TWI574136B - Method of design-based defect classification and system thereof - Google Patents

Description

Method and system for designing defect based on design [Reciprocal Reference of Related Applications]

The present application claims the benefit of priority to U.S. Provisional Patent Application Serial No. 61/594, filed on Jan.

The present invention relates to wafer inspection tools and methods of operating such wafer inspection tools and, in particular, the present invention relates to methods and systems for classifying defects based on design.

In the semiconductor industry, devices are fabricated by a number of processes that produce structures of ever-decreasing dimensions. Thus, such processes as inspection processes, measurement processes, and the like (hereinafter referred to as inspection processes) require increased precision and effectiveness for making samples. The term "sample" as used in this specification shall be interpreted broadly to cover any type of wafer, main reticle and other structures, such wafers used in the manufacture of semiconductor integrated circuits, magnetic heads, flat panel displays, and other thin film devices, A combination and/or portion of the main reticle and other structures.

The inspection process may include parameters for evaluating individual processes and/or Identification and measurement, calibration, monitoring, inspection, reporting, and/or other procedures necessary to provide the necessary structural feedback. Various inspection tools may be based on scanning electron microscopy, atomic force microscopy, optical inspection tools, and the like, such as non-destructive observations as non-limiting examples. The inspection process is important for debugging sample processes, monitoring process changes, and improving product yield.

As design rules continue to shrink (under 28 nm and below 28 nm), the amount of defect-related data reported by high-sensitivity inspection tools is very large (eg, thousands of defects per wafer). In addition, new types of processes (such as immersion lithography, photoresist shrinkage, photoresist trimming, etc.) are introduced to introduce new types of errors due to different proximity effects (optical, chemical mechanical polishing (CMP)). , chemistry, 3D, etc.) are generated and reported as defects by inspection tools. The severity of the reported defects can vary from a catastrophic impact on product yield to an unimportant anomaly that does not affect product quality.

Therefore, there is a need to classify reported defects and separate the defects of interest (DOI) from the defects considered to be nuisances. As manufacturing control requirements become more challenging, the classification of reported defects has become very complex and time consuming and processing power.

The problem of classifying defects during the process has been discovered in the prior art, and various techniques have been developed to provide solutions.

A typical method is to analyze predetermined attributes of defects (eg, size, magnitude, orientation, shape, etc.) and perform classification based on the attributes. Other classification techniques also consider the location of reported defects in the sample (eg, for certain defined regions).

Defects can be based on one or more attributes of the defect, and one or more One or more attribute classifications of the patterned features formed on the sample proximate to the defect. In this manner, not only can one or more attributes be based on defects, but defects can also be classified based on one or more attributes of any patterned features that are close to the defect on the sample.

Defects can be further classified using various methods for combining inspection data with design data.

In accordance with certain aspects of the subject matter disclosed herein, a computer implemented method for classifying defects detected on a production layer of a sample is provided. The method comprises the steps of: using a computer to obtain input data related to the detected defect; using a decision algorithm associated with the production layer and specifying two or more classification operations and the two or more classification operations Serializing the input data; sorting the processed defects according to a predetermined grouping, wherein each group is associated with at least one sorting operation, wherein at least one sorting operation sorts at least a portion of the processed defects to one or more of the classified groups to generate a defect of the final classification, and wherein each of the classification operations except the last operation sorts at least a portion of the processed defects to be processed by one or more of the subsequent classification operations; and stores at least the final classification in the storage medium Defects.

In accordance with other aspects of the subject matter disclosed herein, an inspection system is provided that is capable of classifying defects detected on a production layer of a sample. The inspection system includes a classification unit operatively coupled to the at least one inspection tool and operatively coupled to the design data storage unit, wherein the classification unit includes a defect data interface, the defect data interface being configured to at least one inspection The tool obtains the defect indication data; the design data interface, the design data interface Provided to obtain design instruction data from the design data storage unit; the memory configured to store defects of the final classification; and a processor configured to use the decision algorithm associated with the production layer and specify The two or more classification operations and the sequence of the two or more classification operations process the defect indication data and the designation instructions; the processor is further configured to sort the processed defects according to the predetermined grouping. Each packet is associated with at least one classification operation; wherein at least one classification operation sorts at least a portion of the processed defects to one or more classification packets to produce a defect of the final classification; and wherein each of the last classification operations The sorting operation sorts at least a portion of the processed defects to be processed by one or more of the subsequent sorting operations.

