AI

Semiconductor fabs have been getting smarter and smarter over the past 30 years. It’s a natural evolution – the direct outcome of numerous continuous-improvement efforts. The really important difference on the road to smarter fabs, the one change that’s enabling the Industry 4.0 revolution, is the concept of a cyber-physical system or digital twin. If you don’t have a thorough, detailed, high-fidelity digital twin of your entire fab operation, then you cannot have “Smart Manufacturing.” That’s really the definition of a smart site. A digital twin is simply a requirement for all smart factories of the future. One caveat: No matter what you build today from a smart perspective, your digital twin’s fidelity will improve over the next 20 years. A factory’s digital twin has two facets: the operational aspect and the yield aspect. Each of these two facets places different requirements on a database including the types of data, the frequency of data generated, the retention of data, and even the AI/ML techniques used to analyze the data. A combination of these data requirements are needed to create a digital twin – the virtual representation of your entire factory operation, whether it’s on the wafer-fab front end or the assembly and test back end. What’s most important here is that facility-wide data sets and databases must be able to communicate with each other using refined summary statistics to create a practical digital twin. For example, a lot of information is collected on the yield side to feed the deep-learning models needed to manage processes. However, the factory scheduler, driven largely by the smart operational database, needs only summary statistics from the yield database to be able to act in the next moment or over the next 24 hours. Figure 1 illustrates the needs of and the interaction between a smart operational and a yield database. Figure 1: The Operational and Yield databases in a Smart Factory need to exchange summary statistics. Today, we find that although these databases generally speak to each other in smart factories, they’re still not sufficiently connected to permit the use and analysis of data needed to realize the full potential of a smart factory. That level of interconnectedness is still in the future. Some solution providers have created what is essentially a “smart learning warehouse” (“database” has become too limited a term here). This warehouse collects, analyzes and learns from the extensive amount of information that a fab generates. Game-changing, more holistic applications become possible when this information can be combined in new and informative ways. As it turns out, a data source is just a data source, but users in different factory areas need to extract different information from these common data sources. They need different applications and portals – in other words “views” – that are adapted and adjusted for each area’s needs. Aren’t we smart enough? Some people think that 300mm fabs are already smart. That’s true. They are. But, they could be a lot smarter. No 300mm fab in use today has attained the full, utopian vision of what a smart factory can deliver over the next 10 years. When you finally integrate all of the disparate databases in a fab – when you’re able to use all of those different data sources as one common data source – that’s when your Smart Factory will have the ability to self-optimize its future actions and react quickly to real-time events. The largest semiconductor manufacturers tend to develop these smart factory applications on their own. The remaining semiconductor fabs need to work together with other fabs and their solution providers to develop these smart factory applications. Why now? Why is everyone talking about “Smart” now? It’s because the semiconductor industry has helped to create all of the enabling technology: the compute power, the networking and networking standards, and even the industry’s maturation into a multi-tiered organization of solution providers. We’ve reached the point where we can collect data from a widespread sensor network along with tool-health data and we can then warehouse this data so that it can be applied to more intelligent decision-making. While there may be one or two sensors on a tool today, in the future there will be many such sensors connected over an IoT network or networks that provide mountains of data to the warehouse. All of this data will feed into the digital-twin version of the fab. One of the biggest changes on the horizon made possible by all of this accessible data is advanced scheduling. Despite all of the automation advancements made over the past 25 years, including robotic handling, it’s still hard to decide “where, what, and when?” for every single lot in the factory. Today, no factory in the world is more complex than a semiconductor fab. Optimizing a semiconductor manufacturing process is the most complex manufacturing-optimization task in the world. Do it for ROI ROI is the chief reason for having a digital twin. Once you can make a truly smart, holistic schedule of the fabs operations — not a dispatch or rule-based dispatch list — then you can create an operationally smart factory. Rule-based dispatching systems primarily focus on tools and tool-centric views. Although they incorporate knowledge from current WIP and tool conditions to make decisions better than simple dispatch systems, smart factories are not just about tools and the current WIP at them. Smart factories use the status of every tool and lot in the factory to make fab-centric optimizations instead of tool-specific optimizations. Once you have a digital twin, you’re optimizing for global functions such as line linearity and on-time delivery. These functions are not just about the moment. The transition to a smart factory thus represents a huge philosophical change. When you know exactly what’s going to happen in a factory over the next 12 hours for every single lot, every single wafer carrier, and every single entrance port of every tool in the factory, then you suddenly have control over the factory’s idle time. You know when you can optimally perform PM (preventive maintenance). You know how to best redirect material or labor resources to maximize output. You can create a smart schedule for every maintenance person in the factory that comprehends each person’s skill set and tool downtime so that there’s no negative impact on the factory’s productivity. You can only do all of this when you know the future. Figure 2 illustrates the opportunity. Imagine that a factory contains 1,500 tools. Use of these tools is scheduled for the next twelve hours. The information depicted in Figure 2 encompasses process changes from one chemistry to another, implant changes, reticle changes, and the status of every single consumable for all 1,500 tools. The white spaces that appear between processes in Figure 2 represent opportunities to intelligently schedule events such as maintenance to maximize factory productivity. Figure 2: Smart scheduling permits factory-wide optimization to maximize productivity. Once you have a schedule, you need to translate that schedule into actions or movement. It’s not easy to do this and most material-control systems today make overly simplistic decisions based on modeled assumptions and typical cases rather than the actual time each lot needs to be at a precise location, which can only come from a schedule. Once the data from all of the tools is connected, a smart scheduling system can use the digital twin to make far better process decisions. The larger the factory (or more complex the factory), the more important it is to make smarter decisions. Note: SEMI has a Smart Manufacturing Technology Community. For more information or to get involved, click here. If you would like to discuss Smart Manufacturing more with John directly, he can be contacted at [email protected]. John Behnke is general manager of the Final Phase Systems product line at INFICON.

