Measurement system, measurement method, and measurement process fabrication method

Abstract

The object of this invention is to provide a measurement system that can easily analyze the cause of bonding defects between two joined wafers. Therefore, this invention is a measurement system for measuring semiconductor devices, comprising a pre-joint measurement apparatus for measuring a first wafer and a second wafer before joining, wherein the pre-joint measurement apparatus sets the first wafer as the measurement target for a first measurement point, and sets a second measurement point connected to the first measurement point after joining as the measurement target for the second wafer.

Description

Technical Field

[0001] This invention relates to a measurement system, a measurement method, and a measurement process manufacturing method. Background Technology

[0002] In the manufacturing process of semiconductor devices, various inspections and measurements are performed appropriately. As a technique for generating programs to operate the devices used in inspection and measurement, Patent Document 1 is known, for example. Patent Document 1 discloses that a computer automatically generates a sample program for inspection and measurement by having the user select various items such as image data of the object to be inspected and measured, the area to be inspected and measured, the content to be inspected and measured, and the parts to be inspected and measured using an input device such as a mouse on a computer.

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 2006-65582 Summary of the Invention

[0006] The problem that the invention aims to solve

[0007] In the technology described in Patent Document 1, inspection and measurement procedures are made separately for different types of wafers. Therefore, for example, when inspecting and measuring the top and bottom wafers in a bonding wafer, the layout relationship is sometimes not taken into account, making it impossible to analyze the cause of bonding defects.

[0008] The purpose of this invention is to provide a measurement system that can easily analyze the causes of bonding defects between two joined wafers.

[0009] Methods for solving problems

[0010] To address the aforementioned issues, the present invention provides a measurement system for measuring semiconductor devices, comprising a pre-bonding measurement apparatus for measuring a first wafer and a second wafer before bonding, wherein when the first wafer is designated as the measurement target, the pre-bonding measurement apparatus designates a second measurement target that will be connected to the first measurement target after bonding as the measurement target of the second wafer.

[0011] Invention Effects

[0012] According to the present invention, a measurement system can be provided that can easily analyze the cause of poor bonding between two joined wafers.

[0013] The issues, structures, and effects other than those described above become clear through the following description of the implementation methods. Attached Figure Description

[0014] Figure 1 This is a schematic diagram illustrating the process flow of semiconductor device manufacturing, focusing on the portion related to wafer bonding, and the wafer structure at this stage.

[0015] Figure 2 It is a system structure diagram related to the manufacturing process of semiconductor devices.

[0016] Figure 3 This is a block diagram showing the structure of an SEM-type measuring device, which is one of the measuring devices before bonding.

[0017] Figure 4 This is an illustration of the pattern layout of the bottom and top wafers, which are the objects of measurement.

[0018] Figure 5 It is a flowchart of the manufacturing process.

[0019] Figure 6 It is a flowchart of the process changeover procedure.

[0020] Figure 7 It is a flowchart of the actions performed by the measurement program.

[0021] Figure 8 It is a flowchart of the analysis program's actions.

[0022] Figure 9 Specific examples to indicate comparison.

[0023] Figure 10 Another specific example of comparison.

[0024] Figure 11 Another specific example of contrast.

[0025] Figure 12 Examples of items in a data table.

[0026] Figure 13 This is an example of a GUI displayed when the analysis program is executed.

[0027] Figure 14 It is a schematic diagram of the object's structure, SEM image, structural cross-section, and cross-sectional outline of the SEM image.

[0028] Figure 15 This is a schematic diagram illustrating the calculation of the characteristic quantities of the depression.

[0029] Figure 16 It is a flowchart of the measurement process when the measurement procedure corrects the process according to the measurement results. Detailed Implementation

[0030] The embodiments of the present invention will be described with reference to the accompanying drawings.

[0031] Example 1

[0032] <Instructions for Wafer Bonding and Measurement>

[0033] Figure 1 This is a schematic diagram illustrating the process flow of semiconductor device manufacturing, focusing on the portion related to wafer bonding, and the wafer structure at that point.

[0034] Before wafer bonding, patterns are formed on the bottom wafer 103 (first wafer) and the top wafer 102 (second wafer) through multiple processes 101. To fabricate multiple semiconductor devices from a single wafer, repeating patterns are formed on the wafer. Here, the repeating unit ultimately used as a semiconductor device is referred to as a chip, which is a section divided by vertical and horizontal lines in the bottom wafer 103 and the top wafer 102. The repeating pattern of a portion of the chip on the bottom wafer 103 is shown in structural cross-sectional view 133, and the repeating pattern of a portion of the chip on the top wafer 102 is shown in structural cross-sectional view 132.

[0035] After patterning, a copper (Cu) plating layer 104 is applied to form the wiring portion. The surface of this plating layer is ground using CMP 105 (Chemical Mechanical Polishing), intentionally retracting it to form Cu pads 106. This retraction amount is referred to as a recess 107. In the subsequent bonding process 108, the top wafer 102 is flipped around the Y-axis 110, bringing it into contact with the bottom wafer 103. The Y-axis 110 is determined by a notch 109, which defines the rotational direction within the wafer's surface, and passes through the center of the wafer. At this time, by applying pressure and heating, the Cu pads 106 expand and come into contact with each other, thereby bonding and ensuring conductivity, thus creating the bonded wafer 112. Although the example of retracting the Cu pads 106 is described later, it is not always necessary to retract them. Sometimes Cu pads of other shapes, such as protruding or flat shapes, are bonded, allowing the invention to be applied without being limited by the shape of the Cu pads. Furthermore, since both the surface and back of the bonded wafer 112 have substrates (Si, etc.) for the bottom wafer 103 and the top wafer 102, the repeating pattern is not exposed. As shown in the cross-sectional view 141 near the bonding surface 140 of the bonded wafer 112, a repeating pattern is formed internally. By performing multiple processes 111 on the bonded wafer 112, a semiconductor device is completed.

[0036] In the semiconductor device manufacturing process, appropriate inspection and measurement (hereinafter referred to as measurement) are performed at critical points in the process. Figure 1The diagram shows an example of pre-joining measurement 121 performed after CMP 105 and post-joining measurement 122 performed after joining process 108. Measurements other than these timings can be performed before or after joining.

