Abstract: The invention discloses a modular quick-fit structure of an oscillating incubator, comprising: an incubator body, a combined mounting slot and an accessory mounting module. The combined mounting slot comprises a sliding slot body, a spring retainer is fixedly attached to a surface of one end of the sliding slot body; the accessory mounting module comprises a combined mounting plate, fixing grooves are provided at edges of two sides of the combined mounting plate, and an accessory module is fixedly installed over the combined mounting plate. The invention at least has the following beneficial effects that quick-fit and quick-connection between the combined mounting plate and the combined mounting slot facilitates the quick-fit and quick-connection between the accessory mounting module and the incubator body, so that problems of small space available for installation inside the machine and inconvenience to mount can be avoided.

Abstract: The present application discloses a CMOS image sensor. A pixel cell circuit comprises a photodiode and a CMOS pixel readout circuit. The pixel cell circuit is formed on an SOI substrate, and the photodiode is formed on a bottom semiconductor substrate. The CMOS pixel readout circuit is formed on a top semiconductor substrate. A photo-induced carrier of the photodiode is connected to the CMOS pixel readout circuit by means of an electrotransfer structure passing through a dielectric buried layer. The present application also discloses a method for manufacturing a CMOS image sensor. The present application can increase a pixel cell density without reducing a photodiode area, thus achieving an ultra-high CMOS image sensor density and improving the device quality.

Abstract: The present application discloses a double-layer stacked CMOS image sensor, photo diode and transfer gate transistor of a pixel cell are formed on the first substrate sequentially along a longitudinal direction, and the other pixel transistors of the pixel cell are formed on the second substrate. The first substrate and the second substrate are packaged separately, and the second substrate is stacked on the top side of the first substrate instead of being in juxtaposition. Since the photo diode and the pixel transistors other than the transfer gate transistor of the pixel cell are located on two separate substrates respectively, the area of a photo diode region may be increased significantly, thereby greatly increasing full well capacitance of the image sensor and increasing a dynamic range, and reduce a dark current and image noise significantly, thereby improving the dark line noise and full well capacitance simultaneously.

Abstract: Techniques are described for template-based image generation. An example, method can include processing a first image of a table, the table comprising a cell. The method can further include processing first bounding box information for a first bounding box, the first bounding box information indicating first bounding box dimensions, the first bounding box corresponding to the cell. The method can further include generating text for populating the cell, wherein text dimensions for the text are based at least in part on the first bounding box dimensions. The method can further include generating a second image, the second image displaying the table with the text overlaid on the cell. The method can further include generating second bounding box information indicating second bounding box dimensions for a second bounding box corresponding to the text as displayed in the second image.

Abstract: In one embodiment, a non-transitory computer-readable media stores instructions executable by processors for generating a prompt configured for eliciting outputs from large language models (LLMs) based on information associated with a task, inputting the prompt to a first LLM configured to output a response based on processing the prompt, determining metrics for evaluating the first LLM based on the task, wherein each of the metrics is associated with a scoring guideline, generating metric prompts based on the respective metrics and the scoring guidelines associated with the respective metrics, inputting the response and the metric prompts to second LLMs configured to output scores corresponding to the respective metrics based on processing the response and the metric prompts, and generating an analysis report based on the metrics and their corresponding scores.

Abstract: Techniques are described for automatically, and substantially without human intervention, generating training data where the training data includes a set of training images containing text content and associated label data. Both the training images and the associated label data are automatically generated. The label data that is automatically generated for a training image includes one or more labels identifying locations of one or more text portions within the training image, and for each text portion, a label indicative of the text content in the text portion. By automating both the generation of training images and the generation of associated label data, the techniques described herein are very scalable and repeatable and can be used to generate large amounts of training data, which in turn enables building more reliable and accurate language models.

Abstract: During polishing of a backside conductive layer, a sensor of an in-situ eddy current monitoring system is repeatedly swept across the substrate so that each respective sweep of the sensor generates a respective signal trace that includes a sequence of signal values. For each respective signal trace, the sequence of signal values is converted to a corresponding thickness trace that includes sequence of thickness values for different locations on the substrate, thus generating a sequence of thickness traces. For each respective thickness trace in the sequence of thickness traces, a plurality of minima in the respective thickness trace are identified. A sequence of layer thickness values over time is calculated based on the plurality of minima from the respective traces in the sequence of thickness traces. Conductive vias extend through the semiconductor wafer of the substrate to electrically connect the backside conductive layer to a front-side conductive layer.

Abstract: A chemical mechanical polishing apparatus includes a platen to support a polishing pad, a carrier head to hold a substrate such that a layer on the substrate contacts the polishing pad, an actuator that controls a radial position of the carrier head over the platen, an eddy current monitoring system, and a controller. The eddy current monitoring system includes a first plurality of eddy current sensors supported by the platen and arranged in a first ring at a first distance from an axis of rotation of the platen and a second plurality of eddy current sensors supported by the platen and arranged in a second ring at a larger second distance from the axis of rotation of the platen. The controller is configured to control the actuator such that the second plurality of sensors sweep only across an edge portion of the substrate held by the carrier head.

Abstract: Techniques are disclosed for optical character recognition of extensible markup language content. A method can include a system generating a first training data comprising extensible markup language (XML) content, the first training data comprising a first plurality of training instances, each training instance including a respective image comprising XML content and annotation information for the respective image. The system can train a plurality of machine learning models using the first training data to generate a plurality of trained machine learning models, to perform image-based XML content extraction. The system can generate a plurality of trained machine learning models based at least in part on the training.