In accordance with other aspects of the subject matter disclosed herein, a classification unit is provided that operatively incorporates an inspection system that is capable of classifying defects detected on a production layer of a sample. The classification unit includes a processor operatively coupled to the memory accessible by the processor and the memory is configured to store a defect of the final classification, wherein the processor is configured to use the associated with the production layer The decision algorithm and specifies two or more classification operations and sequences of the two or more classification operations to process the defect indication data and the designation instructions; the processor is further configured to order the processed defects according to the predetermined grouping. Each packet is associated with at least one classification operation; wherein at least one classification operation sorts at least a portion of the processed defects to one or more classification packets to produce a defect of the final classification; and wherein each of the last classification operations The sorting operation sorts at least a portion of the processed defects to be processed by one or more of the subsequent sorting operations.

Further aspects of the subject matter disclosed herein and, as appropriate, In conjunction with any of the above aspects, the defects ordered by the current classification operation to be processed by one or more of the subsequent classification operations are ordered to one or more packets, the one or more packets A group selection associated with the current classification operation and from the following: a grouping of defects that are not classified by the current classification operation and that are removed from further processing by a classification operation defined by the decision algorithm; a packet identified by a current operation identified to a predefined further classification operation and thus a defect to be processed when performing the predetermined classification operation; corresponding to the current classification operation assigned by one of the subsequent classification operations and thus A group of defects to be processed by each next classification operation until further classification. The same defect can be sorted into more than one packet. At least one packet can be associated with at least two classification operations.

According to a further aspect of the subject matter disclosed herein and in combination with any of the above aspects, at least one sorting operation may sort the processed defects based on the matching for each defect, the matching being indicative of all polygons in the vicinity of the defect The data set of the data type and the layer attribute has a match to one of the classification related instructions or a plurality of predefined data sets associated with the combination of the data type and the layer attribute. At least one sorting operation may use a design rule check (DRC) analysis to sort the processed defects by a combination of classifications of design attributes.

According to a further aspect of the subject matter disclosed herein and in combination with any of the above aspects, the order of the sorting operations may be predetermined according to at least one of the following operations: - reduced to be specified by each of the next sorting operations Number of defects; - increase the amount of input data required for each next classification operation; - providing a classification operation that requires a large number of defects to be processed for the classification decision before the classification operation requiring less defects to be processed for the classification decision; and - providing the need after the classification operation requiring less defects for the classification decision to be processed A classification operation for a large number of defects to be processed for classification decision-making, and the classification operation is set to be sorted by processing the appropriate defects for processing aggregation by a classification operation that requires a large number of defects to be processed for the classification decision.

In accordance with a further aspect of the subject matter disclosed herein and in conjunction with any of the above aspects, the design attributes of each of the defects to be processed may be calculated during the inspection scan. Any of the CDA, DRC, and library matching calculations required for the classification operation can be provided during the inspection scan. One or more classification operations including design-based binning may be provided after the inspection scan.