[0037] In the pre-bonding measurement 121, the surface of the Cu pad, one of the measurement objects, is exposed, allowing for direct measurement. Pre-bonding measurement methods include foreign object inspection, SEM (Scanning Electron Microscope) based length measurement, SEM-based step (depression) measurement, AFM (Atomic Force Microscope) measurement, and light-based shape measurement. Foreign object inspection, length measurement, and shape measurement are suitable for revealing systematic defects that repeatedly occur in recurring patterns, especially for randomly generated defects.

[0038] In post-bonding measurement 122, the bonding surfaces of the Cu pads to be measured are not exposed, and measurements need to be performed from the back side (the unpatterned side) of the bottom wafer 103 or the top wafer 102. As a post-bonding measurement method, ultrasonic defect inspection, X-ray defect inspection, infrared defect inspection, etc., can be used. After cutting the back side of the post-bonding wafer 112 to expose the pattern, the same measurements as before bonding, utilizing electrical characteristics, can also be performed.

[0039] In the pre-bonding measurement 121, for example, the amount of recess 107 of the Cu pad 106 is measured (referred to as recess amount). As described above, during the bonding process 108, the Cu pad 106 expands due to overheating. Therefore, if the recess amount is too large, the Cu pads of the bottom wafer 103 and the top wafer 102 will not make contact or the contact area will be insufficient, resulting in poor conductivity. At this time, a void, referred to as a cavity, is generated between the wafers. As a post-bonding measurement 122, for example, if ultrasonic defect inspection is performed, the cavity between the bottom wafer 103 and the top wafer 102 can be detected. Conversely, if the recess amount is too small, the expansion of the Cu pad 106 will exert a repulsive force on the wafers, generating a cavity around the Cu pad 106. When the cavity becomes large, the insulation performance changes, or it affects the conductivity of the surrounding Cu pads. Thus, the recess amount is an important dimension for the conductivity of the left and right devices. For example, in the SEM used in pre-bonding measurement 121, the resolution is sufficiently high (below 0.1 nm) compared to the Cu pad size (hundreds of nanometers), allowing for the measurement of the amount of depression on each pad. However, in the ultrasonic inspection used in post-bonding measurement 122, the resolution is at least around 10 μm, making it difficult to determine whether there are voids on a per-Cu pad basis. Therefore, it becomes important to correlate the measurement results of post-bonding measurement 122 with those of pre-bonding measurement 121 to analyze the causes of voids. Therefore, when the measurement target of the bottom wafer 103 is set as the first measurement point, the second measurement point connected to the first measurement point after bonding is set as the measurement target of the top wafer 102. This allows for analysis (comparison) of which wafer, either the bottom wafer 103 or the top wafer 102, has an abnormal amount of depression. In the case of checking continuity in post-bonding measurement 122, it is also an inspection on a per-chip basis, an inspection of a portion of the chip, and therefore the resolution is different from that of pre-bonding measurement 121. Not only do the resolutions differ, but the accuracy of position determination also varies. Therefore, when making comparisons, both resolution and position determination accuracy need to be considered in the analysis (comparison). For example, if the position determination accuracy and resolution of measurement 122 are low after bonding, it is necessary to extensively analyze (compare) the results of measurement 121 before bonding. If the amount of recess can be determined within the wafer surface, it will help improve the CMP process.

[0040] The following describes measurements other than the amount of concavity. In the pre-bonding measurement 121, not only concavity is measured, but also the surface roughness of the Cu pads, the crystal orientation of the Cu pads, and the size and type of foreign matter. Roughness and crystal orientation affect the amount of expansion and the shape after expansion, thus influencing conductivity. If foreign matter is present, voids are generated near the foreign matter during bonding, resulting in poor conductivity. The size of the void varies depending on the size and type of the foreign matter, and the conductivity also changes accordingly. Therefore, the culpability of the foreign matter can be determined based on its size and type. These measurements are essential for the stable mass production of semiconductor devices. In wafer bonding, it is necessary to understand the relationship between the wafers being bonded. Therefore, measuring the wafers to be bonded at the time of bonding is useful in process development and process management.

[0041] <System Structure Description>

[0042] Figure 2 This is a system architecture diagram of the semiconductor device manufacturing process, particularly focusing on the measurement system. The MES200 (Manufacturing Execution System) provides execution instructions to the pre-bonding measurement device 201, the post-bonding measurement device 202, and other manufacturing apparatus group 203, including CMP devices and bonding devices. The pre-bonding measurement device 201 performs pre-bonding measurement 121, and the post-bonding measurement device 202 performs post-bonding measurement 122. Furthermore, the number of measurement devices is not limited to two; there may be one or more.

[0043] The pre-joining measuring device 201 and the post-joining measuring device 202 read the process that specifies the measurement conditions for performing the measurement procedure from the DB (Data Base) 204 via the network, perform the measurement, and save the measurement results in the DB 204 via the network.

[0044] The analysis server 206 (analysis device) of the analysis program 205 obtains data from the DB 204 and performs analysis (e.g., comparison of measurement results), and displays the results on the display unit 207. The display unit 207 has input / output devices (not shown) that receive analysis conditions from the user and send them to the analysis server 206.

[0045] Figure 3 This is a block diagram showing the structure of an SEM-type measuring device, which is one of the measuring devices used before bonding. For example... Figure 3 As shown, the SEM measuring device 3001 includes an electron microscope 3100 (measuring unit), a control unit 3120, power supply units 3121 and 3122, and a computer 300.

[0046] In the electron microscope 3100, an electron beam 3103 (electron beam) is irradiated onto the sample. The electron microscope 3100 outputs a detection signal based on the irradiation by the electron beam 3103. The SEM-type measuring device 3001 includes the components necessary to form a signal waveform and an image based on the detection signal from the electron microscope 3100. First, refer to... Figure 3 Specifically, an example of an electron microscope 3100 is described.