Artificial intelligence (AI) is on the verge of transforming entire industries as it gears up to power semiconductor industry innovation and growth, thrusting the technology to front and center at SEMICON Japan 2019, December 12-14 at the Tokyo Big Sight (Tokyo International Exhibition Center).The SMART Technology Forum at SEMICON Japan will highlight the latest AI developments and trends. Supported by U.S. Commercial Service in Japan, the forum will feature Yutaka Matsuo of the University of Tokyo. An authority on AI, Matuso will give an overview of both AI business and technology. His presentation will be followed by an AI outlook from Microsoft Japan, Amazon Web Services and DefinedCrowd.A number of Japanese startups are on leading edge of AI innovation in machine and deep learning. One is Preferred Networks Inc., a company that applies cutting-edge deep learning technology to Internet of Things (IoT) applications across transportation, manufacturing and healthcare.In his opening day keynote at SEMICON Japan, Toru Nishikawa, president and CEO of Preferred Networks, Inc., will highlight the latest developments and promise of using deep learning for industrial applications. Nishikawa will unpack how AI companies jockeying for competitive advantage will win by harnessing technologies to process massive amounts of data efficiently and quickly.Following is look at Preferred Networks, Inc. and five other Japanese startups that are driving AI innovation. Within Japan's world of AI, machine learning, and deep dearning, Preferred Networks is likely the most well-known Japanese company. The parent company, Preferred Infrastructure, was founded in March 2006 by Toru Nishikawa and Daisuke Okanohara, who focused on search engine development before turning to machine learning and establishing Preferred Networks to commercialize the technology.Preferred Networks established itself as one of the world’s top providers of machine learning technology with the development of Chainer – an open source deep learning framework that has been offered free of charge since June 2015 and was released before TensorFlow, Google’s renowned Deep Learning framework. Established in 2012, ABEJA is thought to be Japan’s first venture company to specialize in deep learning. ABEJA's core technology is its AI platform ABEJA Platform. Based on this platform, the company offers various solutions to more than 100 client companies. ABEJA also offers ABEJA Insight, a specialized package service for the retail and distribution, manufacturing, and infrastructure industries. Data analytics provider BrainPad Inc. was the first Japanese AI venture listed on the Tokyo Stock Exchange. Established in 2004, before the advent of big data, BrainPad Inc. cultivated a vision of analyzing vast amounts of data in increase the competitiveness of Japanese companies. LeapMind Inc. aims to offer deep learning technology that uses fewer computing resources and draws less power. Both are important capabilities since deep learning requires considerable computing resources to perform image and speech recognition. The company’s answer to this deep learning challenge is a small form factor FPGA with low power consumption.In April 2018, LeapMind started offering the tool DeLTA-Lite to support model construction for Deep Learning. The tool simplifies the development of deep learning design models, eliminating the need for model design, hardware, and software expertise. Hacarus Inc.’s HACARUS-X AI technology, which combines sparse modeling and machine learning technology, features low power consumption and small devices such as FPGAs. In collaboration with semiconductor trading company PALTEK, Hacarus is integrating HACARUS-X algorithms with Xilinx's FPGA Zynq UltraScale + MPSoC. Both companies area also implementing HACARUS-X algorithms in a box computer.Sparse modeling is gaining attention as a modeling method by which humans can understand the judgment process of AI by extracting features from a small amount of learning data. With expertise in life science fields such as medical and biology and image processing technology, LPixel, Inc. develops image analysis systems with original algorithms and machine learning techniques. It has developed a cloud-based AI image analysis platform and an AI medical image diagnosis support technology that streamlines the review of large amounts of research data and detects image fraud in research papers and other documents for the medical and biology fields, freeing researchers to devote more time to their core work. Yoichiro Ando is a marketing director at SEMI Japan.

