TWI755248B - Method for detecting a placement status of wafer in wafer box and detection system of the method - Google Patents

Abstract

A method for detecting a placement status of wafer in wafer box. The method includes providing a wafer box, the wafer box includes a plurality of side-by-side storage units, each storage unit is used for storing a wafer; emitting light towards the wafer box, so that the light passes through the gap between the storage unit and the wafer; collecting an optical signal formed by the light passing through the gap to obtain a image; features information are extracted from the image; the feature informations are compared with a feature informations of a preset model, and the consistency between the image and the preset model is judged according to the comparison results, so as to determine whether the wafer placement status in the wafer box is qualified. The detection method of the invention can automatically and quickly detect the placement state of the wafer in the wafer box. A detection system for performing the method is also disclosed.

Description

Detection method and detection system of wafer placement status in wafer cassette

The invention relates to the technical field of wafer detection equipment, and in particular, to a detection method and detection system for the placement state of wafers in a wafer cassette.

In the wafer packaging and testing plant, the traditional method of wafer status inspection is that the wafer is placed in the wafer box by a robotic arm, and after the cover is placed, the inspector needs to visually check whether the number and position of the wafers placed in the wafer box conform to the requirements. It is stipulated that at the same time, whether the wafer is broken or stacked is detected.

However, the above-mentioned manual detection method has low detection efficiency, and is prone to missed detection and false detection; especially for some translucent wafer boxes, the light intensity needs to be increased to be able to clearly see the wafers in the wafer box. It may cause visual damage to human eyes when it is placed under strong light for a long time; in addition, manual inspection cannot be achieved for opaque wafer cassettes.

In view of this, it is necessary to provide a detection method that can quickly and effectively detect whether the placement state of the wafers in the wafer cassette meets the requirements.

In addition, the present invention also provides a detection system for performing the above-mentioned method for detecting the placement state of wafers in a wafer cassette.

The present invention provides a method for detecting the placement state of wafers in a wafer cassette, comprising the following steps:

providing a wafer cassette, the wafer cassette includes a plurality of storage units arranged side by side, each of the storage units is used for storing wafers;

emitting light toward the wafer cassette so that the light passes through the gap between the storage unit and the wafer;

collecting the light signal formed after the light passes through the gap and obtaining an image;

extracting feature information from the image; and

Compare the feature information with the feature information of the preset model, and determine whether the image is consistent with the preset model according to the comparison result, and then determine whether the placement status of the wafers in the wafer cassette is not qualified.

Further, the characteristic information includes the number of wafers in each memory cell.

Further, the preset model is obtained by training a computer vision model or an AI deep learning model.

The present invention also provides a system for detecting the placement state of wafers in a wafer cassette, comprising:

a wafer cassette, comprising a plurality of side-by-side storage units, each of which is used to store wafers;

a light source module, disposed at one end of the storage unit, the light source module is used for emitting light toward the wafer cassette, so that the light passes through the gap between the storage unit and the wafer;

an image acquisition module, disposed on the side of the storage unit away from the light source module, the image acquisition module is used to collect the optical signal formed after the light passes through the gap and obtain an image;

an identification module for receiving the image and extracting feature information from the image;

a judgment module, configured to compare the feature information with the feature information of the preset model, and judge whether the image is consistent with the preset model according to the comparison result, and then determine the wafers in the wafer cassette Whether the placement status is qualified.

Further, the characteristic information includes the number of wafers in each memory cell.

Further, the preset model is obtained by training a computer vision model or an AI deep learning model.

Further, the wafer cassette includes a casing and a plurality of parallel partitions arranged in the casing, and a storage unit is formed between two adjacent partitions.

Further, the casing is made of a partially transparent material or an opaque material, and the box cover is made of a partially transparent material or an opaque material.

Further, the image acquisition module includes at least one image acquisition device.

Further, the light source module includes at least one light source, and the light emitted by all the light sources covers all the storage units.

