Binning-enhanced defect detection method for 3D wafer structures

Abstract

Location-based binning can separate defects in different channel hole rows in a 3D NAND structure into corresponding bins. A one-dimensional projection of the image is generated and a one-dimensional curve is formed. A mask is generated from the one-dimensional curve. Using the mask, defects in the image are detected and location-based binning is performed.

Description

Technology Field

[0001] Cross-reference regarding related applications

[0002] This application claims priority to a provisional patent application filed on September 4, 2020, and assigned U.S. application number 63 / 074,487, the disclosures of which are incorporated by reference.

[0003] Field of disclosure

[0004] The present disclosure relates to the detection of defects on semiconductor wafers. Background Technology

[0005] Advancements in the semiconductor manufacturing industry are placing greater demands on yield management, particularly on metrology and inspection systems. While critical dimensions continue to shrink, the industry needs to reduce the time required to achieve high-yield, high-value production. Minimizing the total time from detecting yield issues to resolving them determines the return on investment for semiconductor manufacturers.

[0006] Manufacturing semiconductor devices, such as logic and memory devices, typically involves processing a semiconductor wafer using a large number of manufacturing processes to form various features and multiple levels of the semiconductor devices. For example, lithography is a semiconductor manufacturing process involving transferring a pattern from a reticle onto a photoresist arranged on a semiconductor wafer. Additional examples of semiconductor manufacturing processes include, but are not limited to, chemical mechanical polishing (CMP), etching, deposition, and ion implantation. An array of multiple semiconductor devices manufactured on a single semiconductor wafer may be separated into individual semiconductor devices.

[0007] Inspection processes are used at various stages during semiconductor manufacturing to detect defects on wafers and thereby increase yield and, consequently, profit in the manufacturing process. Inspection has always been an important part of manufacturing semiconductor devices, such as integrated circuits (ICs). However, as the dimensions of semiconductor devices decrease, inspection becomes much more critical for the successful manufacturing of acceptable semiconductor devices because smaller defects can cause the devices to fail. For example, as the dimensions of semiconductor devices decrease, it has become necessary to detect defects of a smaller size, because even relatively small defects can cause unwanted aberrations in the semiconductor devices.

[0008] As the demand for smaller semiconductor devices continues to increase, shrinking semiconductor devices, such as memory, has become more difficult due to the rapid increase in lithography-related costs and the numerous process steps associated with pitch splitting technologies. Vertical memory, such as 3D NAND memory, appears to be a promising direction for increasing memory density. The implementation of 3D NAND involves building transistors (bits) vertically rather than orienting memory structures in a planar manner. Compared to the planar approach, increasing the number of bits can be achieved with fewer process steps, relaxed lithography sizes, and lower manufacturing costs.

[0009] 3D NAND has layers containing holes. Semiconductor manufacturers are typically concerned about which row of these channel holes contains a defect. Defects in holes closer to the channel can be more problematic. Many inspection systems lack the resolution to determine where defects are located. There is a need for improved semiconductor wafer inspection systems to be implemented for vertical semiconductor devices, such as 3D NAND memory or other vertical stacks. Previous methods, such as Image-based SuperCell (IBS), have been used to detect defects on 3D NAND structures. However, IBS typically cannot handle local and global gray level (GL) variations that introduce separation errors. Noise can also affect the results.

[0010] Therefore, improved methods and systems for defect detection are needed.

[0011] A method is provided in a first embodiment. The method includes the step of receiving an image from a processor. The image is an image of a three-dimensional structure of a semiconductor wafer and may be generated by a broadband plasma inspection system. A one-dimensional projection of the image is generated using the processor and thereby forms a one-dimensional curve. A mask is generated from the one-dimensional curve of the image using the processor. Defects on the image are detected using the mask using the processor. Location-based binning of the defects is performed using the processor.

[0012] The above three-dimensional structure may be a three-dimensional NAND structure.

[0013] The step of generating the mask may include the step of performing auto-correlation of the one-dimensional curve using the processor and determining the period thereby, and the step of performing auto-convolution and arbitration of the period using the processor and determining the trench center thereby. The trench center may be used as a reference. The trench, edge hole, transition hole, and center hole regions in the mask image may be determined using the processor.

[0014] The defects may be detected among the pixels in the area of ​​the mask. The step of detecting the defects may further include the step of extracting a patch around the location of one of the defects. The method may further include the step of determining the distance to the center of an adjacent trench using the processor. The location-based binning may be the distance to the center of the trench.

[0015] The above position-based binning can separate the defects on different rows of channel holes into corresponding bins.

[0016] A non-transient computer-readable medium may store a program configured to instruct a processor to execute the method of the first embodiment.

