SCREEN Holdings

To provide a manufacturing execution control system capable of shortening a processing lead time of a priority lot in a manufacturing line. In a manufacturing execution control system 1000, a virtualization processing apparatus 110 collects processing information by using a predetermined substrate cleaning apparatus group i as a hub in a wafer processing sequence, and, from an existing MES 100 side, the substrate cleaning apparatus group i is virtually identified and controlled as a single manufacturing apparatus using a single piece of identification information. Information for specifying a priority lot and information indicating priority are shared among group control apparatuses 310.i to 310.i+2 that control the substrate cleaning apparatus group i, and preparation for processing of the priority lot is notified in advance from a group control apparatus of a preceding process to a group control apparatus of a subsequent process. In the substrate cleaning apparatus group i of the subsequent process, preparation for processing the priority lot is performed in advance, thereby enabling the priority lot to be processed by overtaking other normal lots.

A defect inspection apparatus (1) includes a heater unit (9) to emit infrared rays (F2) to a substrate (W) held by a substrate holding unit (8) and heat the substrate (W), an infrared camera (10) to face a first principal surface (Wa) of the substrate (W) and image an imaging region (24) on the first principal surface (Wa); a heat insulating member (31) to absorb the infrared rays (F2) emitted by the heater unit (9), and a heat insulating member moving unit (50) to move the heat insulating member (31) between a heat insulation position (P10) where the heat insulating member (31) and the imaging region (24) overlap each other when viewed from a normal direction (Z) of the first principal surface (Wa) and a retracted position (P2) where the heat insulating member (31) and the imaging region (24) do not overlap each other when viewed from the normal direction (Z). The substrate holding unit (8) may be positioned between the heater unit (9) and the infrared camera (10) in the normal direction (z).

The cover member 51 performs the capping . In this capping, the cover member 51 covers the nozzles 421 from below by causing the ink receiving space 527 open upward to face the nozzle plane 42 from below while being stopped with respect to the discharge heads 44. The cover member 51 includes the liquid retaining member 541 for retaining the cleaning liquid, and the liquid retaining member 541 contacts the non-nozzle plane 43 during the execution of the capping. The ink can be removed from the non-nozzle plane 43 by diffusing the ink adhering to the non-nozzle plane 43 into the cleaning liquid. Since the cover member 51 is stopped with respect to the discharge heads 4 during the execution of the capping, the liquid retaining member 541 continues to continuously contact the non-nozzle plane 43. The ink can be effectively removed from the non-nozzle plane 43.

A dataset for machine learning is initially prepared. Next, labels are assigned to respective pieces of data included in the dataset. Subsequently, a problem related to the labels of the dataset is alleviated. Then, machine learning is performed using the improved dataset, whereby a learning model is created. Thereafter, a difference in tendency between the dataset used for machine learning and operational data to which the learning model is to be applied during actual operation is judged. The dataset is improved or created again if it is judged that there is a difference in tendency between the dataset and the operational data. In this manner, the quality of the dataset for machine learning is improved through multiple steps. Thus, the learning model is created while the quality of the dataset is maintained at a high level.

A method of creating a basic model for multiple machine learning models corresponding to multiple defect detection criteria. The method includes the steps of preparing first learning data (Da) (S101), extracting feature values from respective image data items (Di) of the first learning data (S102), performing clustering of the image data items in accordance with the feature values (S103), extracting some image data items-in accordance with the clustering result (S 104 to S 107), and creating a basic model (Mb) by performing learning using basic data (Db) (S108). The basic data includes data on common defectives identified as defectives in common on all of the defect detection criteria. By performing individual learning processing corresponding to the individual defect detection criteria on the basic model (Mb), it is possible to efficiently create a model with desired accuracy. That is, machine learning models corresponding to the individual defect detection criteria can be created with high accuracy and efficiency.

An electrolytic sulfuric acid supply apparatus includes a collection tank (TK1), a generation tank (TK2), a bidirectional pipe (71), a bidirectional first liquid feeder (72), an electrolytic circulator (6), a water supply (3), and a drain (4). The bidirectional pipe connects the collection tank and the generation tank. The first liquid feeder is disposed in the bidirectional pipe. The electrolytic circulator includes an electrolytic circulation pipe (61) connected to the generation tank and an electrolytic cell (65) disposed in the electrolytic circulation pipe. The water supply includes a water supply pipe (31) supplying water to the collection tank and a water supply valve (32). The drain includes a first drainage pipe (41) connected to the collection tank, a first drainage valve (42), a second drainage pipe (43) connected to the generation tank, and a second drainage valve (44).

A processing liquid is supplied to a substrate by a processing nozzle. A processing image showing the substrate to which the processing liquid is supplied is generated by an imaging device. An input image receiver receives a first input image showing a state of the processing liquid on the substrate when an abnormality occurs in a substrate process. Based on the first input image, a shape model generator generates a shape model representing a three-dimensional shape of the processing liquid supplied to the substrate. A simulation executor simulates a transition of a state of the processing liquid in the shape model. An image generator generates an image showing a transition of the state of the processing liquid in a simulation. A collector collects the generated image as an abnormal image. An abnormality detector detects an abnormality in the substrate process based on a comparison between the processing image and the abnormal image.

Realized is a measurement that accommodates diverse repeating patterns of an object to be measured. A measurement apparatus (1) includes: a first light receiving head (22) which receives first reflected light that is first inspection light having been projected from vertically above and reflected by a sample (100); a second light receiving head (32) which receives second reflected light that is second inspection light having been projected from obliquely above and reflected by the sample; and a processor (60), the first light receiving head having a first shutter (23) capable of adjusting an amount of the first reflected light received, the second light receiving head having a second shutter (33) capable of adjusting an amount of the second reflected light received, and the processor controlling an opening and closing amount of each of the first shutter and the second shutter.

The present invention provides a substrate treating method in which a gap in a peripheral edge portion of a laminate substrate obtained by laminating a plurality of substrates is filled with a filler, the method comprising acquiring a first transmission image of the peripheral edge portion of the laminate substrate, filling the gap in the peripheral edge portion of the laminate substrate after acquisition of the first transmission image with the filler, acquiring a second transmission image of the peripheral edge portion after filling with the filler, and determining a filling state of the filler by comparing the first transmission image and the second transmission image.

Segmented images are prepared by dividing a large image. An intermediate target region is extracted from one segmented image selected, and an intermediate-region information item (82) indicating the range of the intermediate target region and an intermediate feature value (83) of the intermediate target region are acquired. This is repeated using each segmented image as the above one segmented image. Based on the acquired intermediate-region information items (82), an integrated feature value (83a) of a target region that integrates multiple intermediate target regions adjoining one another in the large image is obtained from intermediate feature values (83) of the multiple intermediate target regions.