Compared with the prior art, the method for detecting the placement state of the wafers in the wafer cassette provided by the present invention can realize the automatic detection of the placement state of the wafers, the detection is simple, the operation is convenient, and the detection process is convenient and fast; It can improve product yield, avoid damage to human eyes caused by light, and reduce labor costs; especially for opaque wafer boxes, it can achieve effective detection.

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or intervening elements may also be present. When an element is referred to as being "connected" to another element, it can be directly connected to the other element or intervening elements may also be present. When an element is referred to as being "disposed on" another element, it can be directly disposed on the other element or intervening elements may also be present. The terms "vertical," "horizontal," "left," "right," and similar expressions are used herein for illustrative purposes only.

The system implementation described below is only illustrative, and the division of the modules or circuits is only a logical function division, and other division methods may be used in actual implementation. Furthermore, it is clear that the word "comprising" does not exclude other units or steps and the singular does not exclude the plural. Multiple units or means stated in the system claim may also be implemented by the same unit or means through software or hardware. The terms first, second, etc. are used to denote names and do not denote any particular order.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terms used herein in the description of the present invention are for the purpose of describing specific embodiments only, and are not intended to limit the present invention. As used herein, the term "and/or" includes any and all combinations of one or more of the associated listed items.

Referring to FIG. 1, a method for detecting the placement state of wafers in a wafer cassette provided by an embodiment of the present invention specifically includes the following steps: Step S1, please refer to FIG. 2 and FIG. 3 together, providing a wafer cassette 1. The wafer cassette 1 includes a plurality of storage units 11 arranged side by side, and each of the storage units 11 is used for storing the wafers 2 . The wafer cassette 1 includes an outer casing 12 , a plurality of parallel partitions 13 arranged in the outer casing 12 , and a box cover 14 arranged at one end of the outer casing 12 . A storage unit 11 is formed between two adjacent partitions 13 . The plane where the box cover 14 is located is perpendicular to the partition plate 13 , an opening 15 is provided on the side of the storage unit 11 facing the box cover 14 , and the wafer 2 is put into the storage unit through the opening 15 within 11.

In one embodiment, one side of the box cover 14 is rotatably disposed on the side wall of the casing 12 , and the box cover 14 and the casing 12 are locked and unlocked by the locking device 16 .

Referring to FIG. 4 , in one embodiment, the wafer cassette 1 further includes a tenon 17 disposed on the inner side of the box cover 14 in a one-to-one correspondence with the partition 13 , between two adjacent tenon 17 The distance is approximately equal to the thickness of the wafer 2, and the wafer 2 is clamped between two adjacent tenons 17, and the tenons 17 are used to attach the wafers in the storage unit 11. 2 is fixed to prevent the wafer 2 from shaking in the storage unit 11 .

In one embodiment, the box cover 14 is disposed on the top of the casing 12 , and the opening 15 is disposed upward, which facilitates the placement of the wafers 2 and does not move the wafer box 1 . When the phenomenon that the wafer 2 is unexpectedly pulled out occurs.

In one embodiment, one of the wafers 2 is placed in each of the storage units 11 , which can prevent the wafers 2 from being stacked together and cause damage, and at the same time facilitate the subsequent inspection of the wafers 2 . However, when the manipulator performs the swinging operation, the situation of lamination will inevitably occur. Therefore, the detection method of the present invention can also accurately determine whether there is lamination in the same storage unit 11 . And there is a gap b between the two partitions 13 constituting the storage unit 11 and the wafer 2 , so that light can pass through the wafer cassette 1 .

In one embodiment, the housing 12 can be made of transparent material, translucent material or opaque material, and the box cover 14 can be made of transparent material, translucent material or opaque material. For the opaque casing 12 and the box cover 14, light of a special wavelength band (such as infrared rays) can be selected to irradiate the wafer cassette 1, so that the wafer 2 and the The storage unit 11 is capable of imaging. By adopting the detection method, automatic and efficient detection can be realized, manual false detection and missed detection can be avoided, product yield can be improved, damage to human eyes caused by light can be avoided, and labor costs can be reduced at the same time. Especially for the opaque casing 12 and the box cover 14, this detection method makes up for the defect that manual detection cannot be seen because the naked eye cannot see.