[0017] A system is provided in a second embodiment. The system comprises: a stage configured to hold a semiconductor wafer; a light source configured to direct a beam of light onto the semiconductor wafer on the stage; a detector configured to receive reflected light from the semiconductor wafer on the stage; and a processor that communicates electronically with the detector. The light source may be a broadband plasma source. The detector is configured to receive an image of the semiconductor wafer; generate a one-dimensional projection of the image and thereby form a one-dimensional curve; generate a mask from the one-dimensional curve of the image; detect defects on the image using the mask; and perform location-based binning of the defects.

[0018] The step of generating the mask may include the step of performing autocorrelation of the one-dimensional curve and determining the period by this, and the step of performing magnetic convolution and adjustment of the period and determining the trench center by this.

[0019] The center of the trench can be used as a reference. In the mask image, trench, edge hole, transition hole, and center hole regions can be determined.

[0020] The defects may be detected among the pixels in the area of ​​the mask. Detecting the defects may further include extracting a patch around the location of one of the defects. The method may further include the step of determining the distance to the center of an adjacent trench. The location-based binning may be the distance to the center of the trench.

[0021] The above position-based binning can separate the defects on different rows of channel holes into corresponding bins. Brief explanation of the drawing

[0022] For a more complete understanding of the nature and purposes of the present disclosure, the following detailed description made in conjunction with the accompanying drawings should be referenced. Figure 1 is an exemplary diagram of a 3D NAND structure. Figure 2A illustrates a frame image. Figure 2B illustrates a horizontal projection corresponding to Figure 2A. Figure 2C illustrates a mask image corresponding to Figure 2A. FIG. 3 is a flowchart of an example of a method according to the present disclosure. Figure 4 illustrates examples of programmed omission defects. FIG. 5 illustrates histograms comparing the method embodiment of the present disclosure with IBS. FIG. 6 is a block diagram of a system according to the present disclosure. Specific details for implementing the invention

[0023] Although the claimed subject matter will be described in relation to specific embodiments, other embodiments, including those that do not provide all of the advantages and features described herein, are also within the scope of this disclosure. Various structural, logical, process steps, and electronic modifications may be made without departing from the scope of this disclosure. Accordingly, the scope of this disclosure is defined solely by reference to the appended claims.

[0024] Broadband plasma (BBP) inspection systems, laser inspection systems, or other optical inspection systems can be used for 3D NAND defect detection. To improve sensitivity, semiconductor manufacturers may need to distinguish defects located on different rows of trenches and channel holes, rather than simply reporting all defects in the same area. Missing holes near trenches typically has a greater impact than particles on the trenches, whereas central holes are dummy holes and are generally less significant than other holes.

[0025] The embodiments disclosed herein use a priori information (e.g., periodicity, symmetry, etc.) to overcome the effects of GL variations and autofocus problems. Adjustment methods may be implemented to modify trench location identification. Cross-frame borrowing logic may be used to allow for global GL variations. Image pixels may be segmented based on their distances to the center of their trenches. Individual sensitivities may be applied to each segmentation for defect detection. In the recipe, the distance from the defect location to the center of the trench may be calculated for defect binning. Compared to IBS, the embodiments disclosed herein are robust to wafer process variations and noise, and may also allow for some autofocus problems. Separation purity may also be better than that of IBS. IBS separates images based on neighboring pixels, and local gray level variations will affect binning performance. The embodiments disclosed herein use pixels from the entire care area, which is more robust to allow for local variations.

[0026] An exemplary 3D NAND structure image having various features is shown in FIG. 1. Bins are exemplified, and bin 0 represents the background. Each row of holes may have a height of 100 nm to 200 nm. 3D NAND images are typically periodic and symmetrical vertically, as shown in the frame image in FIG. 2A, while being nearly uniform horizontally. The image in FIG. 2A may be provided by a BBP inspection system or another inspection system. The horizontal projection of the image in FIG. 2A can be used to generate a periodic and symmetrical curve as shown in FIG. 2B, which results in the mask image in FIG. 2C. Such structures can increase the complexity of defect detection. Using the embodiments disclosed herein, cross-correlation may be helpful in identifying pitch and trench centers. Image pixels may be segmented based on their distances to the center of the trench. Individual sensitivities may be applied to each segmentation for defect detection. For quantitative defect binning, the distance from the defect location to the center of the trench can be calculated.

[0027] FIG. 3 is a flowchart of method (100). Some or all of the steps of method (100) may use a processor.

[0028] An image (101) is used in the method (100). The image (101) is an image of a three-dimensional structure of a semiconductor wafer, such as a 3D NAND structure. An exemplary image, such as that shown in FIG. 1, may be from an electron beam inspection system or another type of inspection system.