Marcellino Gemelli, director of global business development at Bosch Sensortec, will present at the upcoming MEMS Sensors Executive Congress on October 29-30, 2018 in Napa, Calif. SEMI’s Maria Vetrano caught up with Gemelli to give MSEC attendees a preview of Gemelli’s feature presentation.Sensor fusion — the integration of different types of sensors through software algorithms to increase overall system performance and/or reduce power consumption— has come a long way since its inception. In those early days, sensor fusion generally involved MEMS inertial sensors only. The advent of new sensor varieties, including environmental sensors, is making new use cases a reality. Gemelli will explore the ways in which the next generation of sensor fusion is improving autonomous mobility devices. SEMI: Why are environmental sensors important to autonomous mobility devices?Gemelli: When most of us think of autonomous systems, we think that they are driven by motion sensors and proximity sensors (e.g., radar, Lidar). When vertical location comes into play, however, in applications such as drones or asset tracking, pressure sensors become an integral part of flight control, navigation and positioning in GPS-challenged areas.While not commonly considered an electronically enabled sense, the ability to “smell” the environment opens new opportunities. The quality of a user’s experience with personal cleaning robots and robo-taxis are good examples of where we might want to enable scent detection.SEMI: I’ve never thought much about using sensors to detect smell. How would a robo-taxi or a cleaning robot benefit from scent detection?Gemelli: Fully autonomous cars will inevitably give rise to robo-taxis. In fact, last month Volvo announced its fully electric robo-taxi, and in March 2018 Waymo announced that Jaguar Land Rover’s SUV would join Fiat Chrysler’s Chrysler Pacifica minivans in its planned fleet of robo-taxis, so we may see robo-taxis in the U.S. within the next five years.With robo-taxis fast-approaching, we need technologies that provide the same level of oversight that a taxi driver once fulfilled. Gas sensors would function like an electronic nose (e-nose) in a robo-taxi to inform the taxi’s owner of prohibited passenger behavior, such as eating, drinking or smoking in the vehicle, which could potentially damage the vehicle’s interior. Camera sensors could record the act as proof of the offense.Cleaning robots would be more sophisticated than they are today. In addition to leveraging image and range-finding sensors to more accurately map the rooms in your house, they could also detect scents from spilled red wine, pet urine or other foreign materials. When the cleaning robot, such as a vacuum, detects the foreign substance, it would navigate around the substance instead of going through it and spreading it all over the carpet.In addition to robo-taxis and cleaning robots, I will also discuss asset tracking and drones.SEMI: What role does sensor fusion play in autonomous mobility devices?Gemelli: Combining sensor fusion with artificial intelligence (AI) will generate new use cases and therefore new markets for sensor suppliers.There is another major benefit as well. With so many connected devices in our lives — including those with cameras, location awareness and always-listening capabilities — we are seeing growing concern about user privacy. Sensor fusion and AI can help to alleviate this concern: By supporting more local processing, they allow for greater control of data, safeguarding personal privacy.SEMI: Who is responsible for the AI part of the sensor-fusion equation?Gemelli: AI is a new frontier for MEMS and sensors suppliers. It benefits us and our customers to embrace AI algorithms through in-house development and/or partnerships.SEMI: What would you like MEMS Sensors Executive Congress attendees to take away from your presentation?Gemelli: I plan to issue a call to action to increase research in hybrid sensor-fusion software architectures, including AI, as suppliers’ collaboration will benefit the industry at large.Marcellino Gemelli is currently based in Palo Alto (CA) responsible for business development of Bosch Sensortec's MEMS product portfolio. He received the ‘Laurea’ degree in Electronic Engineering at the University of Pavia, Italy while in the Italian Army and an MBA from MIP, the Milano (Italy) Polytechnic business school. He previously held various engineering and product management positions at STMicroelectronics from 1995 to 2011 in the fields of MEMS, electronic design automation and data storage. He was contract professor for the Microelectronics course at the Milano (Italy) Polytechnic from 2000 to 2002.Marcellino Gemelli will present Environmental Sensors Systems Enabling Autonomous Mobility on Tuesday, October 30 at MEMS Sensors Executive Congress in Napa Valley, Calif.Register today to learn more about the connection between sensor fusion, AI and next-generation autonomous mobility devices.Maria Vetrano is a public relations consultant at SEMI.

Even for someone who has been in this industry since the days of the TI Datamath 4-function calculator and the TMS1100 4-bit microcontroller (yes, that’s been a LONG time – the movie Grease premiered the same year!), it is sometimes hard to grasp the scope and complexity of what happens in today’s leading-edge semiconductor gigafabs. In fact, the only way to comprehend the enormous volume of transactions that occur is to consider what happens in a single minute – this is illustrated in the infographic we have labeled “The Gigafab Minute.”* It’s amazing enough to think that a single factory can start 100,000 wafers every month on their cyclical journey through 1500 process steps… and have 99%+ of them emerge 4 months later to be delivered to packaging houses and then on to waiting customers. It’s quite another to realize that all of this happens continuously (24 x 7) and automatically. “How is this possible?” you ask.Well, a big part of the solution is the body of SEMI standards which have evolved since the early 80s to keep pace with the ever-changing demands of the industry. From an automation standpoint, many of these standards deal with the communications between manufacturing equipment and the factory information and control systems that are essential for managing these complex, hyper-competitive global enterprises.A significant characteristic of these standards is that they have been carefully designed to be “additive.” This means that new generations of SEMI’s communications standards do not supplant or obsolete the previous generations, but rather provide new capabilities in an incremental fashion. To appreciate the importance of this in actual practice, consider how the GEM, GEM300, and EDA/Interface A standards support the transactions that occur in a single Gigafab Minute.Starting at 1:00 o’clock on the infographic and moving clockwise, you first notice that 2.31 wafers enter the line. Of course, these are actually released in 25-wafer 300mm FOUPs (Front-Opening Unified Pod), but 100K wafers per month translates to 2.31 per minute. Since these factories run continuously, once the line is full, it stays full. And with an average total cycle time of 4 months, this means that there are 400K wafers of WIP (work in process) in he factory at any given time. This number, and the total number of equipment (5000+), drive the rest of the calculations.GEM (Generic Equipment Model) – SEMI E30, etc.The GEM messaging standards were initially defined in the early 90s to support the factory scheduling and dispatching applications that decide what lots should go to what equipment, the automated material handling systems that deliver and pick-up material to/from the equipment accordingly, the recipe management