Milestone-Proposal:The Engineering Data Analysis System sNOVA, 1989-1997
To see comments, or add a comment to this discussion, click here.
Docket #:2023-16
This proposal has been submitted for review.
To the proposer’s knowledge, is this achievement subject to litigation? No
Is the achievement you are proposing more than 25 years old? Yes
Is the achievement you are proposing within IEEE’s designated fields as defined by IEEE Bylaw I-104.11, namely: Engineering, Computer Sciences and Information Technology, Physical Sciences, Biological and Medical Sciences, Mathematics, Technical Communications, Education, Management, and Law and Policy. Yes
Did the achievement provide a meaningful benefit for humanity? Yes
Was it of at least regional importance? Yes
Has an IEEE Organizational Unit agreed to pay for the milestone plaque(s)? Yes
Has an IEEE Organizational Unit agreed to arrange the dedication ceremony? Yes
Has the IEEE Section in which the milestone is located agreed to take responsibility for the plaque after it is dedicated? Yes
Has the owner of the site agreed to have it designated as an IEEE Milestone? Yes
Year or range of years in which the achievement occurred:
1989-1997
Title of the proposed milestone:
The Engineering Data Analysis System sNOVA, 1989-1997
Plaque citation summarizing the achievement and its significance:
sNOVA, an industrial optimization platform originated from Macronix, affects the automation of the entire semiconductor industry chain, making Taiwan a global semiconductor production base. With remarkable features such as shortening production abnormal debugging time, reducing wafer layer cost, improving product quality, and automatically adjusting the precision of production machine process parameters, sNOVA takes the lead in realizing the digital transformation of semiconductor manufacturing through artificial intelligence and big data technology.
Chinese version
源於旺宏的一個產業優化平台sNOVA，牽動著整個半導體產業鏈的自動化，也讓台灣成為全球半導體生產大本營。 sNOVA以縮短生產異常調試時間、降低晶圓層成本、提升產品質量、自動調整生產機器工藝參數精度等顯著特點，開創了半導體製造通過人工智能和大數據技術實現數位轉型的先河。
200-250 word abstract describing the significance of the technical achievement being proposed, the person(s) involved, historical context, humanitarian and social impact, as well as any possible controversies the advocate might need to review.
IEEE technical societies and technical councils within whose fields of interest the Milestone proposal resides.
In what IEEE section(s) does it reside?
IEEE Taipei Section
IEEE Organizational Unit(s) which have agreed to sponsor the Milestone:
IEEE Organizational Unit(s) paying for milestone plaque(s):
Unit: IEEE Taipei Section
Senior Officer Name: Pei-Wen Li
IEEE Organizational Unit(s) arranging the dedication ceremony:
Unit: IEEE Taipei Section
Senior Officer Name: Pei-Wen Li
IEEE section(s) monitoring the plaque(s):
IEEE Section: IEEE Taipei Section
IEEE Section Chair name: Pei-Wen Li
Milestone proposer(s):
Proposer name: Ruey-Beei Wu
Proposer email: Proposer's email masked to public
Please note: your email address and contact information will be masked on the website for privacy reasons. Only IEEE History Center Staff will be able to view the email address.
Street address(es) and GPS coordinates in decimal form of the intended milestone plaque site(s):
No. 16, Lixing Rd., East Dist., Hsinchu City , Taiwan (R.O.C.) GPS coordinates: x 24.80361 y 120.96861
Describe briefly the intended site(s) of the milestone plaque(s). The intended site(s) must have a direct connection with the achievement (e.g. where developed, invented, tested, demonstrated, installed, or operated, etc.). A museum where a device or example of the technology is displayed, or the university where the inventor studied, are not, in themselves, sufficient connection for a milestone plaque.
Please give the address(es) of the plaque site(s) (GPS coordinates if you have them). Also please give the details of the mounting, i.e. on the outside of the building, in the ground floor entrance hall, on a plinth on the grounds, etc. If visitors to the plaque site will need to go through security, or make an appointment, please give the contact information visitors will need.
The intended site is the exhibition hall on the first floor of the Macronix headquarters. It is the closest public accessible building from where sNOVA was originally developed.
