Applied Materials, Inc.

Paypal, Inc.

Amit Kumar Bansal

Juan Carlos Rocha-Alvarez

Tuan Anh Nguyen

Ganesh Balasubramanian

Jianhua Zhou

Mohamad A Ayoub

Ramprakash Sankarakrishnan

Xinhai Han

Masaki Ogata

Sanjeev Baluja

Jian J Chen

Bok Hoen Kim

Thomas E Nowak

Zhijun Jiang

Nitin Pathak

Abdul Aziz Khaja

Dale Robert Du Bois

Patrick Reilly

Shahid Shaikh

Todd J Egan

Nagarajan Rajagopalan

Edward W Budiarto

Tejas Ulavi

Terrance Y Lee

Michael Wenyoung Tsiang

Jeongmin Lee

Heung Lak Park

Badri N Ramamurthi

Sergey Starik

Dmitriy Panasyuk

Kalyanjit Ghosh

Shailendra Srivastava

Juan Carlos Rocha

Yuxing Zhang

Deenesh Padhi

Karthik Janakiraman

Yusheng Zhou

Thomas Rubio

David T Blahnik

Mark A Fodor

Kaushik Comandoor Alayavalli

Robert Kim

Jun Ma

Kwangduk Douglas Lee

Prashant Kumar Kulshreshtha

Ziqing Duan

Zheng John Ye

Sarah Michelle Bobek

Daemian Raj Benjamin Raj

Chiu Chan

Sam Hyungsam Kim

Nikhil Sudhindrarao Jorapur

Sai Susmita Addepalli

Wenjiao Wang

Gregory Eugene Chichkanoff

Kristopher R Enslow

Saket Rathi

Visweswaren Sivaramakrishnan

Hichem M'Saad

Jason M Schaller

Ren-Guan Duan

Ajit Balakrishna

Diwakar N Kedlaya

Edward P Hammond

Hang Yu

Venkata Sharat Chandra Parimi

Eller Y Juco

Bozhi Yang

Kirby Hane Floyd

Siu Fai Cheng

Fang Ruan

Binh Nguyen

Thuy Britcher

Kallol Bera

Kartik Bhupendra Shah

Tza-Jing Gung

Amir Hamed Al-Bayati

Mukund Srinivasan

Derek R Witty

Martin Jay Seamons

Swaminathan T Srinivasan

Praket Prakash Jha

Sudha S Rathi

Satoru Kobayashi

Matthew James Busche

Jonghoon Baek

Mayur G Kulkarni

Scott A Hendrickson

Satish Radhakrishnan

Xing Lin

Sungwon Ha

Jeffrey Charles Blahnik

Ratsamee Limdulpaiboon

Anjana M Patel

Yongjing Lin

Xiaopu Li

Ndanka O Mukuti

Dong Hyung Lee

Anton Baryshnikov

Rupankar Choudhury

Methods and systems are presented for tagging an account associated with a user based on a predicted likelihood of an event associated with the user. A set of features is determined for data associated with the user. Values from the data are aggregated over time intervals for each feature to create time series data. The time series data is used as input to a neural network configured to accept input with the determined features. A predictive value indicating the likelihood of an event associated with the user is received from the neural network and used to determine whether to tag a user account. Determinations regarding the user are made based on the existence of absence of a tag on the user's account.

System and method for tag-directed deep-learning-based features for predicting events and making determinations

Exemplary semiconductor processing systems may include an output manifold that defines at least one plasma outlet. The systems may include a gasbox disposed beneath the output manifold. The gasbox may include an inlet side facing the output manifold and an outlet side opposite the inlet side. The gasbox may include an inner wall that defines a central fluid lumen. The inner wall may taper outward from the inlet side to the outlet side. The systems may include an annular spacer disposed below the gasbox. An inner diameter of the annular spacer may be greater than a largest inner diameter of the central fluid lumen. The systems may include a faceplate disposed beneath the annular spacer. The faceplate may define a plurality of apertures extending through a thickness of the faceplate.

Uniform in situ cleaning and deposition

Embodiments presented herein are directed to radio frequency (RF) grounding in process chambers. In one embodiment, a dielectric plate is disposed between a chamber body and a lid of a process chamber. The dielectric plate extends laterally into a volume defined by the chamber body and the lid. A substrate support is disposed in the volume opposite the lid. The substrate support includes a support body disposed on a stem. The support body includes a central region and a peripheral region. The peripheral region is radially outward of the central region. The central region has a thickness less than a thickness of the peripheral region. A flange is disposed adjacent to a bottom surface of the peripheral region. The flange extends radially outward from an outer edge of the peripheral region. A bellows is disposed on the flange and configured to sealingly couple to the dielectric plate.

Radio frequency power return path

The present disclosure generally relates to an apparatus for improving azimuthal uniformity of a pressure profile of a processing gas. In one example, a processing chamber includes a lid, sidewalls, and a substrate support defining a processing volume. A bottom bowl, a chamber base, and a wall define a purge volume. The purge volume is disposed beneath the processing volume. The bottom bowl includes a first surface having a first equalizer hole. A passage couples the processing volume to the purge volume via the first equalizer hole and an inlet. The passage is positioned above the first equalizer hole. The chamber base has a purge port coupleable to a purge gas line for supplying a purge gas to the purge volume. A baffle is disposed in the purge volume at a height above the purge port, and is configured to deflect a trajectory of the purge gas.

Baffle implementation for improving bottom purge gas flow uniformity

A faceplate for a substrate process chamber comprises a first and second surface. The second surface is shaped such that the second surface includes a peak and a distance between the first and second surface varies across the width of the faceplate. The second surface of the faceplate is exposed to a processing volume of the process chamber. Further, the faceplate may be part of a lid assembly for the process chamber. The lid assembly may include a blocker plate facing the first surface of the faceplate. A distance between the blocker plate and the first surface is constant.