systems that ensure each process step is executed properly, and the MES (Manufacturing Execution System) transactions that maintain the fidelity of the factory system’s “digital twin.”Every minute of every day, GEM messages support and chronicle the following activities: 240 process steps are completed (i.e., 240 25-wafer lots are processed), 300 recipes are downloaded along with a set of run-specific adjustable control parameters, and 600 FOUPs are moved from one place to another (equipment, stockers, under-track storage, etc.). For each of these activities, the factory’s MES is notified instantaneously.GEM300 – SEMI E40, E87, E90, E94, E157With the advent of 300mm manufacturing in the mid-to-late 90s, a global team of volunteer system engineers from the leading chip makers defined the GEM300 standards to support fully automated manufacturing operations. Starting at 5:00 o’clock on the infographic, the number of transactions per minute jumps almost 3 orders of magnitude, from the monitoring of 900 control jobs across 4000 process tools to the tracking of 360,000 individual recipe step change events. This level of event granularity is essential for the latest generation of FDC (Fault Detection and Classification) applications, because precise data framing is a key prerequisite for minimizing the false alarm rate while still preventing serious process excursions. In this context, more than 6000 recipe-, product- and chamber-specific fault models may be evaluated every minute.Simultaneously, the applications that monitor instantaneous throughput to prevent “productivity excursions” and identify systemic “wait time waste” situations depend on detailed intra-tool wafer movement events. In a fab with hundreds of multi-chamber, single-wafer processes, 75,000 or more of these events occur every minute. EDA (Equipment Data Acquisition) – SEMI E120, E125, E132, E134, E164, etc.Rounding out the SEMI standards in our example gigafab is the suite of EDA standards which complement the command and control functions of GEM/GEM300 with flexible, high-performance, model-based data collection. The EDA standards enable the on-demand collection of the volume and variety of “big data” required from the equipment to support the advanced analysis, machine learning, and other AI (Artificial Intelligence) applications that are becoming increasingly prevalent in leading semiconductor manufacturers. As EUV (Extreme Ultraviolet) lithography moves from pilot production to high-volume manufacturing at the 7nm process node and beyond, the litho process area will become a major source of process data by itself, generating 10 GB of data every minute. This is in addition to the 100 GB of data collected from other process areas. The End ResultThe final wedge (12:00 o’clock) in our infographic highlights the real objective – which is producing the millions of integrated circuits that fuel our global economy and provide the technologies that are an integral part of our modern way of life. Assuming a nominal die size of 50 square mm (typical of an 8 GB DRAM), the 2.31 wafers we started at 1:00 o’clock result in almost 3200 individual chips. But none of this would be possible without the pervasive factory automation technology we now take for granted. So, as you finish reading this posting on whatever device you happen to be using, take a micro-moment to acknowledge and thank the hundreds of standards volunteers whose insights and efforts made this a reality!You may not be responsible for running a gigafab anytime soon, but the SEMI standards used in this setting are no less applicable to any Smart Manufacturing environment. Give us a call if you’d like to know more about how these technologies can benefit your operations for many years to come.Alan Weber is Vice President, New Product Innovations, at Cimetrix Incorporated. Previously he served on the Board of Directors for eight years before joining the company as a full-time employee in 2011. Alan has been a part of the semiconductor and manufacturing automation industries for over 40 years. He holds bachelor’s and master’s degrees in Electrical Engineering from Rice University. For more information on SEMI Standards, please click here.

SEMI FabView update for calendar year Q3 2018 Global fab construction investment shows continuing strength, with 19 new fab projects expected to begin construction in 2019 and 2020, based on the latest data published in SEMI’s World Fab Forecast. Fab investment is just one indicator of how growing demand in areas such as high-performance computing, data storage, artificial intelligence (AI), cloud computing, and automotive are driving the fourth consecutive year of spending growth in the semiconductor industry. Below are a few highlights* from September’s SEMI FabView: Memory: Not fading Micron plans to invest $3 billion by 2030 in Manassas, Virginia – These investments, driven by strong demand for automotive applications, are contemplated in Micron's long-term model. The production ramp is anticipated to be in the first half of 2020. SK Hynix to build new DRAM fab in Icheon (Gyeonggi Province), Korea – The construction, to be completed by the end of 2020, will adopt 1znm node (probably EUV). Total investment is estimated to exceed $13 billion. Nanya Technology doubles 2018 capex plan – The increase is for additional DRAM capacity and more 20nm DRAM conversion (from 30nm). 200mm and below: Not leading edge, but continues to draw investment Vanguard changes fab investment strategy – Vanguard will focus on 200 mm and has scrapped its plan for 300mm expansion. Murata to invest into 150mm expansion – Murata announced a 5 billion Yen investment (US$44.6 million) in a new fab extension in Vantaa, Finland. Investment, M A in Analog, Logic, Power and Opto Segments Texas Instruments is looking to invest $3.2 billion in new fab construction in 2019 – Texas Instruments is eyeing Richardson, Texas and also considering sites outside Texas. Bosch 300mm fab in Dresden, Germany – Bosch held a groundbreaking ceremony on April 24. Equipment installation is expected in 2H19. Microchip completes acquisition of Microsemi – Microchip closed its $8.45 billion acquisition of Microsemi on May 29. Microsemi has five fabs in the U.S. with a wide range of semiconductor products and system solutions. New fabs in China keep on coming Shanghai Jita Semiconductor/Huada Semiconductor – Shanghai Jita Semiconductor, a subsidiary of Huada Semiconductor and China Electronics Corporation (CEC), announced plans earlier this month to build both 200 mm and 300 mm semiconductor fabs for analog and power semiconductors in Shanghai. The combined fab investment will total $5.18 billion. Hamamatsu Photonics building 200 mm fab – Hamamatsu announced that it is building a new facility Investment of 2.8 billion Yen (US$25 million) to boost opto semiconductor capacity. Production is anticipated to start in late 2019. * Actual FabView updates provide more detail SEMI FabView, a mobile-friendly, interactive version of SEMI’s popular World Fab Forecast, delivers on-demand fab information such as fab spending and capacity for over 1,200 facilities, including over 60 planned facilities worldwide, across a wide range of product segments including Power, GPU, Memory, Foundry, MEMS and Sensors fabs. Fab data include region, start of construction, operation, construction and equipment spending, capacity, wafer sizes, product types and geometries. SEMI FabView subscribers receive forecast model updates through SEMI’s World Fab Database. Click here for a trial if you want to experience SEMI FabView first hand. Christian G. Dieseldorff is senior principal analyst and Eugenia Liu is senior product marketing manager, Industry Research and Statistics, SEMI, Milpitas, California.