[0047] An electron beam 3103, drawn from an electron source 3101 by an extraction electrode 3102, is accelerated by an accelerating electrode (not shown). The accelerated electron beam 3103 is then focused by a focusing lens 3104, which acts as a focusing lens. The focused electron beam 3103 is then scanned one-dimensionally or in two-dimensionally on a sample 3108 by a scanning electrode 3105. The electron beam 3103 is decelerated by a negative voltage applied to an electrode built into a sample stage 3109 and focused by the lens of an objective lens 3106, illuminating the sample 3108.

[0048] When the electron beam 3103 irradiates the sample 3108, electrons 3110, which are secondary electrons and backscattered electrons, are released from the irradiated area (irradiated area) and / or from the irradiated area into the interior of the sample 3108. These released electrons 3110 are accelerated towards the electron source 3101 by the acceleration effect generated based on the negative voltage applied to the sample 3108, and collide with the conversion electrode 3112, generating secondary electrons 3111. The secondary electrons 3111 released from the conversion electrode 3112 are captured by detectors 3113 and 3114, and the detection signals output by detectors 3113 and 3114 vary according to the amount of secondary electrons 3111 captured. Additionally, detectors for reflected electrons (not shown) are sometimes also configured. The configuration and structure of the detectors are examples; sometimes multiple detectors are used. With multiple detectors, multiple images can be obtained in a single electron beam scan.

[0049] The detection signal output from detector 3113 is provided to computer 300 by control unit 3120. Computer 300 has a display unit (not shown). The brightness of the image displayed on this display unit varies according to the detection signal. That is, the amount of electrons captured by detector 3113 (electron quantity) is displayed on the display unit as brightness.

[0050] For example, when displaying a two-dimensional image in the display unit, synchronization is achieved between the deflection signal supplied to the scanning electrode 3105 and the detection signal output from the detector 3113, thereby displaying the brightness of the image in the scanning area scanned by the deflection signal in the display unit.

[0051] In addition, Figure 3The electron microscope 3100 shown includes a deflector (not shown) that moves the scanning area of ​​the electron beam 3103. This deflector is used to display images of patterns of the same shape existing at different positions on the display unit. This deflector is also called an image shift deflector, which allows the field of view position of the electron microscope 3100 to be moved without moving the sample 3108 via a sample stage (e.g., sample stage 3109) that moves the sample 3108. Alternatively, the image shift deflector and the scanning electrode 3105 can be a shared deflector, and the signal for image shifting and the deflection signal can be superimposed and provided to the deflector.

[0052] The detection signals (image, brightness profile, brightness, etc.) from the electron microscope 3100 are supplied to the computer 300 via the control unit 3120. The computer 300 calculates values ​​related to shape changes of the shape of interest of the observed object based on the supplied detection signals and outputs one or more of these calculated values. Furthermore, the computer 300 may also be integrated with the electron microscope 3100.

[0053] The control unit 3120 controls the power supply units 3121 and 3122 according to instructions from the computer 300. By controlling the power supply unit 3122, the voltage applied to the extraction electrode 3102 and the accelerating electrode (not shown) is changed. Similarly, by controlling the power supply unit 3121, the voltage applied to the sample 3108 is changed. Furthermore, the control unit 3120 controls the deflection signal supplied to the scanning electrode 3105 and the signal supplied to the objective lens 3106 according to instructions from the computer 300. Moreover, as described above, the control unit 3120 supplies the detection signal output from the detector 3113 to the computer 300.

[0054] The computer 300 includes an input / output device 301 for receiving information from the user and displaying the results, a processor 302 for executing programs, and a memory 303 as a storage area. The memory 303 stores a process manufacturing program 311, a process change program 312, a measurement program 313, a process 314, and measurement results 315 with symbols (IDs) for identifying the measured object.

[0055] <Description of the measurement object>

[0056] Figure 4 This is an illustration of the pattern layout of the bottom and top wafers, which are the objects of measurement.

[0057] Chip magnification view 401 shows an enlarged view of the vicinity of a chip on the top wafer 102. Chip magnification view 401 has an assembly of Cu pads that are conductive after wafer bonding, i.e., pattern 402. A layout magnification view 403, which further enlarges chip magnification view 401, is shown. In the pattern within the wafer, there is a part of the assembly of Cu pads, i.e., pattern 402, i.e., Cu pad 404, which is related to the performance of the device that is conductive after wafer bonding, and a virtual pattern 405 that is unrelated to the performance of the device that is not conductive after wafer bonding. In addition, there is an alignment mark 407 within the chip, which is used to specify the alignment when the wafer is mounted on the device and the position of the measurement point, i.e., the point where measurement is performed within the chip. A coordinate system 408 is defined with the feature portion of the alignment mark 407 as the origin and the axis perpendicular and parallel to the Y-axis 110 of the chip as the XY axis.

[0058] Chip magnification view 411 shows an enlarged view of the vicinity of a chip on the bottom wafer 103. Chip magnification view 411 has an assembly of Cu pads, i.e., pattern 412, that will be conductive after wafer bonding. A layout magnification view 403, further enlarged from chip magnification view 411, is shown. Within the pattern on the wafer, there is a portion of the assembly of Cu pads, i.e., pattern 412, i.e., Cu pad 414, which is related to the performance of the conductive device after wafer bonding, and a virtual pattern 415 unrelated to the performance of the non-conductive device after wafer bonding. Additionally, there is an alignment mark 417 within the chip, which is used to define the alignment when the wafer is mounted on a device and the position of the measurement point within the chip. A coordinate system 418 is defined with the characteristic portion of the alignment mark 417 as the origin, and an axis perpendicular and parallel to the Y-axis 110 of the chip as the XY axis.

[0059] The axis parallel to the Y-axis 110 of the chip and passing through the center of the chip is called the top wafer inversion axis 409 and the bottom wafer inversion axis 419. During wafer bonding, the top wafer 102 is inverted with the Y-axis 110 of the chip as the center for bonding. Therefore, the pattern 402 of the top wafer 102 and the pattern 412 of the bottom wafer 103 are inverted via the inversion axis 409 or the inversion axis 419. That is, as shown in layout enlarged views 403 and 413, the Cu pads 404 and 414 have the same spacing. However, the shapes of the Cu pads 404 and 414 are not necessarily the same.