1‧‧‧DRC

2‧‧‧DRC

3‧‧‧DRC

110‧‧‧ wafer

120‧‧‧Design database storage unit

121‧‧‧ scrapbook server

130‧‧‧Manufacture process

140‧‧‧Checking process

150‧‧‧Classification process

210‧‧‧Checking tools

220‧‧‧Classification unit

221‧‧‧Engineering interface

222‧‧‧Design data interface

223‧‧‧Defect data interface

224‧‧‧Processor and memory unit

230‧‧‧ Engineering Station

301‧‧‧Steps

302‧‧‧Steps

303‧‧ steps

304‧‧‧Steps

305‧‧‧Steps

306‧‧‧Steps

401‧‧‧Classification operation

402‧‧‧Classification operation

403‧‧‧ operation

404‧‧‧ operation

405‧‧‧ operation

406‧‧‧Classification operation

410‧‧‧Classification

411‧‧‧Classification

412‧‧‧Classification

414‧‧‧Classification

Group 415‧‧

416‧‧‧Classification

Group 417‧‧

418‧‧‧Classification

Group 419‧‧

Group 420‧‧

Group 421‧‧

600‧‧‧Decision Tree Template

601‧‧‧ cluster classification operation

602‧‧‧Next classification operation

603‧‧‧Library Matching Operations

604‧‧‧ operation

605‧‧‧ operation

606‧‧‧Library Matching Operations/Operations

608‧‧‧ Results

609‧‧‧Table

701‧‧‧Steps

702‧‧‧Steps

703‧‧‧Steps

704‧‧‧Steps

705‧‧‧Steps

In order to understand the present invention and to be able to actually practice the present invention, the embodiments will now be described by way of non-limiting example only in the accompanying drawings in which: FIG. Exemplary Workflow for Sample Manufacturing; FIG. 2 illustrates a schematic functional diagram of an inspection system in accordance with certain embodiments of the disclosed subject matter; FIG. 3 illustrates certain embodiments of the subject matter disclosed in accordance with the present invention. A generalized flowchart of a computerized classification of defects; FIG. 4 illustrates a non-limiting schematic example of a decision tree template generated by computerized classification of defects according to certain embodiments of the disclosed subject matter; FIG . A non-limiting example of an operational sequence table generated by analyzing a decision tree defined by the template in FIG. 4 ; FIG. 6 illustrates a computerized classification of defects in accordance with certain embodiments of the subject matter disclosed herein. generalized flow diagram of the computerized decision algorithm generates a first embodiment and FIG. 7 illustrates in accordance with certain embodiments disclosed in the subject of the present invention; screenshots another non-limiting illustrative examples of decision tree template.

In the following detailed description, numerous specific details are set forth However, it should be understood by those skilled in the art that the invention may be practiced without the specific details. In other instances, well-known methods, procedures, and components are not described in detail to avoid obscuring the invention.

In the drawings and the description, the same element symbols indicate the elements that are common to different embodiments or arrangements.

It will be apparent from the following discussion that unless explicitly stated otherwise, it should be understood that throughout the specification, such as "processing", "calculus", "calculation", "classification", "sorting", "matching", "comparison" The terminology and the like includes the actions and/or processing of a computer that processes and/or processes operational data and/or converts the data into other data that is expressed as physical quantities, such as, for example, electrical quantities, and/or the data. Represents a physical object. The terms "computer," "processor," and "controller" are used broadly to cover any type of electronic device having data processing capabilities, including, by way of non-limiting example, inspection systems that are present in the present disclosure.

The operations in accordance with the teachings herein may be performed by a computer program stored in a non-transitory computer readable storage medium, by a computer specifically constructed for the desired purpose, or by a general purpose computer configured for a particular purpose.

Embodiments of the subject matter disclosed herein are described without reference to any particular programming language. It will be appreciated that the teachings of the subject matter disclosed herein can be implemented in various programming languages.

It is to be understood that in the <Desc/Clms Page number> Conversely, various features of the subject matter disclosed in the &lt;RTI ID=0.0&gt;&gt;

Referring to Figure 1 , there is illustrated an exemplary workflow for sample fabrication in accordance with certain embodiments of the subject matter disclosed herein. For the purposes of illustration only, the following description is provided with respect to inspection of semiconductor wafers. The examples are applicable to other samples and/or other operations that provide information related to defects.

As shown, the wafer 110 can be produced via the fabrication process 130 based on the design data contained in the design repository storage unit 120 . The design repository storage unit 120 can be implemented as a separate, local or remote storage unit, or the design repository storage unit 120 can be integrated with other tools used during the process. By way of non-limiting example, the design data may include a computer automated design (CAD) library. Additionally or alternatively, the design data may be derived from the data stored in the CAD library and/or the design data may be different from the format of the data stored in the CAD library.

During the process, wafer 110 may undergo one or more production processes to produce individual production layers on the wafer. Inspection process 140 may include an inspection process, a defect re-inspection process, or other measurement process for each or some of the production layers of the wafer. The inspection process can be operated simultaneously with the process. The inspection process or part of the inspection process may be provided after the process, as appropriate.

Inspection process 140 may be provided by one or more suitable wafer inspection tools. The inspection process 140 can identify the location of defects in the wafer 110 , and the inspection process 140 can further process the defect data in an attempt to capture information that can be used to gain insight into the design process. By way of non-limiting example, wafer 110 can be inspected using any combination of suitable inspection methods (eg, wafer to wafer, cell to cell, wafer to design, etc.).