[0029] In 102, a one-dimensional projection of the image, as shown in FIG. 2A, is generated, thereby forming a one-dimensional curve as shown in FIG. 2B. The one-dimensional projection accumulates all pixels along one dimension to determine the gray level distribution. The one-dimensional curve represents the gray levels of the one-dimensional projection.

[0030] In one example, this algorithm can take the average of the gray levels of all pixels parallel to the trench in the image to generate a one-dimensional projection. The one-dimensional projection can be converted into a one-dimensional curve by using the average value output. Each row has one average value output. The values ​​from all rows can form a one-dimensional curve.

[0031] In 103, the autocorrelation of the one-dimensional curve is performed to determine the period. An example of the autocorrelation function is given in Equation 6. The autocorrelation of the one-dimensional curve can determine the period length of the unique trench-Hall pitch. The autocorrelation of the one-dimensional curve can also provide candidates for the trench center.

[0032] In one example, autocorrelation determines the pitch from the original profile to the offset profile. The original profile can be a one-dimensional projection profile. The offset profile is profile R x [k] can be. Peak values ​​can be examined, and peak-to-peak values ​​are determined. This may include normalized cross-correlation (NCC).

[0033] In 104, magnetic convolution and adjustment of the period are performed to determine the trench centers. The adjustment method can be used to determine whether the candidates are trench centers or central hole rows. For example, the profile can be flipped during magnetic convolution to find the trench centers. The adjustment can use dark or bright peaks as centers, which can be based on information from semiconductor manufacturers. For example, the user can select a dark peak or a bright peak from the user interface.

[0034] Examples of autocorrelation (103) and autoconvolution (104) are provided. A real discrete signal x[n] is T-periodic if there exists a positive integer T such that for all n ∈ Z, x[n] = x[n + T]. A signal x[n] is M-symmetric if there exists an integer M such that for all n ∈ Z, x[n] = x[M - n]. In this case, M / 2 is one of the centers of symmetry and is not necessarily unique.

[0035] If a signal is both T-periodic and M-symmetric, it is also (M + jT)-symmetric, where j is an arbitrary integer. For a fixed integer j, for all n ∈ Z, x[n] = x[M - n] = x[M + jT - n]. Therefore, x[n] is also (M + jT)-symmetric. This means that a periodic and symmetric signal contains a series of centers of symmetry with half-cycle spacing.

[0036] The autocorrelation function at lag k for a discrete signal x[n] is It can be defined as. The magnetic convolution function at lag k for a discrete signal x[n] is It can be defined as, which can be viewed as the cross-correlation of x[n] and its reversion x[-n]. Therefore, the following mathematical equations apply.

[0037]

[0038]

[0039] If x[n] is M-symmetric, Vx[k] reaches a global maximum at k = M. Using Equation 3,

[0040]

[0041] For any k, V x [k] is V x [M] cannot be exceeded. This is shown in Equation 4.

[0042]

[0043] Therefore, the center of symmetry of the signal can be determined from the peaks in its self-convolution function.

[0044] When x[n] is both T-periodic and M-symmetric, the self-convolution function Vx[k] achieves a global maximum at k = M + jT for all j ∈ Z. This can be used for pitch detection on 3D structures.

[0045] Once the period and trench centers are identified, a mask image is generated from the one-dimensional curve of the image in 106, as in Fig. 2C. The mask image may be based on predefined widths of the trench region and each hole region. Thus, the trench center can be used as a reference. As shown in Fig. 1, trench, edge hole, transition hole, and central hole regions can be determined in the mask image. The transition holes and central holes are outlined by dashed lines in Fig. 1. Edge holes are the outer row of the hole group closest to the trench, central holes are in the center of the hole group, and transition holes are between the edge holes and the central holes. In one example, after the trench center is determined, within the period, the trench center is segmented into subregions based on sensitivity thresholds to generate a mask image.

[0046] In one example, a semiconductor manufacturer may specify sensitivities in different regions of a mask based on a semiconductor structure or other design. Thus, the mask may represent regions of pixels with specific sensitivities. The semiconductor manufacturer may set the sensitivity in the BBP inspection system for each subregion so that each subregion can use an independent threshold value.

[0047] Pixels of the image from 101 can be segmented based on a mask from 106 for the detection of individual segmentation.

[0048] In 107, defects on the image are detected using a mask. For example, defects can be detected among the pixels within the mask area with a desired sensitivity. Detecting defects may involve extracting a patch around the location of one of the defects. For example, the patch size is 32x32 pixels for a BBP inspection system. The distance between the defect and the adjacent trench center can be determined. After the trench center is calculated, the defect peak location can also be found for each defect. This distance is the difference between these two values. This distance can be used for defect binning. For example, a local maximal in the patch difference image can represent a defect.