Are the original buildings extant?
Yes
Details of the plaque mounting:
The intended location is an exhibition hall on the first floor of the headquarters of Macronix. A picture giving an overview. A picture giving an overview of the hall is shown below.
How is the site protected/secured, and in what ways is it accessible to the public?
The building is open on weekdays. The site is protected/secured by a security company and security cameras.
Who is the present owner of the site(s)?
Macronix
What is the historical significance of the work (its technological, scientific, or social importance)? If personal names are included in citation, include justification here. (see section 6 of Milestone Guidelines)
sNOVA leads global semiconductor manufacturing into the era of AI and Big Data
In 1989, Macronix took the lead in Taiwan and the world to employ statistics experts in semiconductor fabs. Combining statistics and semiconductor knowledge, Cheng-yung PENG (彭誠湧), a Statistical Process Control (SPC) engineer and head of the in-house SPC system project, led the team to develop the prototype of the engineering data analysis system sNOVA, which can quickly find problems that cause abnormal quality or low yield and ensure the highest quality products and services for customers.
As a pioneering system in Taiwan, sNOVA ushered in the era of AI and Big Data in global semiconductor manufacturing, and made Macronix a paperless fab from day one.
sNOVA had triggered other high-tech companies in Taiwan to learn the know-how and develop a system of their own to level up the quality of products.
Take Taiwan’s United Microelectronics Corp. and Taiwan Semiconductor Manufacturing Co., Ltd., which has the highest market share in wafer production in the world, for example, they both build computerized systems after Macronix, not only to boost in-house management efficiency, but also to integrate production lines. The breakthrough had paved the way for Taiwan to play a key role in global chip manufacturing.
C.C. Wei, president of TSMC, praised the mindset and practice of Macronix as “very innovative”. [1]
The system has effectively improved product defect rate from PPM to PPB
In the past, engineers in Macronix and in the semiconductor industry often had to spend more than a week to find the cause of error when they faced hundreds of complex semiconductor manufacturing processes. However, with sNOVA and its prototypes in the past, engineers can quickly detect the problem in just a few minutes. The sNOVA system has greatly improved the product quality yield rate to more than 500 PPB (the defect rate per billion is less than 500), Macronix has also become the first memory company in the world to measure product defect rate in parts per billion (PPB) rather than parts per million (PPM) (Fig. 1).
Meanwhile, because of the introduction of the sNOVA system, Macronix products are protected by AI, and the quality of the products is continuously improved. Even in the field of the aerospace industry, which is the most demanding environment, Macronix cooperated with Xilinx in 2015 to launch the industry's first commercial flash memory product with an extreme temperature range of -55°C to +125°C developed in response to the stringent requirements for high reliability in the aerospace and defense market.[2]
In the radiation resistance test of the product, the European Space Agency (ESA, European Space Agency) and Alter technology conducted a radiation resistance test on commercially available flash memory products during the Jupiter Icy Moons Explorer mission , Macronix's NOR Flash and NAND Flash products both won the first place, beating Micron, Kioxia and Infineon and other international manufacturers. The test results were turned into 4 papers and published in RADECS (RADiation and Its Effects on Component and Systems) International Symposium in 2019 and 2021 respectively. [3]-[6]
sNOVA is the Taiwan’s pioneer of AI and Industry 4.0 in the semiconductor industry
The EDA system sNOVA developed more than 30 years ago, is a pioneer in industrial automation, and it is also a pioneer in introducing statistical methods and artificial intelligence in the semiconductor industry. Through decades of implementation and continuous upgrades, the spirit and concept contained in the sNOVA system can be regarded as the starting point of the development of Industry 4.0 in the semiconductor industry in Taiwan.
Take the defect automatic classification system as an example, from taking photos, sampling, classification to removing bad ICs with defects, the whole process is completely intelligent. It not only completes the communication, connection and data integration with existing machines or systems in the semiconductor industry, but also standardizes the way engineers work, further developing AI methods to replace the traditional processes [7].
Compared with the traditional manual operation, the current results have increased the accuracy by four times, with the amount of defect detection being increased by more than 10 times, and the work efficiency being increased by more than 10 times. This operation mode is exactly the Industry 4.0 that becomes more and more popular nowadays.