Defects identified by inspection process 140 may be classified by classification process 150 . As further detailed with reference to Figures 2 through 7 , the classification process can be provided by the classification unit prior to or in conjunction with the defect retest. The inspection and/or classification process (and/or part of the inspection and/or classification process) may be provided according to individual recipes.

The term "recipe" as used in this specification shall broadly describe any parameter setting that specifies the operation of one or more individual tools (eg, the region of interest to be inspected, the location of the region on the wafer, and the repetition) Cycle, pixel size, beam current, charging and image capture, defect detection algorithms, image processing parameters, and/or other parameter settings, etc.).

The term "design material" as used in this specification should be broadly stated to cover any material that indicates the hierarchical entity design (layout) of the sample and/or information from the physical design (eg, via complex simulations, geometry, and Bolling operations, etc.) ). Design materials can be provided in different formats, which are unrestricted Examples are the GDSII format, the OASIS format, and the like. Design data specifies the structural elements of a design. Structural elements can be constructed as geometric shapes that are combined with the insertion of other structural elements as appropriate. By way of non-limiting example, a given structural element may include one or more STRUCTURE elements inserted by a SREF instruction in the GDSII format, an AREF instruction, or a given structural element may be inserted by PLACEMENT and REPETITION (OASIS format) One or more CELL elements.

Referring to Fig. 2 , which illustrates a generalized functional diagram of an inspection system, the inspection system is configured to provide a defect classification process in accordance with certain embodiments of the subject matter disclosed herein.

The system includes one or more inspection tools 210 , engineering stations 230 , classification units 220, and design data storage units 120 operatively coupled between the systems. The design material storage unit 120 includes a clip server 121 . Optionally, the scrapbook server can be implemented as a stand-alone computer, or the functionality of the scrapbook server can be fully or partially integrated with, for example, one or more inspection tools and/or with engineering stations. The classification unit 220 includes a processor and a memory unit 224. The processor and the memory unit 224 are operatively coupled to the engineering interface 221 , coupled to the design data interface 222, and coupled to the defect data interface 223. The engineering interface 221 Configuring to enable data exchange with the engineering station, the design material interface 222 is configured to request and obtain data from the design data storage unit 120 (including the design data scrapbook), and the defect data interface 223 is configured to enable and Data exchange of one or more inspection tools. The processor and memory unit are further operable to store data necessary for operation of the classification unit 220 , and are operative to receive data received from the engineering interface 221 , the design data interface 222, and the defect data interface 223 .

The classification unit can be implemented as a separate tool for use with one or more inspection tools, or the functionality of the classification unit can be at least partially integrated with, for example, one or more inspection tools and/or with engineering stations.

Further details of the system and system functions are detailed with reference to Figures 3 through 7 .

The subject matter of the present disclosure is not limited by the specific architecture shown in FIG . 2 , and equivalent and/or modified functions may be combined or divided in another manner, and equivalent and/or modified functions may be any suitable software, The combination of firmware and hardware is implemented.

For purposes of illustration only, the following description is provided with respect to CAD-based attributes. As such, the embodiments are applicable to other suitable design materials.

Referring to Figure 3 , there is shown a generalized flow diagram of a defect classification process in accordance with certain embodiments of the subject matter disclosed herein.

The classification unit obtains ( 301 ) an input material associated with a production layer on the wafer and associated with defects revealed in the production layer (hereinafter referred to as the production layer of interest). The source and demand of a list of input data can be specified by the user. Defect-related input data associated with the production layer of interest may include defect indication data, design instruction data, and combinations of the above. By way of non-limiting example, the defect-related input data may include, for example, inspection results (eg, defect locations in CAD coordinates or defects in inspection coordinates, and information indicating adjustments to design data); defect images, design data, and layer lists ( For example, the list of CAD layers typically includes a CAD layer visible on the production layer of interest, but may also contain a CAD layer corresponding to the previous production layer and/or some CAD marking layers; the design data may be in the original GDS/OASIS format or Convertible to any other vector format); data indicating design data adjustments (eg, CAD-to-wafer rotation and offset created from CAD-based recipes or from external knowledge, transformation matrices, etc.); patterned libraries (eg Included in an external binary file), rules for generating the resulting CAD layer (eg, DRC scripts), and so on.