[0049] Defect locations can be calculated in 108. Based on the distribution, location-based binning can be performed in 109. The results of the location-based binning can be converted into a histogram, as shown in Fig. 3. In one example, the distribution of defects in relation to the distance to the center of the trench is the final binning result. This can be used to determine the locations of defects of interest.

[0050] This behavior is further explained in the example below. Let x[n] be a real-valued discrete signal. All If there exists a positive integer T such that x[n] = x[n + T] for , this is T-periodic, and all If there exists an integer M such that x[n] = x[M - n] for , it is M-symmetric. In this case, M / 2 is one of the centers of symmetry (though not necessarily unique). It can be verified whether a signal is both T-periodic and M-symmetric. This also applies to all It is (M + jT)-symmetric with respect to . As a result, a periodic and symmetric signal contains a series of symmetry centers with half-period spacing.

[0051] In the case of autocorrelation and autoconvolution, cross-correlation functions can be used to detect periodicity and symmetry. Note that a periodic and symmetric signal will overlay itself at multiples of a period. Therefore, the cross-correlation function of the signal and itself will reach maximum values ​​at a series of points at one-period intervals. This type of cross-correlation function is referred to as the autocorrelation function. Furthermore, this signal will also overlay its inversion at multiples of a period. Thus, the cross-correlation function of the signal and its inversion will contain a series of peaks, each corresponding to the center of symmetry of the original signal. This type of cross-correlation function is defined as the autoconvolution function. In summary, periodicity and symmetry can be determined, respectively, from the autocorrelation function and the autoconvolution function.

[0052] Projection data Let us assume that it is given by. The mean and variance can be defined as follows in Equation 5.

[0053]

[0054] Estimates of autocorrelation and magnetic convolution at a positive lag k can be obtained using Equation 6.

[0055]

[0056] The expression for the negative lag k can be defined in the same way. Using Equation 7, the autocorrelation function R x [k] and self-convolution function V x It may be convenient to normalize [k].

[0057]

[0058] The terms autocorrelation and autoconvolution can represent normalized versions.

[0059] Periodicity and symmetry can be inferred from the autocorrelation function and the autoconvolution function. All peaks higher than a given threshold can be initially identified. Subsequently, the pitch can be identified as the average interval of the peaks.

[0060] Once the pitch T is obtained, the trench center can be further determined. The peaks of the self-convolution function Let's assume it occurs at. Due to the presence of wafer noise, some of the expected peaks may be missed. Some unknown About An M that makes this possible can be found. Therefore, The optimization problem related to is as follows.

[0061]

[0062] Without loss of generality, M can be restricted to the range [T / 2, T / 2). The optimal solution M* to the problem in Equation 8 is the following set of candidates It can belong to, and here and is the remainder m related to T l As a result, the value of the objective function for the candidate set can be compared and the minimizer M* can be obtained.

[0063] As seen in 3D NAND images, half-pitch ambiguity may exist at the center of the trench, i.e., at the bright or dark center. Adjustment may be required to determine the correct trench center. It can be difficult to determine trench polarity based solely on projection data. There are at least three methods for adjustment. The first method is to use brightness from observations. The second method is to use horizontal variance, as hole regions typically have more wafer noise. The third method is to pre-define a template to match the wafer pattern. Since there may be layer-to-layer, wafer-to-wafer, and die-to-die variations, an empirical selection of adjustment methods can be applied to each specific layer.

[0064] The acquired trench center can be used as a reference. The trench, edge hole, transition hole, and center hole regions can be filled one by one in the mask image with corresponding ratios. Using the mask image, segmented MDAT (multi-die automatic thresholding) detection is performed to detect defects for each segmentation.

[0065] When there are global GL variations, some frames may experience difficulties due to incorrect pitch and trench center values. Borrow logic can be implemented to obtain pitch and trench center values ​​from other frames. Therefore, values ​​for period and trench center can be borrowed from neighboring frames. Inter-frame offsets may be taken into account during borrowing.

[0066] In the experiment, the method (100) was tested on a wafer using a BBP inspection system. The performance was compared with that of an IBS. As shown in FIG. 4, there are programmed hole-missing defects on the wafer. Each dot in FIG. 4 is a channel hole. The missing defects, depicted as hollow circles in FIG. 4, are located in rows 1 through 5. The objective is to distinguish defects from distinct rows.

[0067] In prototype development, the distance between a defect and the center of a trench is calculated as an attribute of the defects for binning. The histograms in FIG. 5 show that the embodiments disclosed herein ("new binning algorithm") can find a clear cutline to separate two different types of defects, whereas using IBS results in a large overlap between the distributions of the two types of defects. Due to the clear cutline on the histogram, the overall binning accuracy is improved using the embodiments disclosed herein.