Over the past three decades, most of the world’s innovations have centered largely on business models and involved iterative advances of existing technologies, with none matching the global impact of the top 10 semiconductor industry discoveries and advances, Dr. Morris Chang, founder of TSMC and the IC foundry model, said at SEMICON Taiwan 2018 this week.Few have as clear a perspective on the transformative power of semiconductors as Dr. Chang, founder of TSMC and father of the IC foundry model. Keynoting the IC60 Master Forum celebrating the 60th anniversary of the invention of the integrated circuit (IC), Dr. Chang listed what he considers the 10 key semiconductor industry innovation milestones since 1948:1. Invention of the transistor by Shockley, Bardeen, and Brattain – 19482. Silicon transistor – 19543. Integrated circuit – 19584. Moore’s Law – 19655. MOS technology MOS FET – 1964 Silicon gate – 1967 CMOS – 1970 6. Memory DRAM – 1966 Flash – 1967 7. Outsourced assembly and test (OSAT) – 1960s8. Microprocessor – 19709. VLSI systems design – 1970-1980 IP and design tools – 1980-present 10. Foundry model – 1985 Among the most consequential semiconductor advances may be yet to come, Dr. Chang said, citing innovations including artificial intelligence (AI) and machine learning, new device architectures, Extreme Ultraviolet lithography (EUV), 2.5D/3D packaging, and new materials such as graphene and carbon nanotubes.Dr. Chang argued that because bringing an innovation into production is immensely more expensive than proving a theory in a lab, innovators are not always the ones to implement and benefit from their novel ideas. Today, innovation costs are skyrocketing, driving more consolidation across the supply chain.Michael Droeger is director of marketing at SEMI.

2017 was a good year for the MEMS and sensors business, and that upward trend should continue. We forecast extended strong growth for the sensors and actuators market, reaching more than $100 billion in 2023 for a total of 185 billion units. Optical sensors, especially CMOS image sensors, will have the lion’s share with almost 40 percent of market value. MEMS will also play an important role in that growth: During 2018–2023, the MEMS market will experience 17.5 percent growth in value and 26.7 percent growth in units, with the consumer market accounting for more than 50 percent(1) share overall. Evolution of SensorsSensors were first developed and used for physical sensing: shock, pressure, then acceleration and rotation. Greater investment in R D spurred MEMS’ expansion from physical sensing to light management (e.g., micromirrors) and then to uncooled infrared sensing (e.g., microbolometers). From sensing light to sensing sound, MEMS microphones formed the next wave of MEMS development. MEMS and sensors are entering a new and exciting phase of evolution as they transcend human perception, progressing toward ultrasonic, infrared and hyperspectral sensing.Sensors can help us to compensate when our physical or emotional sensing is limited in some way. Higher-performance MEMS microphones are already helping the hearing-impaired. Researchers at Arizona State University are among those developing cochlear implants — featuring piezoelectric MEMS sensors — which may one day restore hearing to those with significant hearing loss. The visually impaired may take heart in knowing that researchers at Stanford University are collaborating on silicon retinal implants. Pixium Vision began clinical trials in humans in 2017 with its silicon retinal implants.It’s not science fiction to think that we will use future generations of sensors for emotion/empathy sensing. Augmenting our reality, such sensing could have many uses, perhaps even aiding the ability of people on the autism spectrum to more easily interpret the emotions of others.Through my years in the MEMS industry, I have identified three distinct eras in MEMS’ evolution: The “detection era” in the very first years, when we used simple sensors to detect a shock. The “measuring era” when sensors could not only sense and detect but also measure (e.g., a rotation). The “global-perception awareness era” when we increasingly use sensors to map the environment. We conduct 3D imaging with Lidar for autonomous vehicles. We monitor air quality using environmental sensors. We recognize gestures using accelerometers and/or ultrasonics. We implement biometry with fingerprint and facial recognition sensors. This is possible thanks to sensor fusion of multiple parameters, together with artificial intelligence. Numerous technological breakthroughs are responsible for this steady stream of advancements: new sensor design, new processes and materials, new integration approaches, new packaging, sensor fusion, and new detection principles.Global Awareness SensingThe era of global awareness sensing is upon us. We can either view global awareness as an extension of human sensing capabilities (e.g., adding infrared imaging to visible) or as beyond-human sensing capabilities (e.g., machines with superior environmental perception, such as Lidar in a robotic vehicle). Think about Professor X in Marvel’s universe, and you can imagine how human perception could evolve in the future! Some companies envisioned global awareness from the start. Movea (now part of TDK InvenSense), for example, began their development with inertial MEMS. Others implemented global awareness by combining optical sensors such as Lidar and night-vision sensors for robotic cars. A third contingent grouped environmental sensors (gas, particle, pressure, temperature) to check air quality. The newest entrant in this group, the particle sensor, could play an especially important role in air-quality sensing, particularly in wearable devices.Driven by increasing societal concern over mounting evidence of global air-quality deterioration, air pollution has become a major topic in our society. Studies show that there is no safe level of particulates. Instead, for every increase in concentration of PM10 or PM2.5 inhalable particles in the air, the lung cancer rate is rising proportionately. Combining a particle sensor with a mapping application in a wearable could allow us to identify the locations of the most polluted urban zones.The Need for Artificial Intelligence To realize global awareness, we also need artificial intelligence (AI), but first, we have challenges to solve. Activity tracking, for example, requires accurate live classification