The classification unit also obtains ( 302 ) a predetermined decision algorithm. Decision algorithms can be obtained from engineering stations. By way of a non-limiting example, the decision algorithm can be predefined by the user, as will be described in further detail with reference to FIG . The same decision algorithm can be used for the production layers of interest for all individual wafers. Alternatively, for each next wafer, the decision algorithm can be automatically adjusted based on the prior wafer classification process.

The classification unit further processes ( 303 ) the obtained input data using the obtained decision algorithm, the obtained input material being associated with the production layer of interest and related to the revealed defect. In accordance with certain embodiments of the subject matter provided by the present invention, the decision algorithm pre-defines the classification operation for the production layer of interest and ranks the defects into a predetermined grouping order. Each classification operation can be associated with one or more predetermined packets, each packet having a predefined type. The type of the predetermined grouping may be selected as one of the following types: 1) a classification group corresponding to a defect that is ultimately classified by the current classification operation and thus removed from further processing; 2) corresponding to the classification not being performed by the current classification operation And because of the mismatching of the classification operations defined by the decision algorithm (eg, unit-to-cell defects only in the decision algorithm for wafer-to-wafer classification), the grouping of defects removed from further processing 303 ; 3) corresponding to the current The operation is identified as a packet assigned to a further classification operation and thus a defect to be processed when performing the predetermined classification operation; 4) identified by the current classification operation as being assigned to one of the subsequent classification operations and thus to be taken by each next The classification operation processes the grouping of defects up to further classification.

The same defect can be sorted into more than one group, as appropriate. By way of a non-limiting example, defects can be classified into classification groups, and defects can also be sorted into packets that collect certain defects for further statistical analysis.

Depending on the situation, different classification operations may share the same classification group (a nuisance defect with different properties identified as nuisance by different classification operations may be placed in a nuisance group).

By way of non-limiting examples, the decision tree template can be used to pre-define the order of the classification operations and the individual associated groups, as will be further detailed with reference to Figures 4 through 6 .

Thus, the classification operations are provided according to a predefined order; at each current operation, the classification unit classifies ( 304 ) the defects into one or more classification packets and the classification unit ranks ( 305 ) the remaining defects to be used by one or More subsequent sorting operations are processed or filtered from the entire sorting process. During subsequent operations, the classification unit further classifies ( 304 ) the defects and ranks the defects that are to be processed by one or more subsequent classification operations or removed from further processing ( 305 ). The process continues ( 306 ) until all defects are sorted, removed from further processing (eg, unit-to-cell defects in the case of wafer-to-wafer defect classification), and/or stored for another sorting process (eg, manual sorting) .

In some embodiments, the predetermined order of classification operations and the packets associated with the operations may be defined by different considerations. Certain classification operations may be provided in parallel (eg, the same defects are classified according to different design rule checks (DRCs)).

By way of a non-limiting example, the order and grouping can be defined such that the number of defects specified by the processing of each next sorting operation is reduced.

The order of operations and the grouping of such operations can be further defined by increasing the amount of input data required for each next sorting operation. By way of non-limiting example, the classification operation based only on the defect attribute can be followed by the classification operation based on the defect attribute and the defect image; the classification operation based on the defect attribute and the defect image can be further followed by the need for the pattern library and/or other CAD data. Classification operations; the classification operations that require a pattern library and/or other CAD data can be further followed by additional operations that require DRC analysis; Alternatively or additionally, the order of the classification operations may be defined by the number of defects required for the classification process. Via a non-limiting example, operations that require a large number of defects to be processed for classification decisions (eg, statistically based classifications) may precede operations that require fewer defects (eg, classification based on defect attributes). By way of another non-limiting example, a classification operation requiring a large number of defects to be processed for classification decisions may be at the end of the classification process, and the foregoing classification operation may sort the appropriate defects to the group designated for processing by the classification operation (viewing In the case, in addition to sorting the defects into individual classification groups, it is also possible to provide for sorting certain defects to such operations).

Processing can be provided in runtime mode ( 303 ). Depending on the situation, processing can be provided offline and separately for each inspection result file ( 303 ).