[0068] Using a priori information of the wafer layout (e.g., periodicity, symmetry, etc.) can overcome the effects of GL variations and autofocus problems. The embodiments disclosed herein can accurately separate defects on different rows of channel holes into corresponding bins and can help semiconductor manufacturers achieve enhanced sensitivity tuning and precise defect monitoring for better control of wafer yield.

[0069] One embodiment of the system (200) is illustrated in FIG. 6. The system (200) includes an optical-based subsystem (201). Generally, the optical-based subsystem (201) is configured to generate an optical-based output for the sample (802) by directing light toward the sample (202) (or scanning light across the sample (202)) and detecting light from the sample (802). In one embodiment, the sample (202) includes a wafer. The wafer may include any wafer known in the art. In another embodiment, the sample (202) includes a reticle. The reticle may include any reticle known in the art.

[0070] In an embodiment of the system (200) illustrated in FIG. 6, the optical-based subsystem (201) includes an illumination subsystem configured to direct light toward a sample (202). The illumination subsystem includes at least one light source. For example, as illustrated in FIG. 6, the illumination subsystem includes a light source (203). In one embodiment, the illumination subsystem is configured to direct light toward the sample (202) at one or more angles of incidence, which may include one or more oblique angles and / or one or more normal angles. For example, as illustrated in FIG. 6, light from the light source (203) is directed toward the sample (202) through an optical element (204) and then a lens (205) at an oblique angle of incidence. The oblique angle of incidence may include any suitable oblique angle of incidence, which may vary, for example, depending on the characteristics of the sample (202).

[0071] The optical-based subsystem (201) may be configured to direct light toward the sample (202) at different angles of incidence at different times. For example, the optical-based subsystem (201) may be configured to change one or more characteristics of one or more elements of the illumination subsystem so that the light may be directed toward the sample (202) at an angle of incidence different from that shown in FIG. 6. In one such example, the optical-based subsystem (201) may be configured to move the light source (203), the optical element (204), and the lens (205) so that the light may be directed toward the sample (202) at a different oblique angle of incidence or a vertical (or nearly vertical) angle of incidence.

[0072] In some examples, the optical-based subsystem (201) may be configured to direct light toward the sample (202) at more than one angle of incidence simultaneously. For example, the illumination subsystem may include more than one illumination channel, and one of the illumination channels may include a light source (203), an optical element (204), and a lens (205) as shown in FIG. 6, and another of the illumination channels (not shown) may include similar elements that may be configured differently or identically, or at least one or more other components such as a light source and perhaps further described herein. When such light is directed toward the sample simultaneously with other light, one or more characteristics (e.g., wavelength, polarization, etc.) of the light directed toward the sample (202) at different angles of incidence may be different so that the light resulting from illuminating the sample (202) at different angles of incidence can be distinguished from one another at the detector(s).

[0073] In another example, the illumination subsystem may include only one light source (e.g., the light source (203) shown in FIG. 6), and light from the light source may be separated into different optical paths (e.g., based on wavelength, polarization, etc.) by one or more optical elements (not shown) of the illumination subsystem. Light in each of the different optical paths may then be directed toward the sample (202). Multiple illumination channels may be configured to direct light toward the sample (202) simultaneously or at different times (e.g., when different illumination channels are used to sequentially illuminate the sample). In another example, the same illumination channel may be configured to direct light having different characteristics toward the sample (202) at different times. For example, in some examples, the optical element (204) may be configured as a spectral filter, and the properties of the spectral filter may be changed in various different ways (e.g., by swapping out the spectral filter) so that light of different wavelengths may be directed toward the sample (202) at different times. The illumination subsystem may have any other suitable configuration known in the art for directing light having different or the same characteristics toward the sample (202) sequentially or simultaneously at different or the same angles of incidence.

[0074] In one embodiment, the light source (203) may include a BBP source. In this way, the light generated by the light source (203) and directed toward the sample (202) may include broadband light. However, the light source may include any other suitable light source, such as a laser. The laser may include any suitable laser known in the art and may be configured to generate light of any suitable wavelength or wavelengths known in the art. Additionally, the laser may be configured to generate light that is monochromatic or nearly monochromatic. In this way, the laser may be a narrowband laser. The light source (203) may also include a polychromatic light source that generates light of a number of discrete wavelengths or wavebands.

[0075] Light from the optical element (204) can be focused onto the sample (202) by the lens (205). Although the lens (205) is illustrated in FIG. 6 as a single refractive optical element, it will be understood that in practice, the lens (205) may include a plurality of refractive and / or reflective optical elements and that these may be combined to focus light from the optical element onto the sample. The illumination subsystem illustrated in FIG. 6 and described herein may include any other suitable optical elements (not illustrated). Examples of such optical elements include, but are not limited to, polarizing components(s), spectral filters(s), spatial filters(s), reflective optical elements(s), apodizer(s), beam splitters(s) (such as beam splitter (213)), aperture(s), etc., which may include any such suitable optical elements known in the art. Additionally, the optical-based subsystem (201) may be configured to change one or more of the elements of the lighting subsystem based on the type of lighting to be used to generate the optical-based output.

