Abstract: A semiconductor processing system includes a gas delivery module, and a chamber body connected to the gas delivery module. The divider has an aperture, is fixed within an interior of the chamber body, and separates an interior of the chamber body into upper and lower chambers, the aperture fluidly coupling the lower chamber to the upper chamber. A susceptor is arranged within the aperture. A controller is operably connected to the gas delivery module to purge the lower chamber with a first purge flow including an etchant while etching the upper chamber, purge the lower chamber with a second purge flow including the etchant while depositing a precoat in the upper chamber, and purge the lower chamber with a third purge flow including the etchant while depositing a film onto a substrate in the upper chamber. Film deposition methods and lower chamber etchant purge kits are also described.

Abstract: The present disclosure discloses a nonlinear all-optical deep-learning system and method with multistage space-frequency domain modulation. The system includes an optical input module, configured to convert input information to optical information, a multistage space-frequency domain modulation module, configured to perform multistage space-frequency domain modulation on the optical information generated by the optical input module so as to generate modulated optical information, and an information acquisition module, configured to transform the modulated optical information onto a Fourier plane or an image plane, and to acquire the transformed optical information so as to generate processed optical information.

Abstract: Systems of reducing devitrification within a chemical vapor deposition system can include a susceptor support ring that is configured to be positioned between a gas inlet and a gas outlet of a chamber passage. An example system can also include a getter support comprising a support base and one or more recesses therein. Each of the one or more recesses can be arranged to receive corresponding one or more support elements that are configured to support the getter plate. At least a portion of the getter support may include a coating comprising silicon carbide (SiC) having a thickness of at least about 50 microns. The getter support may be arranged to be disposed a maximum distance of between about 1 mm and about 10 mm from the susceptor support ring.

Abstract: A reflector includes a reflector body arranged to overlap a reaction chamber of a semiconductor processing system. The reflector body has a grooved surface and a reflective surface extending between a first longitudinal edge of the reflector body and a second longitudinal edge of the reflector body, the reflective surface spaced apart from the grooved surface by a thickness of the reflector body. The grooved surface and the reflective surface define a pyrometer port, two or more elongated slots, and two or more shortened extending through the thickness of the reflector body. The shortened slots outnumber the elongated slots to bias issue of a coolant against the reaction chamber toward the second longitudinal edge of the reflector body. Cooling kits, semiconductor processing systems, and methods of cooling a reaction chamber during deposition of a film onto a substrate supported within the reaction chamber are also described.

Abstract: A device and a method for generating a machine translation model and a machine translation device are disclosed. The device inputs a source training sentence of a source language and a dictionary data to a generator network so that the generator network outputs a target training sentence of a target language according to the source training sentence and the dictionary data. Then, the device inputs the target training sentence and a correct translation of the source training sentence to a discriminator network so as to calculate an error between the target training sentence and the correct translation according to the output of the discriminator network, and trains the generator network and the discriminator network respectively. The trained generator network is the machine translation model.

Abstract: An all-optical Diffractive Deep Neural Network (D2NN) architecture learns to implement various functions or tasks after deep learning-based design of the passive diffractive or reflective substrate layers that work collectively to perform the desired function or task. This architecture was successfully confirmed experimentally by creating 3D-printed D2NNs that learned to implement handwritten classifications and lens function at the terahertz spectrum. This all-optical deep learning framework can perform, at the speed of light, various complex functions and tasks that computer-based neural networks can implement, and will find applications in all-optical image analysis, feature detection and object classification, also enabling new camera designs and optical components that can learn to perform unique tasks using D2NNs. In alternative embodiments, the all-optical D2NN is used as a front-end in conjunction with a trained, digital neural network back-end.

Abstract: A system and method for removing both carbon-based contaminants and oxygen-based contaminants from a semiconductor substrate within a single process chamber is disclosed. The invention may comprise utilization of remote plasma units and multiple gas sources to perform the process within the single process chamber.