of AI data. Relegating all AI processing to a main processor, however, would consume significant CPU resources, reducing available processing power. Likewise, storing all AI data on the device would push up storage costs. To marry AI with MEMS, we must do the following: Decouple feature processing from the execution of the classification engine to a more powerful external processor. Reduce storage and processing demands by deploying only the features required for accurate activity recognition. Install low-power MEMS sensors that can incorporate data from multiple sensors (sensor fusion) and enable pre-processing for always-on execution. Retrain the model with system-supported data that can accurately identify the user’s activities. There are two ways to add AI and software in mobile and automotive applications. The first is a centralized approach, where sensor data is processed in the auxiliary power unit (APU) that contains the software. The second is a decentralized approach, where the sensor chip is localized in the same package, close to the software and the AI (in the DSP for a CMOS image sensor, for example). Whatever the approach, MEMS and sensors manufacturers need to understand AI, although they are unlikely to gain much value at the sensor-chip level.Heading to an Augmented WorldWe have achieved massive progress in sensor development over the years and are now reaching the point when sensors can mimic or augment most of our perception: vision, hearing, touch, smell and even emotion/empathy as well as some aesthetic senses. We should realize that humans are not the only ones to benefit from these developments. Enhanced perception will also allow robots to help us in our daily lives (through smart transportation, better medical care, contextually aware environments and more). We need to couple smart sensors’ development with AI to further enhance our experiences with the people, places and things in our lives.About the authorWith almost 20 years’ experience in MEMS, sensors and photonics applications, markets, and technology analyses, Dr. Eric Mounier provides in-depth industry insight into current and future trends. As a Principal Analyst, Technology Markets, MEMS Photonics, in the Photonics, Sensing Display Division, he contributes daily to the development of MEMS and photonics activities at Yole Développement (Yole). He is involved with a large collection of market and technology reports, as well as multiple custom consulting projects: business strategy, identification of investment or acquisition targets, due diligence (buy/sell side), market and technology analyses, cost modeling, and technology scouting, etc.Previously, Mounier held R D and marketing positions at CEA Leti (France). He has spoken in numerous international conferences and has authored or co-authored more than 100 papers. Mounier has a Semiconductor Engineering Degree and a PhD in Optoelectronics from the National Polytechnic Institute of Grenoble (France).Mounier is a featured speaker at SEMI-MSIG European MEMS Sensors Summit, September 20, 2018 in Grenoble, France. (1) Source: Status of the MEMS Industry report, Yole Développement, 2018

FD-SOI was a very important topic during the recent Mount Qingcheng China IC Ecosystem Forum. To situate things, Mount Qingcheng, with its lush hills and waterways, is located just outside of Chengdu. That of course is where GlobalFoundries is building its new fab, which will be the first in China to run FD-SOI. Chengdu is also a key city in China's automotive electronics landscape. [caption id="attachment_12236" align="alignright" width="300"] (Image Courtesy: VeriSilicon)[/caption] The theme of the forum was Building a Smart Automotive Electronics Industry Chain. Over 260 decision-makers from government, academia and industry attended – and the SOI Consortium had a significant presence. The event was chaired by Wayne Dai, CEO/Founder of consortium member VeriSilicon, and tireless champion of the the FD-SOI ecosystem in China and worldwide. Morning keynotes were given by: Carlos Mazure, Soitec CTO and SOI Consortium Executive Co-Director; Mark Granger, GF's VP of Automotive Product Line Management; and Tony King-Smith, Executive Advisor at AImotive, a GF 22FDX customer. BTW, transcripts of all the talks are available through Gasgoo, China's largest automotive B2B marketplace. You can click here to access them. (They're in Chinese – but you can open them in the language of your choice using the major translation websites.) Chengdu Officials Affirm Support for FD-SOI Fan Yi, Deputy Mayor of Chengdu, spoke extensively of FD-SOI in his keynote on the importance of rapidly developing smart cars. He heralded the “spectacular” new GlobalFoundries fab there. Following a meeting with the company's top brass the day before, he affirmed GF's confidence in their investment. There is a solid roadmap for FD-SOI, he noted, and efforts are underway to accelerate the move into production and expand education and training. He cited the benefits of FD-SOI for the entire supply chain, from design through package and test, raising the level of the entire IC industry to new heights. The government, he said, attaches great importance to this enterprise. Their thinking regarding intelligent transport in China is integrated with the overall approach to smart cities. SOI Consortium Leads Industry Keynotes [caption id="attachment_12232" align="alignleft" width="300"] Wayne Dai, VeriSilicon Founder and CEO (Photo courtesy VeriSilicon)[/caption] In his opening remarks, Wayne Dai emphasized the need for China to seize the advantage in the next round of development opportunities in the automotive electronics industry. This year's Qingcheng forum, he noted, brought together key representatives from across the supply chain, from of the highest to the deepest reaches of the smart car electronics industry, and across markets, technologies, solutions, industrial ecosystem, standards and regulations. In his talk on how FD-SOI is boosting the accelerated development of automotive electronics, Carlos Mazure presented the SOI Industry Consortium. He noted that the Consortium promotes mutual understanding and development across the ecosystem. SOI is already present throughout automotive applications, he noted. There are currently about 100mm2 of SOI per car, in such diverse areas power systems, transmissions, entertainment, in-vehicle networking and more. SOI will experience especially high growth in