[0076] The optical-based subsystem (201) may also include a scanning subsystem configured to allow light to be scanned across the sample (202). For example, the optical-based subsystem (201) may include a stage (206) on which the sample (202) is placed during the generation of an optical-based output. The scanning subsystem may include any suitable mechanical and / or robotic assembly (including the stage (206)) which may be configured to move the sample (202) so that light can be scanned across the sample (202). Additionally or alternatively, the optical-based subsystem (201) may be configured such that one or more optical elements of the optical-based subsystem (201) perform partial scanning of the light across the sample (202). The light may be scanned across the sample (202) in any suitable manner, for example, in a serpentine-like path or in a spiral path.

[0077] The optical-based subsystem (201) further includes one or more detection channels. At least one of the one or more detection channels includes a detector configured to detect light from the sample (202) caused by illumination of the sample (202) by the subsystem and to generate an output in response to the detected light. For example, the optical-based subsystem (201) illustrated in FIG. 6 includes two detection channels, one formed by a collector (207), an element (208), and a detector (209), and the other formed by a collector (210), an element (211), and a detector (212). As illustrated in FIG. 6, the two detection channels are configured to collect and detect light at different collection angles. In some examples, both detection channels are configured to detect scattered light, and the detection channels are configured to detect light scattered from the sample (202) at different angles. However, one or more of the detection channels may be configured to detect other types of light (e.g., reflected light) from the sample (202).

[0078] As further illustrated in FIG. 6, both detection channels are depicted as being located within the plane of the paper, and the illumination subsystem is also depicted as being located within the plane of the paper. Thus, in this embodiment, both detection channels are located within the plane of incidence (e.g., centered on the plane of incidence). However, one or more of the detection channels may be located outside the plane of incidence. For example, a detection channel formed by a collector (210), an element (211), and a detector (212) may be configured to collect and detect light scattered outside the plane of incidence. Thus, such a detection channel may be typically referred to as a "side" channel, and such a side channel may be centered on a plane substantially perpendicular to the plane of incidence.

[0079] Although FIG. 6 illustrates an embodiment of an optical-based subsystem (201) comprising two detection channels, the optical-based subsystem (201) may comprise a different number of detection channels (e.g., only one detection channel or two or more detection channels). In one such example, the detection channel formed by the collector (210), the element (211), and the detector (212) may form one side channel as described above, and the optical-based subsystem (201) may include an additional detection channel (not shown) formed as another side channel located on the opposite side of the incident plane. Thus, the optical-based subsystem (201) may include a detection channel comprising a collector (207), an element (208), and a detector (209) and configured to collect and detect light at scattering angle(s) centered on the incident plane and perpendicular or nearly perpendicular to the surface of the sample (202). Accordingly, this detection channel may be typically referred to as a "top" channel, and the optical-based subsystem (201) may also include two or more side channels configured as described above. Accordingly, the optical-based subsystem (201) may include at least three channels (i.e., one top channel and two side channels), each of the at least three channels having its own collector, and each of these collectors is configured to collect light at scattering angles different from each of the other collectors.

[0080] As further explained above, each of the detection channels included in the optical-based subsystem (201) may be configured to detect scattered light. Thus, the optical-based subsystem (201) illustrated in FIG. 6 may be configured to generate a dark field (DF) output for the sample (202). However, the optical-based subsystem (201) may also or alternatively include detection channel(s) configured to generate a bright field (BF) output for the sample (202). In other words, the optical-based subsystem (201) may include at least one detection channel configured to detect specularly reflected light from the sample (202). Thus, the optical-based subsystems (201) described herein may be configured for DF imaging only, BF imaging only, or both DF imaging and BF imaging. Although each of the collectors is illustrated as a single refractive optical element in FIG. 6, it should be understood that each of the collectors may include one or more refractive optical die(s) and / or one or more reflective optical element(s).

[0081] One or more detection channels may include any suitable detectors known in the art. For example, detectors may include photo-multiplier tubes (PMTs), charge-coupled devices (CCDs), time-delay integration (TDI) cameras, and any other suitable detectors known in the art. Detectors may also include non-imaging detectors or imaging detectors. In this way, if the detectors are non-imaging detectors, each detector may be configured to detect specific characteristics of scattered light, such as intensity, but may not be configured to detect such characteristics as a function of position within the imaging plane. Accordingly, the output generated by each detector included in each of the detection channels of the optical-based subsystem may be signals or data, but may not be image signals or image data. In such examples, a processor such as the processor (214) may be configured to generate images of the sample (202) from the non-imaging outputs of the detectors. However, in other cases, the detectors may be configured as imaging detectors configured to generate imaging signals or image data. Thus, the optical-based subsystem may be configured to generate optical images or other optical-based outputs described herein in a number of ways.