The material of the casing 12 may be plastic or glass, and the material of the box cover 14 may be plastic or glass. In one embodiment, the casing 12 and the box cover 14 are made of plastic, and the tenon 17 is made of plastic or other elastic materials, and the tenon 17 will deform when squeezed.

Step S2, emit light a towards the wafer box 1, so that the light a passes through the gap b between the storage unit 11 and the wafer 2, that is, the partition 13 and the The gap b between the wafers 2.

In one embodiment, a light source module 3 is used to emit the light a, and the light source module 3 is located at the bottom of the wafer cassette 1. At this time, the wafer 2 is arranged perpendicular to the light source module 3, and the light source The light ray a emitted by the module 3 is substantially parallel to the plane where the wafer 2 is located. The wafer box 1 is placed in a suitable position of the light source module 3, the light source module 3 is turned on, the light a is blocked by the wafer 2 after encountering the wafer 2, and the other parts of the The light a passes through the gap b between the partition 13 and the wafer 2, so that the wafer 2 can be imaged, and at the same time adjust the light intensity and color temperature of the light source module 3 to make the wafer box 1 The wafer 2 inside can be clearly imaged.

In one embodiment, the light source module 3 adopts LED light bars, and a plurality of the LED light bars are evenly arranged at the bottom of the wafer box 1, which can illuminate the entire wafer box 1 and avoid dark spots. Area.

Step S3, collect the light signal formed after the light a passes through the gap b, and obtain an image.

In one embodiment, the image is captured by an image capture module 4 , and the image capture module 4 is disposed on the side of the wafer cassette 1 away from the light source module 3 .

The image acquisition module 4 includes at least one image acquisition device, and the image of the wafer cassette 1 is acquired by arranging the at least one image acquisition device along the transmission direction of the light a. Specifically, in this embodiment, in order to obtain the images of the wafer cassette 1 at different angles, a plurality of the image collectors may be arranged at different positions on the top of the wafer cassette 1, so as to facilitate obtaining the images of the wafer cassette 1. images of the storage unit 11 and the wafer 2 at different positions in the wafer cassette 1, and images of the wafer 2 at the same position from different angles.

The image collector includes, but is not limited to, a general optical camera or a CCD camera, as long as the image captured by the camera can clearly show the light and shadow images of the storage unit 11 and the wafer 2 inside the wafer cassette 1. Can.

In one embodiment, the image acquisition device is a CCD camera, and by disposing a plurality of CCD cameras at different positions on the top of the wafer cassette 1 , the images at different positions in the wafer cassette 1 can be collected. Clear images of the storage unit 11 and the wafer 2 facilitate subsequent feature extraction and comparison and judgment.

In step S4, feature information extraction is performed on the image.

In one embodiment, the feature information extraction process is performed by an identification module 5 , and the identification module 5 is connected to the image acquisition module 4 in a signal connection for receiving the information transmitted by the image acquisition module 4 . image.

The recognition module 5 receives the images, arranges and combines the images at different positions and angles, and then extracts the corresponding feature information of each of the images.

In one embodiment, the characteristic information includes the number of wafers 2 in each memory cell 11 . The number of wafers 2 in each storage unit 11 can be reflected by the image, which refers to the number of wafers 2 placed in each storage unit 11. In the image Including the clear image of the storage unit 11, the clear image of the tenon 17 and the clear image of the wafer 2, by calculating the two adjacent tenons of each of the storage units 11 The characteristic information of the number of wafers 2 in each storage unit 11 can be obtained from the number of the wafers 2 stuck between 17 .

Step S5, compare the feature information with the feature information of the preset model, and determine whether the image is consistent with the preset model according to the comparison result, and then determine the swing of the wafer in the wafer cassette. Whether the release status is qualified.