electrification, information/entertainment, networking, 5G, AI/edge computing and ADAS. He then went on to give some history and an extensive overview of the major trends and highlights we've seen over recent years. He finished by giving examples of convergence across the supply chain with IC manufacturers working with automakers to lower power, increase processor performance and advance 5G. [caption id="attachment_12233" align="alignright" width="665"] Carlos Mazure, Soitec CTO and SOI Consortium Executive Co-Director; Tony King-Smith, Executive Advisor at AImotive and Mark Granger, GF's VP of Automotive Product Line Management (Photo courtesy VeriSilicon)[/caption] GF's Mark Granger addressed the rapid development of automotive electronics. In certain areas, he said, he sees growth rates of over 20%. They are working on building the Chengdu ecosystem, especially for design, and in cooperation with the rest of the supply chain. Furthermore, he reminded the audience, when you talk about cars, travel implies that you also talk about IoT as well as things like infotainment and integrated radar ICs. In addition to cost and power efficiencies, the AEC-Q100 standard for IC reliability in automotive applications is also pushing designers to turn to FD-SOI. In the GF meeting with Chengdu government officials (referenced above in deputy mayor Fan Yi's talk), he too confirmed their support of FD-SOI as a key technology for China. GF is currently cooperating with about 75 automotive partners, he said, and the company is looking to increase cooperation with partners in the Chengdu region. Tony King-Smith talked about the 22FDX test chip AImotive is doing with Verisilicon and GF. In case you missed it, in June 2017 AImotive announced its AI-optimized hardware IP was available to global chip manufacturers for license. AiWare is built from the ground up for running neural networks, and the company says it is up to 20 times more power efficient than other leading AI acceleration hardware solutions on the market. In the same announcement, they revealed that VeriSilicon would be the first to integrate aiWare into a chip design,and that aiWare-based test chips would be fabricated on GF's 22FDX. The chip is expected to debut this year. While the afternoon agenda was not specific to FD-SOI, it did focus on the "smart cockpit" and "intelligent driving", with talks by nine leading players in China's automotive IC and investment communities. ~ ~ ~Note: Many thanks to the folks at VeriSilicon, who wrote up this event for their WeChat feed, and shared photos with us here at ASN.

The march to greater precision, efficiency and safety – the lifeblood of high-technology manufacturing facilities – has taken on a new urgency as emerging applications such artificial intelligence (AI), the Internet of Things (IoT) and Industry 4.0 give new meaning to smart factories. Facing fiercer competition and ever more sophisticated fabrication processes, semiconductor fabs are under intense pressure to keep pace with new technologies as they work to upgrade. Nowhere are the stakes higher than in Taiwan, where high-tech manufacturing contributes mightily to the region’s GDP growth. To help Taiwan fabs confront the challenges and opportunities of designing smarter factories, SEMI and its High-Tech Facility Committee hosted the High-Tech Facility Workshop in June. SEMICON Taiwan 2018 High-Tech Facility Pavilion exhibitors gathered to explore how they can build smarter factories by deploying smart surveillance and disaster prevention technologies along with smart communications systems that better use manufacturing data to drive new safety and product quality efficiencies.During the workshop, SEMI High-Tech Facility Committee representatives shared strides it has made upgrading overseas facilities and developing standards to help establish smart factories in Taiwan.SEMICON Taiwan – 5-7 September at Taipei’s Nangang Exhibition Center – is also an important event for advancing smart manufacturing in Taiwan. Nearly 30 leading global manufacturers will exhibit at the SEMICON Taiwan High-Tech Facility Pavilion. The venue covers operational aspects of semiconductor manufacturing vital to becoming smarter including energy savings, nano-contamination control, facility information modeling, precision instrumentation and control, fire protection, mechatronics, and automation control. The pavilion will also feature a series of theme events offering a comprehensive overview of topics including the latest practices for integrating smart facility capabilities from the perspective of an advanced fab designer.At the TechXPOT stage, High-Tech Facility Pavilion exhibitors will also demonstrate the latest technology breakthroughs and cutting-edge smart factor solutions.The September 6th High-Tech Facility International Forum at SEMICON Taiwan will again gather factory experts and thought leaders from industry and academia to examine “Effective Ways to Make a Facility Smart.“ Experts from industry heavyweights in the fields of wafer foundry, LCD, memory and semiconductor packaging including TSMC, UMC, Innolux, ASE, Micron Taiwan, Winbond and VIS will offer insights into key areas of high-tech facilities including facility electricity, machinery, water management, vaporization and automation systems. On the same day as the forum, the High-Tech Facility Get-Together and High-Tech Facility VIP Dinner will bring together industry elites, academic professionals, and government officials to explore partnership opportunities. SEMI Taiwan and the High-Tech Facility Committee share HTF market trends information, technology updates and standards with SEMI members and exhibitors. Founded in 2013, the High-Tech Facility Committee now has 85 corporate members. Dedicated to accelerating industry collaboration through the integration of Taiwan industrial, government and academic resources, the committee each year holds several group meetings focusing on topics including energy savings, earthquake and fire protection, nano-contamination control, and precision instrumentation and control to advance critical technologies and facilitate standardization. The committee also aims to help the industry become more competitive faster by promoting technology standards that boost productivity and reduce production costs.Please visit www.semi.org and www.semicontaiwan.org for more information about SEMI’s high-tech facility initiatives.Iris Tsou is a marketing specialist at SEMI Taiwan.