Abstract: A susceptor has a circular pocket portion, an annular ledge portion, and an annular rim ledge portion. The circular pocket portion is arranged along a rotation axis and has a perforated surface. The annular ledge portion extends circumferentially about pocket portion and has ledge surface that slopes axially upward from the perforated surface. The rim portion extends circumferentially about the ledge portion and is connected to the pocket portion by the ledge portion of the susceptor. The susceptor has one or more of a tuned pocket, a contact break, a precursor vent, and a purge channel located radially outward of the perforated surface to control deposition of a film onto a substrate supported by the susceptor. Semiconductor processing systems, film deposition methods, and methods of making susceptors are also described.

Abstract: A packing module includes first and second flow paths which exchange heat between water sprayed from above and air flowing from below; a first import portion for importing water sprayed from one side of the packing module into the first flow path; a second import portion for importing water sprayed from the other side of the packing module into the second flow path; a first export portion for guiding water flowing out from the first water path to one side of the packing module for discharging; and, a second export portion for guiding water flowing out from the second water path to the other side of the packing module for discharging. Two flow paths in the packing module operate in different operation modes, so that water or wind flows through one flow path to cool water while air flows through the other flow path for heat exchange through partition walls.

Abstract: In some embodiments, a method for semiconductor processing preclean includes removing an oxide layer from a substrate using anhydrous hydrogen fluoride in combination with water vapor. A system for the preclean may be configured to separate the anhydrous hydrogen fluoride and the water vapor until they are delivered to a common volume near the substrate. Corrosion within components of the system may be limited by purification of anhydrous hydrogen fluoride, passivation of components, changing component materials, and heating components. Passivation may be achieved by filling a gas delivery component with anhydrous hydrogen fluoride and allowing the anhydrous hydrogen fluoride to remain in the gas delivery component to form a passivation layer. Consistent water vapor delivery may be achieved in part by heating components using heaters.

Abstract: A method of operating a reactor system to provide multi-zone substrate temperature control. The method includes, with a first pyrometer, sensing a temperature of a first zone of a substrate supported in the reactor system, and, with a second pyrometer, sensing a temperature of a second zone of the substrate. The method further includes, with a controller, comparing the temperatures of the first and second zones to setpoint temperatures for the first and second zones and, in response, generating control signals to control heating of the substrate. The method also includes controlling, based on the control signals, operations of a heater assembly operating to heat the substrate.

Abstract: A reactor system designed to provide accurate monitoring of wafer temperatures during deposition steps. The reactor system includes a pyrometer mounting assembly supporting and positioning three or more pyrometers (e.g., infrared (IR) pyrometers) relative to the reaction chamber to measure a center wafer temperature and an edge wafer temperature as well as reaction chamber temperature. The pyrometer mounting assembly provides a small spot size or temperature sensing area with the edge pyrometer to accurately measure edge wafer temperatures. A jig assembly, and installation method for each tool setup, is provided for use in achieving accurate alignment of the IR pyrometer sensing spot (and the edge pyrometer) relative to the wafer, when the pyrometer mounting assembly is mounted upon a lamp bank in the reactor system or in tool setup. The wafer edge temperature sensing with the reactor system assembled with proper alignment ensures accurate and repeatable measurement of wafer temperatures.

Abstract: A method of operating a reactor system to provide wafer temperature gradient control is provided. The method includes operating a center temperature sensor, a middle temperature sensor, and an edge temperature sensor to sense a temperature of a center zone of a wafer on a susceptor in reaction chamber of the reactor system, to sense a temperature of a middle zone of the wafer, and to sense a temperature of an edge zone of the wafer. The temperatures of the center, middle, and edge zones of the wafer are processed with a controller to generate control signals based on a predefined temperature gradient for the wafer. First, second, and third sets of heater lamps are operated based on the temperature of the center, middle, and edge zones to heat the center, the middle, and the edge zone of the wafer. Reactor systems are also described.