[0060] The measurement points in this specification are explained. Measurement points are information indicating the location and extent of the object being measured, such as the XY number of the chip being measured within a wafer plane, the center coordinates of the measurement, and the measurement range relative to its center. Figure 4In this diagram, measurement points 421 and 422 represent the measurement points (blacked-out portions) of the top wafer 102, and measurement points 431 and 432 represent the measurement points (blacked-out portions) of the bottom wafer 103. Therefore, measurement points 421 and 431 are reversed relative to the chip's Y-axis 110, and measurement points 422 and 432 are reversed relative to either the reverse axis 409 or the reverse axis 419; that is, when considered in X-coordinates, this results in a relationship that transforms X into -X. This allows for the measurement of Cu pad pairs that are in contact with each other during bonding in the bonding process 108. Consequently, when defects or other anomalies occur in the post-bonding measurement 122, it is possible to analyze which of the top wafer 102 or the bottom wafer 103 is at fault. Thus, in this specification, the analysis of measurement results based on the positional relationship of the measurement points is referred to as a comparison.

[0061] When the positional relationship between the Y-axis and the inversion axis 409 of coordinate system 408 when measuring the top wafer 102 by the pre-bonding measurement device 201 is different from the positional relationship between the Y-axis and the inversion axis 419 of coordinate system 418 when measuring the bottom wafer 103 by the pre-bonding measurement device 201, the difference in positional relationship needs to be taken into account when performing the transformation.

[0062] <Explanation of Manufacturing Processes and Process Changeovers>

[0063] The fabrication and transformation of processes that include measurement point information are explained.

[0064] Figure 5 This is an example of the operation flow diagram for process fabrication procedure 311. First, process fabrication procedure 311 receives layout data 501, which includes wafer design data, image data, etc., and measurement condition information 502, such as coordinates of measurement points and shooting conditions, input by the user. Next, process fabrication procedure 311 uses the layout data 501 to register alignment marks (step S510). Then, process fabrication procedure 311 uses the measurement condition information 502 to set shooting conditions (FoV (Field of View), number of pixels, acceleration, current, cumulative scans, scanning method) (step S511). Additionally, process fabrication procedure 311 uses the layout data 501 and measurement condition information 502 to set the shooting area (start coordinates, XY steps, XY movement amount) (step S512). Furthermore, process fabrication procedure 311 uses the measurement condition information 502 to evaluate conditions (e.g., as described later). Figure 14 The evaluation conditions for the depression shown are set (step S513). After that, the process fabrication program 311 registers the chip to be photographed (step S514), and saves the process 314 containing these settings in DB204.

[0065] Hereinafter, the process used to measure the bottom wafer 103 (first wafer) will be referred to as the first process, and the process used to measure the top wafer 102 will be referred to as the second process. It is also possible to use... Figure 5 The process fabrication procedure 311 described herein is used to fabricate the first process and the second process respectively. However, considering the difference in layout between the bottom wafer 103 and the top wafer 102, inputting the measurement condition information 502 in a way that represents an inverted measurement point relationship is cumbersome and may lead to errors. Therefore, by using the process change procedure 312, it is possible to maintain the inverted measurement point relationship and perform measurements efficiently and reliably.

[0066] Figure 6 This is an example of the operation flowchart for process transformation procedure 312. Here, we will describe an example of transforming a first process related to the bottom wafer 103 into a second process related to the top wafer 102 based on information obtained by reversing the measurement point information. Furthermore, the shooting conditions and other settings in the first process are directly used in the fabrication of the second process.

[0067] First, the process conversion program 312 reads the first process 504 for measuring the bottom wafer 103 from DB204 and obtains the first measurement point information of the bottom wafer 103 as the measurement object (step S602). The first measurement point information is, for example, the chip position [x, y] as coordinate information on coordinate system 408, and the chip position [ChipX, ChipY] indicating the chip number when the center of the bottom wafer 103 is set to [0, 0]. In addition, the first measurement point information is not limited to information provided as coordinate information, but also includes information assigned by other methods such as vectors. In addition, the chip is not necessarily located at the center of the wafer, and there may be cases where the center of the wafer is, for example, the chip dividing line. On the other hand, the process conversion program 312 obtains the coordinate system information 601 from the layout data 501 (step S601). Then, the process conversion program 312 uses the coordinate system information 601 to perform coordinate transformation (reverse) on the chip position 603 and chip position 604 of the bottom wafer 103 obtained in step S602 (step S605). Furthermore, the process changeover procedure 312 reflects the chip position 606 and chip position 607 obtained by coordinate transformation into the second process 609 as the measurement object of the top wafer 102, i.e., the second measurement point information (step S608). Afterwards, the process changeover procedure 312 saves the obtained second process 609 in DB204 (step S610).

[0068] exist Figure 6The example illustrates how a first process 504 is transformed into a second process 609 to create a new second process 609, but the reverse is also possible. Furthermore, if both a first process 504 and a second process 609 exist, only the changed portion—the inverted second measurement point information from the first process 504—can be reflected in the second process 609. Alternatively, the process fabrication program 311 and the process transformation program 312 can be integrated, or the coordinate transformation 605 can be performed during the fabrication of the first process 504 without going through DB204, allowing the second process 609 to be fabricated in parallel with the first process 504. Additionally, the measurement point information from the first process can be used to fabricate a third process for measuring the bonded wafer 112.

[0069] <Instructions for performing measurements using the manufacturing process>

[0070] Figure 7 This is an example of an action flow diagram based on measurement program 313. Here, measurement program 313 for the bottom wafer 103 before bonding is designated as the first measurement program 710, measurement program 313 for the top wafer 102 before bonding is designated as the second measurement program 720, and measurement program 313 for the wafer 112 after bonding is designated as the third measurement program 730. Furthermore, the following example illustrates how the results measured using the first to third measurement programs are stored in DB204. However, measurements from other processes can also be stored in DB204, allowing for analysis across multiple processes.