Abstract: Methods and apparatuses are provided herein for independently adjusting flowpath conductance. One multi-station processing apparatus may include a processing chamber, a plurality of process stations in the processing chamber that each include a showerhead having a gas inlet, and a gas delivery system including a junction point and a plurality of flowpaths, in which each flowpath includes a flow element, includes a temperature control unit that is thermally connected with the flow element and that is controllable to change the temperature of that flow element, and fluidically connects one corresponding gas inlet of a process station to the junction point such that each process station of the plurality of process stations is fluidically connected to the junction point by a different flowpath.

Abstract: In one embodiment, a non-transitory computer-readable media stores instructions executable by processors for generating a prompt configured for eliciting outputs from large language models (LLMs) based on information associated with a task, inputting the prompt to a first LLM configured to output a response based on processing the prompt, determining metrics for evaluating the first LLM based on the task, wherein each of the metrics is associated with a scoring guideline, generating metric prompts based on the respective metrics and the scoring guidelines associated with the respective metrics, inputting the response and the metric prompts to second LLMs configured to output scores corresponding to the respective metrics based on processing the response and the metric prompts, and generating an analysis report based on the metrics and their corresponding scores.

Abstract: A method of manufacturing a semiconductor structure including the following steps is provided. A substrate is provided. The substrate has a first region and a second region. A stacked structure is formed on the substrate in the first region. The stacked structure includes a first dielectric layer, a charge storage layer, a second dielectric layer, a first conductive layer, and a first hard mask layer. A dielectric material layer is formed on the substrate in the second region. A second conductive layer is formed on the dielectric material layer in the second region. A first patterned photoresist layer is formed. The first hard mask layer exposed by the first patterned photoresist layer and a portion of the dielectric material layer exposed by the first patterned photoresist layer are removed by using the first patterned photoresist layer as a mask.

Abstract: An accurate positioning mechanism of a high-speed biological shaking table, which is characterized by comprising a servo motor, an eccentric shaft, a motor connecting disc, a positioning balancing weight, a positioning magnet and a Hall sensor, an eccentric shaft is installed at the top main body end of the servo motor, a motor connecting disc is installed on the surface of the eccentric shaft, a positioning balancing weight is arranged at the top end of the eccentric shaft, a positioning magnet is arranged on the surface of the side, away from the eccentric shaft, of the positioning balancing weight, and a Hall sensor is installed on the surface of the rocking shaft upper plate.

Abstract: A method for processing substrates includes subjecting each respective first substrate of a first plurality of substrates to a process that modifies a thickness of an outer layer of the respective first substrate, generating a plurality of groups of process parameter values; generating a plurality of removal profiles, training an artificial neural network by backpropagation using the plurality of groups of process parameter and plurality of removal profiles as training data where the artificial neural network has a plurality of input nodes to receive respective removal values from the removal profile and a plurality of output nodes to output control parameter values, for each respective second substrate of a second plurality of substrates determining a target removal profile, determining respective control parameter values to apply by applying the target removal profile to the input nodes, and subjecting each respective second substrate to the process using the respective control parameter values.

Abstract: Techniques are described for HTML-based image generation. An example, method can include generating hypertext markup language (HTML) code for a table comprising a table structure of a set of rows and columns. The method can further include generating HTML code for a text to populate a cell of the table. The method can further include generating a rendered image of the table using the HTML code. The method can further include detecting a first pixel of the rendered image comprising the first color, and a second pixel of the rendered image comprising the second color. The method can further include detecting the text on the rendered image. The method can further include generating a bounding box, surrounding the detected text. The method can further include generating annotation comprising a bounding box parameter and a text parameter.

Abstract: A chemical mechanical polishing apparatus includes a platen to support a polishing pad, a carrier head to hold a surface of a substrate against the polishing pad, a motor to generate relative motion between the platen and the carrier head so as to polish an overlying layer on the substrate, an in-situ acoustic monitoring system comprising an acoustic sensor that receives acoustic signals from the surface of the substrate, and a controller configured to determine a angular orientation of the substrate based on received acoustic signals from the in-situ acoustic monitoring system.

Abstract: Techniques for evaluating the output of a large language model are disclosed. A training data set that includes deterministic computational metrics that measure features of large language model output and qualitative metrics that provide a non-deterministic measure of large language model output quality may be used to train a ML model. The ML model may then be used to estimate the qualitative metrics of large language model output by using deterministic computational metrics as input.

Abstract: A body is brought into contact with a polishing pad of a polishing system, a polishing liquid is supplied to the polishing pad, relative motion between the body and the polishing pad is generated while the body contacts the polishing pad, a signal from an in-situ eddy current monitoring system during the relative motion while the body contacts the polishing pad, generating, and mechanical vibrations in the polishing system are detected based on a signal from the in-situ eddy current monitoring system.

Abstract: Methods and apparatus for lamp housing crack detection are provided herein. For example, a method associated with a process chamber having a lamp housing comprises sequentially processing a plurality of substrates in the process chamber, during processing of the plurality of substrates, collecting lamp housing data indicative of a fluid leak in the lamp housing from a sensor operably connected to the lamp housing of the process chamber, determining from the lamp housing data whether lamp housing crack is present, and responsive to determining lamp housing crack is present, at least one of triggering an alert or stopping processing of the process chamber.

Abstract: Methods and apparatuses for controlling precursor flow in a semiconductor processing tool are disclosed. A method may include flowing gas through a gas line, opening an ampoule valve(s), before a dose step, to start a flow of precursor from the ampoule to a process chamber through the gas line, closing the ampoule valve(s) to stop the precursor from flowing out of the ampoule, opening a process chamber valve, at the beginning of the dose step, to allow the flow of precursor to enter the process chamber, and closing the process chamber valve, at the end of the dose step, to stop the flow of precursor from entering the process chamber. A controller may include at least one memory and at least one processor and the at least one memory may store instructions for controlling the at least one processor to control precursor flow in a semiconductor processing tool.