Faceplate having a curved surface

Implementations disclosed herein generally relate to systems and methods of protecting a substrate support in a process chamber from cleaning fluid during a cleaning process. The method of cleaning the process chamber includes positioning in the process chamber a cover substrate above a substrate support and a process kit that separates a purge volume from a process volume. The method of cleaning includes flowing a purge gas in the purge volume to protect the substrate support and flowing a cleaning fluid to a process volume above the cover substrate, flowing the cleaning fluid in the process volume to an outer flow path, and to an exhaust outlet in the chamber body. The purge volume is maintained at a positive pressure with respect to the process volume to block the cleaning fluid from the purge volume.

Method and system for cleaning a process chamber

Exemplary semiconductor processing systems include a processing chamber defining a processing region. The semiconductor processing systems may include a foreline coupled with the processing chamber. The foreline may define a fluid conduit. The semiconductor processing systems may include a foreline trap coupled with a distal end of the foreline. The semiconductor processing systems may include a removable insert provided within an interior of the foreline trap. The semiconductor processing systems may include a throttle valve coupled with the foreline trap downstream of the removable insert.

Actively cooled foreline trap to reduce throttle valve drift

Machine learning models are powerful artificial intelligence tools that can make determinations based on a variety of factors. Unlike a simple linear model, however, determining the contribution of each variable to the outcome of a machine learning model is a challenging task. It may be unclear which factors contributed heavily toward a particular outcome of the machine learning model and which factors did not have a major effect on the outcome. Being able to accurately determine the underlying causative factors for a machine learning-based decision, however, can be important in several contexts. The present disclosure describes techniques that allow for training and use of non-linear machine learning models, while also preserving causal information for outputs of the models. Relative weight calculations for machine learning model variables can be used to accomplish this, in various embodiments.

Preservation of causal information for machine learning

The present disclosure relates to a cleaning assemblies, components thereof, and methods associated therewith for substrate processing chambers. In one example, a cleaning assembly for a substrate processing chamber includes a distribution ring. The distribution ring comprises a body with an inlet and an outlet. The outlet is fluidly coupled to an internal volume of the substrate processing chamber via a sidewall of the substrate processing chamber. The cleaning assembly includes a cleaning conduit configured to fluidly couple a gas manifold to the distribution ring for diverting a first portion of cleaning fluid from the gas manifold to the distribution ring.

Cleaning assemblies for substrate processing chambers

Exemplary semiconductor processing chambers may include a substrate support including a top surface. A peripheral edge region of the top surface may be recessed relative to a medial region of the top surface. The chambers may include a pumping liner disposed about an exterior surface of the substrate support. The chambers may include a liner disposed between the substrate support and the pumping liner. The liner may be spaced apart from the exterior surface to define a purge lumen between the liner and the substrate support. The chambers may include an edge ring seated on the peripheral edge region. The edge ring may extend beyond a peripheral edge of the substrate support and above a portion of the liner. A gap may be formed between a bottom surface of the edge ring and a top surface of the liner. The gap and the purge lumen may be fluidly coupled.

Hardware to prevent bottom purge incursion in application volume and process gas diffusion below heater

A processing chamber may include a gas distribution member, a substrate support, and a pumping liner. The gas distribution member and the substrate support may at least in part define a processing volume. The pumping liner may define an internal volume in fluid communication with the processing volume via a plurality of apertures of the pumping liner circumferentially disposed about the processing volume. The processing chamber may further include a flow control mechanism operable to direct fluid flow from the internal volume of the pumping liner into the processing volume via a subset of the plurality of apertures of the pumping liner during fluid distribution into the processing volume from the gas distribution member.

Semiconductor processing chambers and methods for cleaning the same

Embodiments described herein relate to apparatus and techniques for radio frequency (RF) phase control in a process chamber. A process volume is defined in the process chamber by a faceplate electrode and a support pedestal. A grounding bowl is disposed within the process chamber about the support pedestal opposite the process volume. The grounding bowl substantially fills a volume other than the process volume below the support pedestal. A phase tuner circuit is coupled to an RF mesh disposed in the support pedestal and the faceplate electrode. The tuner circuit adjusts a phase difference between a phase of the faceplate electrode and a phase of the RF mesh.

Device and method for tuning plasma distribution using phase control

A method of processing a substrate according to a PECVD process is described. Temperature profile of the substrate is adjusted to change deposition rate profile across the substrate. Plasma density profile is adjusted to change deposition rate profile across the substrate. Chamber surfaces exposed to the plasma are heated to improve plasma density uniformity and reduce formation of low quality deposits on chamber surfaces. In situ metrology may be used to monitor progress of a deposition process and trigger control actions involving substrate temperature profile, plasma density profile, pressure, temperature, and flow of reactants.

PECVD process

A semiconductor processing system includes a remote plasma source (RPS), a faceplate, and an output manifold positioned between the RPS and the faceplate. The output manifold is characterized by a plurality of purge outlets that are fluidly coupled with a purge gas source and a plurality of deposition outlets that are fluidly coupled with a deposition gas source. A delivery tube extends between and fluidly couples the RPS and the faceplate. The delivery tube is characterized by a generally cylindrical sidewall that defines an upper plurality of apertures that are arranged in a radial pattern. Each of the upper apertures is fluidly coupled with one of the purge outlets. The generally cylindrical sidewall defines a lower plurality of apertures that are arranged in a radial pattern and below the upper plurality of apertures. Each of the lower apertures is fluidly coupled with one of the deposition outlets.

Gas mixer to enable RPS purging

Growing community of inventors

Milpitas, CA, United States of America

Amit Kumar Bansal