[0071] When the MES200 receives a measurement wafer ID 799 as user input, it identifies the ID of the wafer to which it will be bonded, and sends the user-input wafer ID and the wafer ID of the bonding object to the first measurement program 710 and the second measurement program 720. This enables reliable measurement of the bonding object accordingly.

[0072] The first measurement procedure 710 receives the identification code (ID) of the target wafer and a measurement start instruction from the MES 200, thereby initiating the operation. Next, the first measurement procedure 710 reads the first process 504 from the DB 204 (step S711) and executes the first process 504 (step S712), thereby measuring the bottom wafer 103. Afterwards, the first measurement procedure 710 saves the first measurement result 714 with the ID in the DB 204 (step S713).

[0073] The second measurement procedure 720 receives the identification code (ID) of the target wafer and a measurement start instruction from the MES 200, thereby initiating operation. Next, the second measurement procedure 720 reads the second process 609 from the DB 204 (step S721) and executes the second process 609 (step S722), thereby measuring the top wafer 102. At this time, the second measurement procedure 720 performs a coordinate transformation on the measurement result through the inverse transformation of the coordinate transformation in step S605 of the process transformation procedure 312 (step S725). In this coordinate transformation, the coordinate system information 601 obtained from the layout data 501 is also used. This coordinate transformation (step S725) can be skipped, but it facilitates data comparison and makes it easier to visually compare the measurement results of the top and bottom wafers when displaying the data. Of course, it is not necessary to perform this in the second measurement procedure 720; it can be performed after saving to the DB 204 or during comparison. Then, the second measurement procedure 720 saves the second measurement result 724 with ID in DB204 (step S723).

[0074] The third measurement procedure 730 receives the identification code (ID) of the target wafer and the measurement start instruction from the MES 200, thereby initiating the operation. Next, the third measurement procedure 730 reads the third process 735 from the DB 204 (step S731) and executes the third process 735 (step S732), thereby measuring the top wafer 102. Afterwards, the third measurement procedure 730 saves the third measurement result 734 with the ID in the DB 204 (step S734).

[0075] <Explanation of Data Analysis (Comparison) and its Flexible Application>

[0076] Figure 8 This is an example of the operation flowchart of the analysis program 205. For example, the analysis program 205 accepts analysis condition 801 via user input and queries DB 204 for the data to be considered (step S802). Next, the analysis program 205 receives the first measurement result 714, the second measurement result 724, and the third measurement result 734 obtained from the query, corresponding to the bottom wafer 103, the top wafer 102, and the bonded wafer 112 respectively, and compares them (step S803). Afterwards, the analysis program 205 displays and saves the comparison results (step S804). At this time, since the measurement results are assigned wafer IDs, it is possible to query the data of the actual bonded pairs of top and bottom wafers and the bonded wafer from multiple wafer data sets. Furthermore, if there is a situation where multiple measurement results are queried for a single wafer, it is possible to compare the data of multiple wafer pairs with the data of the bonded wafer.

[0077] Figures 9-11 express Figure 8A specific example of the comparison in step S803.

[0078] exist Figure 9 In the diagram, the recesses representing the retraction amount of the Cu pads are extracted from the first measurement result 714 of the bottom wafer 103 and the second measurement result 724 of the top wafer 102, and used as the horizontal and vertical axes of the comparison chart, respectively. Figure 9 In the comparison chart, white-painted point 901 represents Cu pads determined to be defect-free based on the third measurement result 734, while black-painted point 902 represents Cu pads determined to be defective. Thus, based on the relationship between the amount of depression on the top wafer 102 before bonding, the amount of depression on the bottom wafer 103 before bonding, and the presence or absence of defects on the wafer 112 after bonding, the allowable depression management range 903 to prevent defects can be determined. This can be used for process management during mass production, etc.

[0079] Furthermore, the resolution of the devices may not be the same, so resolution and positioning accuracy need to be considered when making comparisons. For example, when using an ultrasonic inspection device to label the measurement results of the indentation amount of each pad with whether there is a defect (901 and 902), the binary values ​​of "defect" (901 and 902) can be used instead of representing it as a probability of presence or absence of a defect. Alternatively, the presence or absence of defects can be displayed on the pads in the images of the first and second measurement results according to these labels (901 and 902) (not shown).

[0080] Figure 10 Is with Figure 9 Examples of comparison charts showing Cu pads of different sizes and layouts. Even Figure 9 Comparison charts and Figure 10 The wafer IDs analyzed in the comparison charts are the same, but depending on the size and layout of the Cu pads, the amount of depression and the distribution of defects generated in the bonding process may not be the same. Figure 10 In the comparison chart, white-painted point 1001 indicates a Cu pad that is determined to be without defects based on the third measurement result 734, while black-painted point 1002 indicates a Cu pad that is determined to be defective. Figure 10 The recess management range 1003 in the Cu pad is located in the same position as... Figure 9 The depression management scope covers 903 different areas.

[0081] For example, during process development, the recess management ranges 903 and 1003 are determined. During mass production, in the pre-bonding measurement 121 after CMP105, if it is confirmed that the recess amount of the Cu pad is outside the recess management range 903, there are methods to issue alarms. Of course, if data is accumulated during mass production, the recess management range can be reconsidered each time. When the recess amount only varies within a sufficiently small range compared to the recess management range, throughput can be improved by reducing the number of measurement points monitored during process production.

[0082] In addition, during process development, by adjusting the process parameters of CMP105, it is possible to determine the parameters of CMP105 and whether there are defects in the wafer 112 after bonding. Figure 11 The process parameters of CMP105 and the presence or absence of defects in wafer 112 after bonding are described. Figure 11 The horizontal and vertical axes of the graph represent parameters 1 and 2 of the CMP105. White-painted points 1101 and 1102 indicate Cu pads determined to be defect-free based on the third measurement result 734, while black-painted points 1103 and 1104 indicate Cu pads determined to be defective. The shape of the plotted points indicates the type of pad. When making this comparison, although not shown, the first measurement result 714 of the bottom wafer 103 and the second measurement result 724 of the top wafer 102 can be correlated with the respective defect-free plotted points. That is, the conditions of the CMP105, the shape of the Cu pads before bonding, the surface condition, and the presence or absence of defects can be analyzed in conjunction with these factors, thus shortening the process development period. Furthermore, it is possible to... Figure 11 The chart is used to determine the allowable range of variation in process parameters, 1105.