With artificial intelligence (AI) rapidly evolving, look for applications like voice recognition and image recognition to get more efficient, more affordable, and far more common in a variety of products over the next few years. This growth in applications will drive demand for new architectures that deliver the higher performance and lower power consumption required for widespread AI adoption. “The challenge for AI at the edge is to optimize the whole system-on-a-chip architecture and its components, all the way to semiconductor technology IP blocks, to process complex AI workloads quickly and at low power,” says Qualcomm Technologies Senior Director of Engineering Evgeni Gousev, who will provide an update on the progress of AI at the edge in a Data and AI program at SEMICON West, July 10-12 in San Francisco. Qualcomm Snapdragon 845 uses heterogeneous computing across the CPU, GPU, and DSP for power-efficient processing for constantly evolving AI models. Source: QualcommA system approach that optimizes across hardware, software, and algorithms is necessary to deliver the ultra-low power – to a sub 1-milliwatt level, low enough to enable always-on machine vision processing – for the usually energy-intensive AI computing. From the chip architecture perspective, processing AI workloads with the most appropriate engine, such as the CPU, GPU, and DSP with dedicated hardware acceleration, provides the best power efficiency – and flexibility for dealing with rapidly changing AI models and growing diversity of applications.“So far it’s been largely a brute force approach using conventional architectures and cloud-based infrastructure,” says Evgeni. “But we’re going to run out of brute force options, so future opportunities lie in developing innovative architectures, dedicated hardware, new algorithms, and new software. Innovation will be especially important for AI at the edge and applications requiring always-on functionality. Training is mostly in the cloud now, but in the near future it will start migrating to the device as the algorithms and hardware improve. AI at the edge will also remove some privacy concerns, an increasingly important issue for data collection and management.”Practical AI applications at the edge where resources are constrained run the gamut, spanning smartphones, drones, autonomous vehicles, virtual reality, augmented reality and smart home solutions such as connected cameras. “More AI on the edge will create a huge opportunity for the whole ecosystem – chip designers, semiconductor and device manufacturers, applications developers, and data and service providers. And it’s going to make a significant impact on the way we work, live, and interact with the world around us,” Evgeni said.Future generations of chips may need more disruptive systems-level change to handle high data volumes with low power A next-generation solution for handling the massive proliferation of AI data could be a nanotechnology system, such as the collaborative N3XT (Nano-Engineered Computing Systems Technology) project, led by H.S. Philip Wong and Subhasish Mitra at Stanford. “Even with next-generation scaling of transistors and new memory chips, the bottlenecks in moving data in and out of memory for processing will remain,” says Mitra, another speaker in the SEMICON West program. “The true benefits of nanotechnology will only come from new architectures enabled by nanosystems. One thing we are certain of is that massively more capable and more energy-efficient systems will be necessary for almost any future application, so we will need to think about system-level improvements.” Major improvement in handling high volumes of data with low high energy use will require system-level improvements, such as monolithic 3D integration of carbon nanotube transistors in the multi-campus N3XT chip research effort. Source: Stanford UniversityThat means carbon nanotube transistors for logic, high density non-volatile MRAM and ReRAM for memory, fine-grained monolithic 3D for integration, new architectures for computation immersed in memory, and new materials for heat removal. “The N3XT approach is key for the 1000X energy efficiency needed,” says Mitra.Researchers have demonstrated improvements in all these areas, including multiple hardware nanosystem prototypes targeting AI applications. The researchers have transferred multiple layers of as-grown carbon nanotubes to the target wafer to significantly improve CNT density and have also developed a low-power TiN/HfOx/Pt ReRAM. The low-temperature CNT and ReRAM processes enable multiple vertical layers to be grown on top of one another for ultra-dense and fine-grained monolithic 3D integration. Other speakers at the Data and AI TechXpot include Fram Akiki, VP Electronics, Siemens; Hariharan Ananthanarayanan, motion planning engineer, Osaro; and David Haynes, Sr. director, strategic marketing, Lam Research. See SEMICONWest.org.Paula Doe, SEMI