiments, the cell is an MJT.
[0113] The features of several embodiments are summarized above so that those skilled in the art can better understand aspects of the present invention. Those skilled in the art should understand that they can easily use the present invention as a basis to design or modify other processes and structures for implementing the same purpose and / or achieving the same advantages as the embodiments introduced herein. Those skilled in the art should also appreciate that such equivalent constructions do not deviate from the spirit and scope of the present invention, and that they can make various changes, substitutions and modifications herein without departing from the spirit and scope of the present invention.
Claims
1. A method for non-destructive inspection of unit etching redeposition, comprising: capturing a grayscale image of a plurality of cells on a wafer, wherein the grayscale image provides a top view and is captured upon completion of etching to form the plurality of cells; identifying the plurality of cells in the grayscale image; subtracting from the grayscale image a region of the grayscale image corresponding to the identified plurality of cells; scoring an amount of etch residue on sidewalls of the plurality of cells and in grooves between the plurality of cells based on the grayscale levels of the remaining pixels at the remaining portion of the grayscale image; and processing the wafer differently depending on whether a score of the rating exceeds a first threshold, Therein, the scoring includes determining a percentage of the gray levels of the remaining pixels between a second threshold and a maximum gray level, and the score is based on and proportional to the percentage.
2. The method according to claim 1, wherein: The plurality of cells include magnetic tunnel junctions.
3. The method according to claim 1, wherein: The plurality of cells are gate stacks of a logic device.
4. The method according to claim 1, further comprising: Prior to the scoring, a peripheral area of the grayscale image extending in a closed path along the periphery of the grayscale image is subtracted.
5. The method according to claim 1, further comprising: An average gray level of the remaining pixels is determined, wherein the second threshold is the average gray level plus a non-zero offset.
6. The method according to claim 1, further comprising: The etching is performed in a process chamber, wherein the capturing is performed in the process chamber, and wherein the wafer is continuously in the process chamber from the start of the etching to the end of the capturing.
7. The method according to claim 1, wherein: The capture was performed by scanning electron microscopy.
8. The method according to claim 1, wherein: The processing includes reforming the plurality of cells in response to the score exceeding the first threshold.
9. The method according to claim 1, wherein: The processing includes depositing a protective layer covering the cell in response to a threshold exceeding the score.
10. A method for non-destructive inspection of cell etching redeposition, comprising: capturing a grayscale image of a plurality of cells on a wafer, wherein the grayscale image provides a top view and is captured after etching to form the plurality of cells; Performing image processing on the grayscale image, the image processing comprising: identifying the plurality of cells in the grayscale image; determining region of interest pixels based on the identifying, wherein the region of interest pixels include pixels between the identified plurality of cells but not at the identified plurality of cells; determining an average gray level between pixels in the region of interest; and determining a first ratio of a first number of the ROI pixels to a total number of the ROI pixels, wherein the first number is the number of the ROI pixels having a gray level exceeding a first threshold, and the first threshold is greater than the average gray level; determining a second ratio of a second number of the ROI pixels to the total number of the ROI pixels, wherein the second number is the number of the ROI pixels having a gray level exceeding a second threshold and less than the first threshold; and combining the first ratio and the second ratio into a score that increases and decreases with the first ratio and the second ratio; and The wafer is processed based on the score.
11. The method according to claim 10, further comprising: Integrated circuits are formed on the wafer according to a series of processing steps, wherein the series includes the etching, and wherein a next processing step in the series is performed in parallel with the image processing.
12. The method according to claim 10, further comprising: Integrated circuits are formed on the wafer according to a series of processing steps, wherein the series includes the etching, and wherein a next processing step in the series is performed serially with the image processing.
13. The method according to claim 10, wherein: The plurality of cells are arranged in an array including a plurality of rows and a plurality of columns.
14. The method according to claim 10, wherein: The plurality of cells have a periodic pattern.
15. The method according to claim 10, further comprising: providing the wafer, wherein the wafer comprises a plurality of integrated circuit dies covered by a multi-layer stack; and Etching is performed on the multi-layer stack to form the plurality of cells, wherein the plurality of cells are at each of the integrated circuit dies.
16. The method according to claim 10, further comprising: A plurality of grayscale images corresponding to different subsets of the plurality of cells are captured, wherein the plurality of grayscale images include the grayscale image and are captured after the etching, and wherein the image processing is performed separately on each of the grayscale images.
17. A system for non-destructive inspection of cell etch redeposition, comprising: an etching process tool configured to etch a multi-layer film on top of a wafer to form a plurality of cells from the multi-layer film; an imaging device configured to capture a grayscale image of the plurality of cells, wherein the grayscale image provides a top view of the plurality of cells; and The image processing device is configured as follows: identifying the plurality of cells in the grayscale image; subtracting non-region-of-interest pixels from the grayscale image, wherein the non-region-of-interest pixels include pixels at the identified plurality of cells; and generating a score for an amount of etch residue on sidewalls of the plurality of cells and in grooves between the plurality of cells based on the grayscale levels of the remaining pixels at the grayscale image; and a process controller configured to process the wafer differently depending on whether the score exceeds a first threshold, Wherein generating the score comprises determining a percentage of the remaining pixels having a gray level between a second threshold and a maximum gray level, and the score is based on and proportional to the percentage.
18. The system of claim 17, wherein: The etch processing tool includes a process chamber, and wherein the imaging device is configured to capture the grayscale image while the wafer is in the process chamber.
19. The system of claim 17, wherein: The imaging device is a scanning electron microscope.
20. The system of claim 17, wherein: The plurality of cells are magnetic tunnel junctions.
Citation Information
Patent Citations
Apparatus and method for contact failure inspection in semiconductor devices
CN1239321A
Semiconductor wafer inspecting method using a scanning electron microscope (SEM)
DE19860704A1
Method and system for monitoring IC process
TW200527572A
Image defect inspection method, image defect inspection apparatus and appearance inspection apparatus
US20040062432A1
Method and system for monitoring IC process
US20050152594A1