By way of a non-limiting example, the classification operation and the sequence of groupings of the classification operations can be predefined using a decision tree template. Figure 4 illustrates a non-limiting schematic example of a decision tree template generated by computerized classification of defects in accordance with certain embodiments of the subject matter disclosed herein.

The illustrated decision tree may also be part of a more general classification/sorting recipe based on a template decision tree group of production layers of interest. Defects to be processed by the sorting operation 401 may be defects that have not been classified by prior classification processing using defect size and shape attributes, defect images, signature analysis, and the like.

The classification operations in the illustrated decision tree are provided using design-based attributes.

Optionally, the classification unit calculates CAD attributes for each of the defects to be processed prior to the classification operation.

During the sorting operation 401 (library matching operation), each defect is analyzed based on the matching of any of the predefined library patterns in the vicinity of each defect. If the match occurs, the defect is sorted into one of the classification packet 410 to the classification packet 411 . If no match occurs, the defect is ordered for further processing by the sort operation 402 .

Classification operation 402 provides a classification of design attributes (CDA) - based on the CAD layer and the data type of the polygons near the defect. For each defect to be processed, the classification operation produces a data set indicating the data type and layer properties of all polygons in the vicinity of the defect (for example, each polygon in the CAD file of the GDS/OASIS format is characterized by a layer Number and number of data types). Comparing the generated data set with one or more predefined data sets, the one or more predefined data sets including classification related instructions associated with a combination of data types and layer attributes .

When generating a decision algorithm, the user can predefine a predefined set of data, or a predefined set of data can be generated by operation 402 in accordance with a predefined user's settings.

In the non-limiting example illustrated, the resulting data set indicating the data type and layer attributes of all polygons near the defect to be processed by operation 402 is analyzed based on the matching of the four predefined data sets. If a defect corresponds to the first predefined set of data, operation 402 classifies the defect into a classification packet 412 . The defects corresponding to the second predefined data set, the third predefined data set, and the fourth predefined data set are sorted to be processed by operation 403 , operation 404 , operation 405, respectively . The remaining defects are ordered to be processed by the classification operation 406 . Operations 403 through 405 (with Design Rule Check (DRC) operations in conjunction with Classification of Design Attributes (CDA)) process the defects to check if the defects meet the criteria predefined in the individual design rules (DRC 1 is used for operation 403) DRC 2 is used for operation 404 and DRC 3 is used for operation 405 ). The classification by operation 403 to operation 405 is provided in accordance with the correlation between the design rules and the revealed defects. Defects that satisfy the individual rules are correspondingly classified into a classification packet 414 , a classification packet 416, and a classification packet 418 . The remaining defects are ordered to be processed by operation 406 (packet 415 , packet 417, and packet 419, respectively ). Operations 403 through 405 are set to be provided in parallel.

Classification operation 406 provides a design-based packet (DBB) to packet 420 to packet 421 for defects previously assigned to packet 415 , packet 417, and packet 419 and by operation 402 to be processed by operation 406 . The DBB packets are arranged to group defects according to the similarity of patterns at the vicinity of the defects. Contrary to the library, the pattern for the DBB is not predefined. The defects sorted to the DBB packet can be further sampled according to the user policy.

By way of a non-limiting example, the execution of a decision algorithm according to the illustrated tree can be provided as follows: - CAD-based attributes can be calculated by the classification unit during the inspection scan; - the scrapbook can be imported into the classification unit via the runtime scrapbook server; - CDA, DRC and library matching calculations can be performed during the inspection scan; The DBB that can be sampled when the inspection scan is completed.

DRC analysis can be provided for each scrapbook. The advantages of this method are as follows: no special DRC station is required; no export CAD layer is required for each DRC rule at the DRC station; the DRC sequence can be applied; the DRC script development program is simple; the changes in the parameters do not need to be re-established new Export the CAD layer.

Figure 5 illustrates a non-limiting example of an operational sequence table generated by analyzing a decision tree defined by the template in Figure 4 . At the beginning, each defect is initialized by the initial classification code (for example, 10). Step _K includes collecting a defect having a classification code according to the row "input classification code"; obtaining an input parameter, and performing a classification algorithm on the individual update classification code of the defect and the input defect collected at step _K before proceeding to step _K+1 (Line "algorithm").