In one embodiment, the above-mentioned determination process is performed by a determination module 6, which is connected to the identification module 5 in a signal connection for receiving the characteristic information transmitted by the identification module 5, and converts the characteristic information to the identification module 5. The information is compared with the feature information of the preset model, and according to the comparison result, it is judged whether the image is consistent with the preset model, and then it is determined whether the placement state of the wafers 2 in the wafer cassette 1 is qualified.

Regarding the characteristic information of the number of wafers 2 in each storage unit 11 , when the number of the wafers 2 between two adjacent tenons 17 corresponding to a certain storage unit 11 is 0 When the storage unit 11 is missing or the wafer 2 is damaged; when the number of the wafers 2 between the two adjacent tenons 17 corresponding to a certain storage unit 11 is greater than 1 , then it is judged that the storage unit 11 is laminated. It can be understood that when the wafer 2 in a storage unit 11 is laminated, since the tenon 17 is made of plastic material or other elastic material, the tenon 17 will It is squeezed and deformed. When the image of the tenon 17 on the image is squeezed and deformed, it can be judged that the phenomenon of lamination occurs in the storage unit 11 at this time; Only when the number of the wafers 2 between the two adjacent tenons 17 corresponding to the unit 11 is 1, the characteristic information of the number of the wafers 2 in each storage unit 11 can be determined to be the same as the preset number. The corresponding feature information of the model is consistent.

Therefore, by judging the above feature information, finally only when the feature information conforms to the corresponding feature information of the preset model can it be determined that the placement state of the wafers 2 in the wafer cassette 1 is qualified of.

Specifically, the preset model is obtained by training a computer vision model or an AI deep learning model.

Firstly, the preset model is established, a large number of the collected images are arranged and combined and feature extraction is performed, and the feature information that meets the control requirements is obtained to establish a model.

Then, the wafer cassette 1 to be inspected is inspected under the same conditions as the inspection environment when the preset model is established.

The detection result is compared with the preset model, and the extracted feature information is compared with the feature information corresponding to the preset model by the judgment module 6 . The placement state of the wafers 2 in the wafer cassette 1 that conform to all the corresponding feature information of the preset model is qualified, and for the corresponding feature information that does not conform to the preset model, the processing system will automatically label on the above image.

Subsequently, the placement state of the wafers 2 in the wafer cassette 1 can be improved according to the labeling on the image, for example, there is lamination in the storage unit 11, and the wafers 2 are damaged. The corresponding wafers 2 can be rejected, and if there is a phenomenon of missing wafers in the storage unit 11, patching is performed. With the above detection method, the unit point to be improved can be accurately obtained, the detection efficiency is improved and the product yield is improved.

Please refer to FIG. 2 and FIG. 3 , an embodiment of the present invention provides a detection system 100 for the state of placement of wafers in a wafer cassette, including a light source module 3 , and an image capture device disposed on the light-emitting side of the light source module 3 . Module 4, an identification module 5 signally connected to the image acquisition module 4, and a judgment module 6 signally connected to the identification module 5, the identification module 5 and the judgment module 6 are integrated into a control host 7.

When the wafer cassette 1 containing the wafers 2 needs to be detected, the wafer cassette 1 only needs to be placed on the light-emitting side of the light source module 3, and the light source module 3 faces the wafer cassette. 1. The light a is emitted, so that the light a is substantially parallel to the surface of the wafer 2, and the light a can pass through the gap b between the spacer 13 and the wafer 2.

The light source module 3 includes at least one light source 31, and by arranging the light source 31 at the corresponding position of the storage unit 11, the light a from all the light sources 31 can cover all the storage units 11, and at the same time The light a is parallel to the surface of the wafer 2, so that all the storage units 11 and the wafer 2 in the wafer cassette 1 can be illuminated.

The light source 31 can clearly image the storage unit 11 and the wafer 2 inside the opaque wafer cassette 1 by adjusting the light intensity and color temperature.

The light source 31 includes, but is not limited to, a light board and a light bar, as long as it can illuminate the opaque wafer cassette 1 with any suitable wavelength band, color temperature, and light intensity that can clearly present the internal image.