[0082] Note that FIG. 6 is provided herein to generally illustrate the configuration of an optical-based subsystem (201) capable of generating an optical-based output that may be included in or used by the system embodiments described herein. The configuration of the optical-based subsystem (201) described herein may be modified to optimize the performance of the optical-based subsystem (201) as is typically done when designing commercial output acquisition systems. Additionally, the systems described herein may be implemented using existing systems (e.g., by adding the functions described herein to existing systems). For some such systems, the methods described herein may be provided as optional functions of the system (e.g., in addition to other functions of the system). Alternatively, the systems described herein may be designed as entirely new systems.

[0083] To enable the processor (214) to receive output, the processor (214) may be coupled to components of the system (200) in any suitable manner (e.g., via one or more transmission media, which may include wired and / or wireless transmission media). The processor (214) may be configured to perform a number of functions using the output. The system (200) may receive instructions or other information from the processor (214). The processor (214) and / or the electronic data storage unit (215) may optionally communicate electronically with a wafer inspection system, a wafer metrology system, or a wafer review system (not exemplified) to receive additional information or transmit instructions. For example, the processor (214) and / or the electronic data storage unit (215) may communicate electronically with a scanning electron microscope.

[0084] The processor (214), other system(s), or other subsystem(s) described herein may be part of various systems, including personal computer systems, image computers, mainframe computer systems, workstations, network appliances, internet appliances, or other devices. The subsystem(s) or system(s) may also include any suitable processor known in the art, such as a parallel processor. Additionally, the subsystem(s) or system(s) may include a platform having high-speed processing and software as a standalone tool or a networked tool.

[0085] The processor (214) and the electronic data storage unit (215) may be placed within the system (200) or another device, or otherwise be part of it. In one example, the processor (214) and the electronic data storage unit (215) may be part of a standalone control unit or within a centralized quality control unit. Multiple processors (214) or electronic data storage units (215) may be used.

[0086] The processor (214) may actually be implemented by any combination of hardware, software, and firmware. Additionally, its functions as described herein may be performed by a single unit or divided among different components, each of which may be implemented by any combination of hardware, software, and firmware. Program code or instructions for the processor (214) to implement various methods and functions may be stored in readable storage media, such as memory within an electronic data storage unit (215) or other memory.

[0087] If the system (200) includes more than one processor (214), different subsystems may be combined with each other so that images, data, information, instructions, etc. can be transmitted between the subsystems. For example, one subsystem may be combined to additional subsystem(s) by any suitable transmission medium, which may include any suitable wired and / or wireless transmission medium known in the art. Two or more of such subsystems may also be substantially combined by a shared computer-readable storage medium (not shown).

[0088] The processor (214) may be configured to perform a number of functions using the output of the system (200) or other outputs. For example, the processor (214) may be configured to transmit the output to an electronic data storage unit (215) or other storage medium. The processor (214) may be configured according to any of the embodiments described herein. The processor (214) may also be configured to perform other functions or additional steps using the output of the system (200) or using images or data from other sources.

[0089] Various steps, functions, and / or operations of the system (200) and methods disclosed herein are performed by one or more of electronic circuits, logic gates, multiplexers, programmable logic devices, ASICs, analog or digital controls / switches, microcontrollers, or computing systems. Program instructions for implementing methods such as those described herein may be transmitted through a carrier medium or stored on a carrier medium. The carrier medium may include a storage medium such as read-only memory, random access memory, magnetic or optical disk, non-volatile memory, solid-state memory, magnetic tape, etc. For example, various steps described throughout this disclosure may be performed by a single processor (214) or, alternatively, by a plurality of processors (214). Furthermore, different subsystems of the system (200) may include one or more computing or logic systems. Accordingly, the above description should not be interpreted as a limitation to this disclosure but merely as an example.

[0090] In one example, the processor (214) communicates with the system (200). Location-based binning using the processor (214) can separate defects on different rows of channel holes into corresponding bins. The processor (214) is configured to receive an image of a semiconductor wafer; generate a one-dimensional projection of the image and thereby form a one-dimensional curve; generate a mask from the one-dimensional curve of the image; detect defects on the image with the mask; and perform location-based binning of the defects.

[0091] The step of generating a mask may include the step of performing autocorrelation of a one-dimensional curve and determining the period by this, and the step of performing magnetic convolution and adjustment of the period and determining the trench center by this. The trench center may be used as a reference. In the mask image, trench, edge hole, transition hole, and central hole regions may be determined.