Figure 6 illustrates a screenshot of another non-limiting illustrative example of a decision tree template 600 and individual results 608 to the defect classifications in the predetermined groupings specified in table 609 . In the illustrated decision tree template, the cluster classification operation 601 filters out individual defects to the group "cluster". The next classification operation 602 filters out the unit-to-unit defects to the group "C2C". The remaining defect library system by matching the operating handle 603, 603 to the library matching operation defect classification to "Kept by Lib" by grouping or sorting operation for processing 604, 604 operate in conjunction to provide design rules by classifying design attributes of checks (DRC+CDA). Operation 604 sorts the processed defects into defects to be processed by operation 605 and defects to be processed by operation 606 . If the defect matching size criteria processed by operation 605 is processed, the defects are sorted into "DRC" classification packets. Otherwise, the defects are ordered to be processed by operation 606 . The library matching operation 606 processes the defects sorted to the operation by operations 604 and 605 , and classifies the matching defects into the classification packet "Filtered by Lib." Sort the remaining defects to the group "Others".

An advantage of certain embodiments of the disclosed subject matter is to enable runtime identification and processing defects that occur as a result of interaction between the design layout and the process. An online design rule check can be provided near each defect that matches the identified defect of the predefined standard, while multiple design rule checks (DRCs) can be provided in a single execution.

Figure 7 illustrates a generalized flow diagram of computerized generation of a decision algorithm in accordance with certain embodiments of the disclosed subject matter. A decision algorithm for the production layer of interest can be generated at the engineering station 230 . The user can set ( 702 ) the CAD attributes used during the sorting process. CAD attribute settings may include selection of CAD layers (eg, a list of CAD layers may include CAD layers visible on the production layer of interest, CAD layers corresponding to previous production layers, CAD mark layers, etc.) for library matching operations Selection and/or generation of CAD patterns, design inspection rules, and the like. The user can further pre-define the grouping, set ( 703 ) the sorting operations, and the order of the sorting operations, and associate the groupings with the sorting operations. The engineering station automatically generates ( 704 ) a decision algorithm based on the user's settings. The engineering station may further obtain ( 701 ) inspection results (eg, defect files) having classified defects (eg, classified into DOI, errors, and nuisance defects) from the inspection tool. If necessary, the data received from the inspection tool can be used to validate ( 705 ) the resulting decision algorithm and adjust the user's settings.

In the description provided, numerous specific details are set forth to provide a thorough understanding of the invention. However, it should be understood by those skilled in the art that the invention may be practiced without the specific details. In other instances, well-known methods, procedures, components, and circuits are not described in detail to avoid obscuring the invention.

It will also be appreciated that the system in accordance with the present invention can be a suitably programmed computer. As such, the present invention contemplates a computer program readable by a computer for performing the methods of the present invention. The present invention further contemplates a machine readable memory tangibly embodying a program of instructions executable by a machine for performing the methods of the present invention.

It will be readily understood by those skilled in the art that the various modifications and variations can be applied to the embodiments described above without departing from the scope of the embodiments as defined by the appended claims.

110‧‧‧ wafer

120‧‧‧Design database storage unit

130‧‧‧Manufacture process

140‧‧‧Checking process

150‧‧‧Classification process

Claims (15)