In one embodiment, the light source 31 is an LED light bar or an LED light board.

In one embodiment, the LED light bar or the LED light board is installed on the bottom of the wafer cassette 1, and the light a from the light source 31 is emitted upward in a vertical direction and irradiates the wafer The light a on 2 is blocked by the wafer 2, and the remaining light a can pass through the gap b between the partition 13 and the wafer 2, so that each wafer 2 can be form a clear image.

The image acquisition module 4 is configured to acquire the light signal formed after the light passes through the gap b to obtain the image. The image includes the image information of the storage unit 11 , the tenon 17 and the wafer 2 , and the distribution state of the storage unit 11 and the tenon 17 can be seen from the image. , whether there is a wafer 2 in the storage unit 11, and whether there is only one wafer 2 between the two adjacent tenons 17, the image can also be used to see the distribution state of the wafer 2. Whether the circle 2 is stuck between two adjacent tenons 17 corresponding to the storage unit 11 , and whether there is a lamination phenomenon in the same storage unit 11 , that is, whether there are multiple adjacent tenons 17 . A wafer 2, whether the tenon 17 is squeezed and deformed, and whether the wafer 2 is damaged, etc.

The image acquisition module 4 includes at least one image acquisition device, and the image acquisition device is arranged along the transmission direction of the light a to acquire the image of the wafer cassette 1 .

There may be multiple image collectors (the image acquisition module 4 ). By setting the image collectors at the corresponding positions of the wafer cassette 1 , the storage unit 11 and all the image collectors can be photographed. Images of the wafer 2 from different angles.

In one embodiment, the image acquisition device (the image acquisition module 4 ) includes, but is not limited to, a general optical camera or a CCD camera, as long as the image captured by the camera can clearly show the storage inside the wafer cassette 1 The light and shadow images of the unit 11 and the wafer 2 can be used.

The identification module 5 is configured to receive the image transmitted by the image acquisition module 4, and extract feature information from the image.

The image acquisition module 4 collects many images of the storage unit 11 and the wafer 2 inside the wafer cassette 1, arranges and combines the images at different positions and angles, and then extracts each image. Corresponding feature information of the image.

In one embodiment, the feature information includes the number of wafers 2 in each storage unit 11 , and the specific meaning is the same as above.

The judging module 6 is used to compare the feature information with the feature information of the preset model, and determine whether the image is consistent with the preset model according to the comparison result, and then determine the wafer cassette 1 Whether the placement state of the wafers 2 inside is qualified.

The preset model is obtained by training a computer vision model or an AI deep learning model in advance. Arrange and combine a large number of the collected images and extract features, and obtain the feature information that meets the control requirements to build a model. In the subsequent inspection process, only the wafer cassette 1 to be inspected is placed in the desired location. On the light source module 3, an image is collected by the image acquisition module 4, and the judgment module 6 can use the judgment module 6 to determine the feature information contained in the image corresponding to the preset model. feature information for comparison. The placement state of the wafers 2 in the wafer cassette 1 that conform to all the corresponding feature information of the preset model is qualified, and for the corresponding feature information that does not conform to the preset model, the processing system will automatically The above image is marked, and subsequent improvements can be made according to the corresponding marking on the image. For example, if there is stacking phenomenon in the storage unit 11, or the wafer 2 is damaged, the corresponding wafer can be removed. 2. If there is a phenomenon of missing pieces in the storage unit 11, make a patch. With the above detection method, the unit point to be improved can be accurately obtained, the detection efficiency is improved and the product yield is improved.

In one embodiment, the identification module 5 and the judgment module 6 are integrated in a control host 7 , by performing image display, feature extraction and judgment process in the control host 7 , and displaying the judgment result.

By means of the detection system 100 for the state of placement of the wafers in the wafer cassette, fast, effective and automatic state detection of the wafer 2 can be realized, problems such as manual error detection, missed detection, and inability to be detected can be avoided, and the product yield can be improved. Avoid damage to human eyes caused by light, and reduce labor costs at the same time.