[0092] Defects can be detected among the pixels in the area of ​​the mask. Detecting defects may further involve extracting a patch around the location of one of the defects. For each defect, the distance to the center of a neighboring trench can be determined. For example, this may be the distance from the edge of the defect, the patch, or the center of the patch to a neighboring trench. Location-based binning may be the distance to the center of the trench.

[0093] Additional embodiments relate to a non-transient computer-readable medium storing program instructions executable on a controller for performing a computer-implemented method for fault detection as disclosed herein. In detail, as illustrated in FIG. 6, an electronic data storage unit (215) or other storage medium may comprise a non-transient computer-readable medium containing program instructions executable on a processor (214). The computer-implemented method may include any step(s) of any method(s) described herein, including method (100).

[0094] Program instructions may be implemented in any of the various ways, including, among other things, procedure-based technologies, component-based technologies, and / or object-oriented technologies. For example, program instructions may be implemented using ActiveX controls, C++ objects, JavaBeans, Microsoft Foundation Classes (MFC), Streaming SIMD Extension (SSE), or other technologies or methodologies, as desired.

[0095] Each of the steps of the method may be performed as described herein. The methods may also include any other step(s) that may be performed by the processor and / or computer subsystem(s) or system(s) described herein. The steps may be performed by one or more computer systems that may be configured according to any of the embodiments described herein. Additionally, the methods described above may be performed by any of the system embodiments described herein.

[0096] Although the present disclosure has been described in relation to one or more specific embodiments, it will be understood that other embodiments of the present disclosure may be made without departing from the scope of the present disclosure. Accordingly, the present disclosure is to be considered limited only by the appended claims and a reasonable interpretation thereof.

Claims

Claim 1 A method comprising the steps of: receiving an image from a processor—the image being an image of a three-dimensional structure of a semiconductor wafer—; generating a one-dimensional projection of the image using the processor and thereby forming a one-dimensional curve; generating a mask from the one-dimensional curve of the image using the processor; detecting defects on the image using the mask using the processor; and performing location-based binning of the defects using the processor. Claim 2 A method according to claim 1, wherein the image is generated by a broadband plasma inspection system. Claim 3 A method according to claim 1, wherein the three-dimensional structure is a three-dimensional NAND structure. Claim 4 A method according to claim 1, wherein the step of generating the mask comprises: a step of performing auto-correlation of the one-dimensional curve using the processor and determining the period thereby; and a step of performing auto-convolution and arbitration of the period using the processor and determining the trench center thereby. Claim 5 A method according to claim 4, wherein the center of the trench is used as a reference, and trench, edge hole, transition hole, and center hole regions are determined in the mask image using the processor. Claim 6 A method according to claim 1, wherein the defects are detected among the pixels in the area of ​​the mask. Claim 7 A method according to claim 1, wherein the step of detecting the defects further comprises the step of extracting a patch around the location of one of the defects. Claim 8 In claim 7, the method further comprises the step of determining the distance from each of the defects to the center of an adjacent trench using the processor. Claim 9 A method according to claim 8, wherein the position-based binning is the distance-to-trench-center. Claim 10 A method according to claim 1, wherein the position-based binning separates the defects on different rows of channel holes into corresponding bins. Claim 11 A non-transient computer-readable medium storing a program configured to instruct a processor to execute the method of claim 1. Claim 12 A system comprising: a stage configured to hold a semiconductor wafer; a light source configured to direct a beam of light onto the semiconductor wafer on the stage; a detector configured to receive reflected light from the semiconductor wafer on the stage; and a processor electronically communicating with the detector, wherein the processor is configured to receive an image of the semiconductor wafer; generate a one-dimensional projection of the image and thereby form a one-dimensional curve; generate a mask from the one-dimensional curve of the image; detect defects on the image using the mask; and perform location-based binning of the defects. Claim 13 A system according to Clause 12, wherein the light source is a broadband plasma source. Claim 14 A system according to claim 12, wherein generating the mask comprises: performing autocorrelation of the one-dimensional curve and determining the period by the same; and performing magnetic convolution and adjustment of the period and determining the trench center by the same. Claim 15 A system according to claim 14, wherein the center of the trench is used as a reference, and trench, edge hole, transition hole, and central hole regions are determined in the mask image. Claim 16 A system in which the defects are detected among the pixels in the area of ​​the mask in paragraph 12. Claim 17 A system according to claim 12, wherein detecting the defects further comprises extracting a patch around the location of one of the defects. Claim 18 In paragraph 17, the system is configured such that the processor is also configured to determine the distance from each of the faults to the center of an adjacent trench. Claim 19 In paragraph 18, the system wherein the above location-based binning is the distance to the center of the trench. Claim 20 In paragraph 12, the position-based binning is a system that separates the defects on different rows of channel holes into corresponding bins.

Images