A computer implemented method for classifying defects detected on a production layer of a sample, the method comprising: using a computer to perform the steps of: obtaining input data related to the detected defects; using and The production layer is associated with a decision algorithm to process the defects to perform the steps of: specifying two or more classification operations and one of the two or more classification operations; and An ordered ranking of defects, wherein each of the predetermined packets is associated with at least one classification operation, wherein at least one classification operation ranks at least a portion of the processed defects to one or more classification packets to produce a final a defect of the classification, and wherein each of the classification operations except the last classification operation sorts at least a portion of the processed defects to be processed by one or more of the subsequent classification operations; and in a storage medium Store at least the defects of the final classification. The method of claim 1, wherein the defects sorted by one of the current classification operations to be processed by one or more of the subsequent classification operations are sorted to one or more packets, the one or more More groupings are associated with the current classification operation and are selected from a group comprising: a classification operation that is not classified by a current classification operation and that is defined by the decision algorithm because of a mismatch since the further Processing the grouping of such defects removed; Corresponding to a packet identified by a current operation assigned to one of the predefined further classification operations and thus a defect to be processed when performing the predetermined classification operation; corresponding to being assigned to one of the subsequent classification operations The grouping of defects identified by the current classification operation and thus to be processed by each of the next classification operations up to further classification. The method of claim 1, wherein the same defect is ordered to more than one packet. The method of claim 1, wherein at least one of the packets is associated with at least two classification operations. The method of claim 1, wherein the input material includes design instruction data. The method of claim 1, wherein the at least one sorting operation sorts the processed defects based on the matching for each defect, the matching being a data indicating a data type and a layer attribute of all polygons in the vicinity of the defect A set of matches containing one or more predefined data sets associated with a combination of a data type and a layer attribute. The method of claim 1, wherein at least one of the classification operations uses design rule check (DRC) analysis and by design attributes A combination of classification by design attributes (CDA) ranks the processed defects. The method of claim 1, wherein the order of the classification operations is predefined according to at least one of: a. reduced to the number of defects specified by each next classification operation; b. The amount of input data required for the classification operation; c. provide a classification operation for one of the large number of defects to be processed for the classification decision before the classification operation for which the classification decision is to be processed; and d. After the classification operation of the decision to process fewer defects, providing a classification operation that requires a large number of defects to be processed for the classification decision, and the classification operation is set to be processed by the classification operation that needs to process a large number of defects for a classification decision, The appropriate defect ordering of the aggregates is processed. The method of claim 1, wherein the at least two classification operations use the Design Rule Check (DRC) analysis to sequentially order the processed defects substantially simultaneously, each of the classification operations corresponding to a different design rule. The method of claim 1, the method further comprising the step of calculating a design attribute of each of the defects to be processed during an inspection scan. The method of claim 10, wherein any one of a classification of design attributes (CDA), a DRC, and a library matching calculation required for the classification operations is provided during the inspection scan. The method of claim 10, wherein the at least one classification operation comprising the design-based grouping is provided after the inspection scan. An inspection system capable of classifying defects detected on a production layer of a sample, the inspection system including a sorting unit operatively coupled to the at least one inspection tool and operatively coupled to a design a data storage unit, wherein the classification unit comprises: a defect data interface configured to obtain defect indication data from the at least one inspection tool; and a design data interface configured to store from the design data The unit obtains designation instructions; a memory configured to store a defect of the final classification; and a processor configured to process the defect indication using a decision algorithm associated with the production layer And the designation instruction data to: specify two or more classification operations and one of the two or more classification operations; the processor is further configured to process the processed according to a predetermined grouping Defect indication information and sorting of the design instruction data, Wherein each of the predetermined groupings is associated with at least one classification operation; wherein at least one of the classification operations sorts the processed defect indication material and at least a portion of the designation instruction material to one or more classification groups to produce a final classification a defect; and wherein each of the classification operations except the last classification operation sorts the processed defect indication material and at least a portion of the design indication data to be one or more of the subsequent classification operations deal with. The system of claim 13, wherein the defects sorted by one of the current classification operations to be processed by one or more of the subsequent classification operations are ordered to one or more packets, the one or more More groupings are associated with the current classification operation and are selected from a group comprising: a classification operation that is not classified by a current classification operation and that is defined by the decision algorithm because of a mismatch since the further Processing a group of the defects that are removed; corresponding to a packet identified by a current operation assigned to a predefined further classification operation and thus a defect to be processed when performing the predetermined classification operation; corresponding to by assigning to A grouping of defects of one of the subsequent classification operations that are identified by the current classification operation and thus are to be processed by each of the next classification operations up to further classification. A classification unit operatively coupled to an inspection system capable of classifying defects detected on a production layer of a sample, the classification unit including a processor operatively coupled to The processor has access to a memory and the memory is configured to store a defect of a final classification, wherein the processor is configured to process the defect indication data and design instructions using a decision algorithm associated with the production layer Data for performing the steps of: specifying two or more classification operations and a sequence of one or more of the two or more classification operations; the processor being further configured to process the defect indication data and the Designing an instruction ordering; wherein each of the predetermined groupings is associated with at least one classification operation; wherein at least one classification operation sorts the processed defect indication material and at least a portion of the design indication data to one or more categories Grouping to produce defects in the final classification; and each classification operation except the last classification operation will be This design defect indication information and the information indicative of at least a portion to be treated by ordering one or more of the classification operations.