[0083] Having divided the process until stable mass production of semiconductor devices into three phases—process development, ramp-up production, and stable production—examples of the effective application of comparison results during process development and stable production have been described. However, comparison results can also be effectively utilized during ramp-up production. For instance, during process development, the number of CMP105 and wafer bonding devices is small, but it is envisioned that multiple identical devices will be installed side-by-side during ramp-up production, with batch switching. In such cases, the depression management ranges 903 and 1003 can be used to determine the depression management range for each combination of devices, or by effectively applying the comparison results during device adjustments, the ramp-up period can be shortened, allowing for a faster transition to stable production. Furthermore, the comparison results can be used to derive the optimal combination of multiple devices.

[0084] Furthermore, the resolution in the X and Y directions may not be consistent across all measurement results. Therefore, not all data may correspond one-to-one. By including the type, serial number, shooting conditions, and evaluation conditions of the actual measuring devices used in the measurement results, it is possible to compare resolution and accuracy. In addition, by combining multiple measuring devices, for example, it is possible to analyze the correlation between the presence or absence of defects after joining and the shape before joining. By combining different devices, it is possible to comprehensively evaluate information known only to each other. By comparing not only the pre-joining inspection measurements with each other, but also the post-joining inspection measurements, it is possible to advance the understanding of phenomena that cannot be clearly understood by each measuring device alone, which helps to improve the CMP process 105, etc.

[0085] Figure 12 This is an example of a data sheet item. The process includes the name of the device being measured, the name of the process step, alignment mark information, chip layout information, imaging conditions, measurement points, evaluation conditions, and information about the chip being measured. The inspection results, in addition to containing the same information copied from the process, also include the actual wafer ID measured, the evaluation result, the acquired image, and the type and number of the measuring device used. These data allow for the aforementioned comparisons. Furthermore, regarding measurement results, there are various data sheet formats, such as averages, median values, standard deviations, etc., per wafer; per chip; per pad; and per specific coordinate and range. Refer to the appropriate data sheet for comparison.

[0086] Figure 13This is an example of a GUI (Graphical User Interface) displayed on the display unit 207 when the analysis program 205 is executed. The GUI 1300 consists of an analysis condition input unit 1310, a data display unit 1320, an analysis result unit 1330, and an execution button group 1340. Analysis conditions include, for example, device name, wafer ID, analysis chip, analysis object data, comparison item, output data, and data output path. When the image display button 1341 is operated, the analysis object data is displayed on the data display unit 1320. The wafer mapping 1321 is the distribution of the amount of depression obtained by analyzing the second measurement result 724 of the top wafer 102. Since the position measured in the bottom wafer 103 is reversed relative to the axis 1344, the wafer mapping 1322, which is the distribution of the amount of depression obtained by analyzing the first measurement result 714 of the bottom wafer 103, can measure the position symmetrical with respect to the axis 1344. Additionally, a defect-free wafer mapping 1323 obtained by analyzing the third measurement result 734 of the bonded wafer 112 is displayed. Here, if a chip is selected as indicated by arrow 1324, an arbitrary image 1326 of the bottom wafer 103, an arbitrary image 1325 of the top wafer 102, and an arbitrary image 1327 of the bonded wafer are displayed. When the next image display button 1329 is pressed, the next image within the chip is displayed. Furthermore, coordinate transformations of the results for the top wafer 102 are performed (see reference...). Figure 7 In step S725), when the invert / non-invert switch button 1328 is operated, the wafer mapping 1321 and image 1325 of the top wafer 102 are inverted relative to axes 1344 and 1345. Therefore, it is easy to identify the bonding chips and the pairing of bonding pads. The analysis result section 1330 displays the analysis results. When the analysis button 1342 is operated, the analysis result section 1330 displays the analysis chart 1331, which changes the color of the plotted points according to the presence or absence of defects based on the results of data C, based on the amount of depression in data A and data B. Furthermore, the analysis result section 1330 displays the depression management range 1332 (analysis result) calculated based on data A to data C. Afterwards, when the save button 1343 is operated, the analysis chart 1331 and the depression management range 1332 are saved in the data output path.

[0087] <Specific Examples of Measurement>

[0088] As a specific example of pre-bonding measurement 121, the measurement of Cu pad depression by SEM will be described. Here, an SEM apparatus with four detectors (N detector, E detector, S detector, and W detector) at the same height and spaced at 90-degree azimuth intervals is used. The N detector and S detector are arranged in opposite positions, and the E detector and W detector are arranged in opposite positions.

[0089] Figure 14 It is a schematic diagram of the object's structure, SEM image, structural cross-section view, and cross-sectional outline of the SEM image.

[0090] The top view of the periphery of the Cu pad is shown in diagram 1400, and the cross-sectional view at line 1404 is shown in diagram 1410. The periphery of the Cu pad is composed of an insulating portion 1401, a Cu pad 1402, and a blocking metal 1403 between the Cu pad and the insulating portion. Observing diagram 1410, it can be seen that the insulating portion 1401 is approximately flat, the Cu pad 1402 is approximately flat, but has uneven surfaces, and the blocking metal 1403 is inclined. The height difference between the insulating portion 1401 and the Cu pad 1402 is the amount of indentation to be measured. Additionally, as schematically shown in diagram 1410, the E detector 1414 and the W detector 1415 are positioned opposite each other. Furthermore, although not shown, the N detector and the S detector are also positioned opposite each other. Figure 14 The positions opposite each other in the front and back directions.

[0091] The E-image 1420 detected by E-detector 1414 shows a higher signal intensity (bright area 1421) when the normal of the blocking metal 1403 is oriented towards the E-detector, and conversely, a lower signal intensity (dark area 1422) when the normal is oriented in the opposite direction to the E-detector. At this point, when observing the cross-sectional profile 1430 at line 1424, a bright peak 1431 and a dark peak 1432 are observed.