Compared with the prior art, the method for detecting the placed state of the wafers in the wafer cassette provided by the present invention can realize the automatic detection of the placed state of the wafers, the detection is simple, the operation is convenient, and the detection process is convenient and fast; The phenomenon of missed inspection and manual inability to detect, etc., improves product yield; avoids damage to human eyes caused by light, and reduces labor costs; especially for opaque wafer boxes, it can achieve effective inspection.

In addition, for those of ordinary skill in the art, various other corresponding changes and deformations can be made according to the technical concept of the present invention, and all these changes and deformations should belong to the protection scope of the claims of the present invention.

100: Inspection system for the placement of wafers in a wafer cassette S1~S5: Steps 1: Wafer box 11: Storage unit 12: Shell 13: Clapboard 14: Box cover 15: Opening 16: Locking device 17: Tenon 2: Wafer 3: Light source module 31: Light source 4: Image acquisition module 5: Identify the module 6: Judgment module 7: Control the host a: light b: gap

FIG. 1 is a flowchart of a wafer state detection method provided by an embodiment of the present invention.

FIG. 2 is a functional block diagram of a wafer state inspection system according to an embodiment of the present invention.

FIG. 3 is a schematic structural diagram of a wafer state detection system according to an embodiment of the present invention.

FIG. 4 is a schematic structural diagram of a wafer state detection system according to another embodiment of the present invention.

S1~S5: Steps

Claims (10)

A method for detecting the state of placement of wafers in a wafer cassette, comprising the following steps: providing a wafer cassette, the wafer cassette includes a plurality of storage units arranged side by side, each of the storage units is used for storing wafers; emitting light toward the wafer cassette so that the light passes through the gap between the storage unit and the wafer; collecting the light signal formed after the light passes through the gap and obtaining an image; extracting feature information from the image; and Compare the feature information with the feature information of the preset model, and determine whether the image is consistent with the preset model according to the comparison result, and then determine whether the placement status of the wafers in the wafer cassette is not qualified. The method for detecting the placement state of wafers in a wafer cassette according to claim 1, wherein the characteristic information includes the number of wafers in each storage unit. The method for detecting the placement state of wafers in a wafer cassette according to claim 1, wherein the preset model is obtained by training a computer vision model or an AI deep learning model. A system for detecting the placement state of wafers in a wafer cassette, comprising: a wafer cassette, comprising a plurality of side-by-side storage units, each of which is used to store wafers; a light source module, disposed at one end of the storage unit, the light source module is used for emitting light toward the wafer cassette, so that the light passes through the gap between the storage unit and the wafer; an image acquisition module, disposed on the side of the storage unit away from the light source module, the image acquisition module is used to collect the optical signal formed after the light passes through the gap and obtain an image; an identification module for receiving the image and extracting feature information from the image; a judgment module, configured to compare the feature information with the feature information of the preset model, and judge whether the image is consistent with the preset model according to the comparison result, and then determine the wafers in the wafer cassette Whether the placement status is qualified. The system for detecting the placement state of wafers in a wafer cassette according to claim 4, wherein the characteristic information includes the number of wafers in each storage unit. The system for detecting the placement state of wafers in a wafer cassette according to claim 5, wherein the preset model is obtained by training a computer vision model or an AI deep learning model. The system for detecting the state of placement of wafers in a wafer cassette according to claim 4, wherein the wafer cassette comprises a casing, a plurality of parallel partitions disposed in the casing, and one end of the casing that is arranged with the casing. The storage unit is formed between two adjacent partitions. The system for detecting the placement state of wafers in a wafer cassette according to claim 7, wherein the casing is made of a partially transparent material or an opaque material, and the box cover is made of a partially transparent material or an opaque material. The system for detecting the placement state of wafers in a wafer cassette according to claim 4, wherein the image acquisition module includes at least one image acquisition device. The system for detecting the placement state of wafers in a wafer cassette according to claim 4, wherein the light source module includes at least one light source, and the light rays emitted by all the light sources cover all the storage units.

Images