Abstract: Methods, non-transitory machine readable media, and computing devices that provide improved dictionary-based compression are disclosed. With this technology, a first portion of an input data stream is compressed using a first dictionary. A second dictionary is trained when the first dictionary is determined to be stale. The dictionary can be determined to be stale based on a size of the input data stream compressed using the first dictionary or a compression ratio decreasing by a threshold, for example. The first dictionary can be stored with metadata associated with the compressed first portion of the input data stream. Accordingly, this technology improves compression ratios, eliminates the need for reference counting, and facilitates improved reclamation of orphan dictionaries, among other advantages.

Abstract: An all-optical Diffractive Deep Neural Network (D2NN) architecture learns to implement various functions or tasks after deep learning-based design of the passive diffractive or reflective substrate layers that work collectively to perform the desired function or task. This architecture was successfully confirmed experimentally by creating 3D-printed D2NNs that learned to implement handwritten classifications and lens function at the terahertz spectrum. This all-optical deep learning framework can perform, at the speed of light, various complex functions and tasks that computer-based neural networks can implement, and will find applications in all-optical image analysis, feature detection and object classification, also enabling new camera designs and optical components that can learn to perform unique tasks using D2NNs. In alternative embodiments, the all-optical D2NN is used as a front-end in conjunction with a trained, digital neural network back-end.

Abstract: A fixture includes a frame, a leveling plate, a bracket, and a laser profiler. The frame is arranged for fixation above a reaction chamber arranged to deposit a film onto a substrate. The leveling plate is supported on the frame. The bracket is supported on the leveling plate. The laser profiler is suspended from the bracket, overlays the reaction chamber, and has a field of view that extends through the leveling plate and the frame to determine position of a target within the reaction chamber. Semiconductor processing systems and methods of determining position of targets within reaction chambers in semiconductor processing systems are also described.

Abstract: A substrate support and lift assembly configured to support and lift a substrate from a susceptor is disclosed. The substrate support and lift assembly can include a susceptor support and a lift pin. The susceptor support can be configured to support the susceptor thereon. The susceptor support includes a plurality of support arms each extending radially from a central portion of the susceptor support to a terminus. Each of the plurality of support arms includes an aperture extending therethrough. The lift pin can be configured to fit through the aperture of a corresponding support arm to lift a substrate on the susceptor.

Abstract: An optical diffractive processing unit includes input nodes, output nodes; and neurons. The neurons are connected to the input nodes through optical diffractions. Weights of connection strength of the neurons are determined based on diffractive modulation. Each optoelectronic neuron is configured to perform an optical field summation of weighted inputs and generate a unit output by applying a complex activation to an optical field occurring naturally in a photoelectronic conversion. Each neuron is a programmable device.

Abstract: In order to reduce write tail latency, a storage system generates redundant write requests when performing a storage operation for an object. The storage operation is determined to be effectively complete when a minimum number of write requests have completed. For example, the storage system may generate twelve write requests and also generate four redundant write requests for a total of sixteen write requests. The storage system considers the object successfully stored once twelve of the sixteen writes complete successfully. To generate the redundant writes, the storage system may use replication or erasure coding. For replication, the storage system may issue a redundant write request for each of n chunks being written. For erasure coding, the storage system may use rateless codes which can generate unlimited number of parity chunks or use an n+k+k? erasure code which generates an additional k? encoded chunks, in place of an n+k erasure code.

Abstract: A touch panel and a touch display are disclosed. The touch panel mainly comprises a light transmissive substrate and two sensing units. The two sensing units are disposed oppositely through the light transmissive substrate. A mesh width of a metal mesh in the two sensing units is large, which can increase a variation of capacitance value and improve the touch sensitivity. In addition, an angle between metal lines with sensing function is configured in a random manner under a certain condition to reduce an area of nodes, thus a rounding effect at the nodes can be effectively reduced or eliminated. The metal meshes of the two sensing units disposed oppositely are staggered with a non-conductive node-free pattern to form a staggered pattern with irregular polygons, so that no interference fringes would be generated when the touch panel is used together with display panels of various mainstream sizes.