[0092] The W image 1440 detected by the W detector 1415 has a higher signal intensity than its surroundings when the normal of the blocking metal 1403 is oriented towards the W detector, thus appearing whiter (bright area 1442). Conversely, when the normal is oriented in the opposite direction to the W detector, the signal intensity decreases, thus appearing darker (dark area 1441). At this time, when observing the cross-sectional profile 1450 at line 1444, bright peak 1452 and dark peak 1451 are observed.

[0093] When the difference between the images E and W is taken and normalized to a range of 0-255 for display, the EW image 1460 is obtained. The brightness of the EW image 1460 varies depending on the tilt direction of the blocking metal 1403. The bright region 1461 is the region where the normal of the blocking metal 1403 faces the direction of the E detector 1414, and the dark region 1462 is the region where the normal of the blocking metal 1403 faces the direction of the W detector 1415. The intermediate color region 1463 represents the region where there is no difference in the detection signals of the E detector 1414 and the W detector 1415, i.e., the region of vertical information towards the cross-sectional schematic diagram 1410. Schematic, it is not the three stages of bright, dark, and intermediate, but rather a continuous value including values ​​between them, as in the cross-sectional contour 1470 at line 1464.

[0094] Figure 15 This is a schematic diagram used to calculate the feature quantities of the depression. The EW feature quantity extraction schematic diagram 1500 is... Figure 14 An image of the EW region 1460 is overlaid with images of the bright area extraction region 1501 and the dark area extraction region 1502. The calculated values ​​(e.g., difference) of the statistics (e.g., average) of the bright area extraction region 1501 and the statistics (e.g., average) of the dark area extraction region 1502 are used as EW depression indices to determine the size relationship of the depressions. A schematic diagram of NS feature extraction 1510 is an image of the NS image, which is the difference image between the N detector and the S detector, overlaid with images of the bright area extraction region 1514 and the dark area extraction region 1513, containing the dark area 1511 and bright area 1512. The calculated values ​​(e.g., difference) of the statistics (e.g., average) of the bright area extraction region 1514 and the statistics (e.g., average) of the dark area extraction region 1513 are used as NS depression indices to determine the size relationship of the depressions. By using the calculated values ​​(e.g., average) of the WE depression index and the NS depression index as depression indices, a depression amount can be calculated for a Cu pad. Therefore, in SEM, the amount of depression on each Cu pad can be calculated. Furthermore, data at different resolutions can be obtained, including the average depression amount of Cu pads in the surrounding area, the average for each pad type, the average for pads within the chip, the average for pads across the entire wafer, and the distribution (trend) of the average depression within the chip. Comparison with the resolution of post-bonding measurements 122 is possible.

[0095] exist Figure 14 as well as Figure 15The example described uses four detectors arranged perpendicularly to the line of the Cu pad boundary. However, when the effective arrangement of the detectors in the four directions is not perpendicularly parallel, bright areas 1421 and dark areas 1422 are also generated at the upper and lower boundaries of the Cu pad, for example. The same applies when the Cu pad 1602 is circular instead of quadrilateral. In such cases, this can be addressed by deforming the bright area extraction region 1501, the dark area extraction region 1502, the bright area extraction region 1514, and the dark area extraction region 1513. Furthermore, these extraction regions do not need to encompass all bright and dark areas; a portion is sufficient.

[0096] In this embodiment, the example of calculating the average brightness as a feature quantity is illustrated, but the value can also be obtained by integrating the difference image (EW, NS image) of the opposing detector from a reference point toward the measurement point at the desired height.

[0097] <Explanation of variations>

[0098] In Example 1 described above, an example was shown where the results were obtained by performing a single measurement on the bottom wafer 103, the top wafer 102, and the bonded wafer 112. However, the number of measurements for each wafer is arbitrary, allowing for multiple analyses (comparisons) of the measurement results. By performing multiple analyses (comparisons), it becomes easier to investigate the causes of void defects and improve the CMP process, for example. Of course, the same wafer can be measured using different devices, or devices with different measurement accuracies, speeds, and resolutions can be combined. For example, a wide distribution of the amount of depression can be obtained through optical measurement, which enables low-resolution, high-speed measurement, and the amount of depression in local pads can be evaluated using high-resolution SEM or AFM. This can be applied to the analysis of the causes of void defects or the characteristics of CMP.

[0099] If the shape of each pad is evaluated and its mean and standard deviation are compared with electrical characteristics, the relationship between electrical characteristics and shape can also be derived. Comparisons with statistical measures such as mean and standard deviation are meaningful in combinations of devices with different resolutions.

[0100] Furthermore, while wafer bonding has been described as an example, this embodiment can also be used in the process known as D2W (Dai to wafer), where Dai (chips) diced and removed from the wafer are bonded to the wafer. Additionally, although not shown, the first measurement result 714 of the bottom wafer 103 before bonding can be compared with the second measurement result 724 of the top wafer 102 before bonding to derive wafer pairs with fewer bonding defects, dynamically changing the wafer pairs to be bonded. In the case of D2W, replacement can also be performed on a Dai-by-Dai basis.

[0101] Example 2

[0102] Example 2 demonstrates that by correcting the process based on the measurement results, measurement time can be shortened and detailed analysis can be performed.

[0103] Figure 16 This is an example of the action flow when a measurement procedure corrects the process according to the measurement results before performing a measurement. For example, the second measurement procedure 720 corrects the second process 609 read from DB204 based on the first measurement result 714 (step S1601), and performs the measurement based on the corrected second process. Furthermore, the third measurement procedure 730 corrects the third process 735 read from DB204 based on the first measurement result 714 and the second measurement result 724 (step S1602), and performs the measurement based on the corrected third process.

[0104] For example, during defect inspection in mass production, if foreign matter is present in the first measurement result 714 of the bottom wafer 103, the measurement of the corresponding measurement points of the top wafer 102 and the bonded wafer 112 can be skipped through process corrections in steps S1601 and S1602. In this case, by not measuring the location of randomly generated defects caused by foreign matter, the measurement time can be shortened.

[0105] For example, during process development, the bottom wafer 103 can be measured at many measurement points. If a shape defect is identified, the second and third processes can be corrected so that the measurement of that defect is concentrated on the top wafer 102 and the bonded wafer 112. In this case, a correspondence between shape defects and the presence or absence of defects can be established, enabling process development in a short time.

[0106] Explanation of reference numerals in the attached figures

[0107] 102…Top wafer, 103…Bottom wafer, 106, 404, 414, 1402…Cu pads, 110…Y-axis, 112…Wafer after bonding, 107…Recess, 207…Display section, 401, 411…Chip magnification, 402, 412…Pattern, 403, 413…Layout magnification, 405, 415…Virtual pattern, 407, 417…Alignment marks, 408, 418…Coordinate system, 409, 419…Reversal axis, 421, 422, 431, 432…Measurement points, 1300…GUI, 1400…Top view, 1410…Cross-sectional view, 1401…Insulation section, 1403…Barrier metal, 1404, 1424, 1444 …line, 1414…E detector, 1415…W detector, 1420…E image, 1421, 1442, 1461, 1512…bright area, 1422, 1441, 1462, 1511…dark area, 1430, 1450, 1470…profile, 1431, 1452…bright peak, 1432, 1451…dark peak, 1440…W image, 1460…EW image, 1500…EW feature extraction schematic diagram, 1501, 1514…bright area extraction region, 1502, 1513…dark area extraction region, 1510…NS feature extraction schematic diagram, 3001…SEM-type measuring device, 3100…electron microscope, 3120…control unit.

Claims

1. A measurement system for measuring semiconductor devices, characterized in that, The measurement system includes a pre-bonding measurement device for measuring the first and second wafers before bonding. When the first wafer is set as the first measurement point, the pre-bonding measurement device will set the second measurement point, which is connected to the first measurement point after bonding, as the measurement point of the second wafer.

2. The measurement system according to claim 1, characterized in that, After measuring the first measurement point on the first wafer, the pre-bonding measurement device measures the second measurement point on the second wafer.

3. The measurement system according to claim 1, characterized in that, The measurement system also features: Post-bonding measurement apparatus, which measures the first wafer and the second wafer after bonding; and The analysis device compares the combined measurement results with the measurement results of the first measurement point and the measurement results of the second measurement point.

4. The measurement system according to claim 3, characterized in that, The measurement result compared by the analytical device is any one of the following: the measurement value of the pad itself, the average of the measurement values ​​of some or all the pads within the chip, or the tendency of the measurement values ​​of the pads within the wafer.

5. The measurement system according to claim 3, characterized in that, The analysis device determines the permitted management range of the measurement results at the first measurement point and the permitted management range of the measurement results at the second measurement point by comparison.

6. The measurement system according to claim 3, characterized in that, The analysis device displays the measurement result of the first measurement point or the image corresponding to the measurement result, or both, or the measurement result of the second measurement point or the image corresponding to the measurement result, or both, in reverse.

7. The measurement system according to claim 3, characterized in that, The resolution of the measuring device after joining is lower than that of the measuring device before joining.

8. The measurement system according to claim 7, characterized in that, The post-joint measurement device is a device that uses SEM or AFM. The pre-joining measurement device is a device that uses any one of ultrasound, X-rays, or infrared radiation.

9. The measurement system according to claim 1, characterized in that, The pre-joining measuring device has: Computers, which are manufactured to specify measurement conditions; and The measurement unit performs measurements based on the aforementioned process. The computer reads in the first process related to the first wafer and obtains the first measurement point information, which is the measurement object of the first wafer. The computer reads the layout data of the first wafer and / or the second wafer, and obtains coordinate system information from the layout data. The computer uses the coordinate system information to reverse the first measurement point information, and uses the reversed first measurement point information to fabricate a second process related to the second wafer.

10. The measurement system according to claim 1, characterized in that, The pre-joining measuring device has: Computers, which are manufactured to specify measurement conditions; and The measurement unit performs measurements based on the aforementioned process. Based on the layout data of the first wafer and the measurement conditions of the first wafer input by the user, the computer creates a first process that includes information on the first measurement points that are the measurement objects of the first wafer. The computer uses coordinate system information obtained from the layout data to reverse the first measurement point information, and based on the reversed first measurement point information, manufactures a second process that includes second measurement point information as the measurement object of the second wafer.

11. The measurement system according to claim 1, characterized in that, The pre-joining measuring device has: Computers, which are manufactured to specify measurement conditions; and The measurement unit performs measurements based on the aforementioned process. The computer fabricates a first process that includes first measurement point information as the measurement object of the first wafer, and a second process that includes second measurement point information as the measurement object of the second wafer. After performing measurements on the first wafer based on the first process, the measurement unit performs measurements on the second wafer based on the second process corrected using the measurement results from the first measurement point.

12. A measurement method for measuring a first wafer and a second wafer before bonding, characterized in that, The measurement method includes: The step of measuring the first wafer by taking the first measurement point as the measurement object; and The step of measuring the second wafer by taking the second measurement point, which is connected to the first measurement point after bonding, as the measurement object.

13. A method for manufacturing a measurement process, characterized in that, The measurement process manufacturing method includes: The step of reading in the first process related to the first wafer and obtaining the first measurement point information as the measurement object of the first wafer; The steps of reading the layout data of the first wafer and / or the second wafer and obtaining coordinate system information from the layout data; as well as The coordinate system information is used to reverse the first measurement point information, and based on the reversed first measurement point information, a second process related to the second wafer is fabricated.

14. A method for manufacturing a measurement process, characterized in that, The measurement process manufacturing method includes: Based on the layout data of the first wafer and the measurement conditions of the first wafer input by the user, a first process step is to manufacture a first process that includes information on the first measurement point as the measurement object of the first wafer. as well as Using coordinate system information obtained from the layout data to invert the first measurement point information, and based on the inverted first measurement point information, a second process step is performed to manufacture a second process that includes second measurement point information as the measurement object of the second wafer.

Images