Inviting statistical professionals to invest in semiconductor industry applications
More than 30 years ago, for students who graduated from statistics-related departments, most of them went into financial and insurance-related fields. At that time, it was unheard of that statistics graduates worked in semiconductor factories. However, the development of several generations of sNOVA system has initiated the demand for statistical talents in the semiconductor industry, and it has also opened up new application fields for graduates of statistics-related departments. Cheng-yung PENG, a statistics expert and a SPC engineer at Macronix, endeavored to educate and recruit statistics talents from campuses that he had introduced and promoted the concepts in several universities in Taiwan. [8]
As Macronix successfully introduced data analysis and statistics into production lines, sNOVA ensures that innovative methods can be effective for engineering problems in terms of computer simulation and engineering experiment verification. For example, multivariate analysis, regression analysis, reliability analysis, statistical verification, etc. In particular, the semiconductor industry is very important to Taiwan's economic development. The government had also invested a lot of resources and attached great importance to it.
What obstacles (technical, political, geographic) needed to be overcome?
Overcoming the challenges of semiconductor’s 600-700 complex processes
It is unprecedented that Macronix overcomes the challenges of 600-700 complex semiconductor manufacturing processes. The early semiconductor production line used handwriting to write data such as the flows and parameters of each machine operation on "paper cards, however, the semiconductor manufacturing process often involved 600-700 complicated procedures. The traditional method is not only inefficient, but also prone to errors. When global fabs were still operating their production lines in the traditional way, Macronix developed the previous generations of sNOVA system. With no precedent to follow and no established model to refer to, it was 100% independently developed and successfully combined 600-700 complex semiconductor manufacturing processes, demonstrating the groundbreaking AI innovation capabilities of sNOVA system. Therefore, in 1992, after three years of development of the prototype of sNOVA, Macronix was awarded the "Award on Industrial Automation" by the Chinese National Federation of Industries in 1992 (Fig. 2) [9].
Allowing statistics, IT and engineering talents to cooperate effectively
Another obstacle in building up sNOVA is to make statistics and semiconductors who are not familiar with each other to cooperate. When statistical talents first entered the fab, it’s hard for them to communicate with colleagues of IT and engineering backgrounds. However, after continuous adaptation, Macronix gradually found the common ground between the parties. In the meantime, while the sNOVA system was continuously upgraded and optimized and production data accumulated over the first three to five years being utilized, more positive effects were brought about.
In 1989, Macronix established the General Engineering Department - Engineering Data Analysis Section, and in 1999, the company integrated the team into the Department of Data Value Development. The extension and optimization of the teams present how statistics and semiconductors could be implemented perfectly altogether.
What features set this work apart from similar achievements?
In recent years, the technology of AI and Big Data has developed vigorously. There are many products that can provide AI technology in the market, but the technology of AI itself is not the key factor to the success of the sNOVA system. "Sustainable management”, “generating value” and “integrating workflow with environment” are the three main reasons why the sNOVA system has been able to promote effective operation for 30 years since its inception.
Sustainable development
The AI Model and method of the sNOVA system can adapt to Macronix’s process technology innovation, and quickly adjust to changes in machine conditions to ensure that the AI measures continue to be correct and helpful. The sNOVA system has millions of various AI Models operating on the assembly line. It actively detects abnormal models in an automatic and effective way and corrects them in time. It can also be updated quickly in conjunction with changes in process technology or machine conditions, ensuring sNOVA grows with manufacturing technology and continues to contribute to product quality.
With the rapid development of information technology in the market, "how to upgrade the technology of information development in AI system" is not only a challenge in sNOVA's investment decision, but also has a direct impact on the improvement of sNOVA's computing speed and computing cost. The development of sNOVA adopts the most advanced component design. An AI function may have millions of lines of code, but through the component design, only a thousand lines of code related to daily maintenance and operation are required, which greatly saves development time and enables the reference to be more flexible.
The early development of sNOVA evolved from IBM's IT technology software - NOVA system. After more than 20 years of continuous research and upgrading, it has become a system developed by Macronix completely in-house.
Creating Value
To avoid the development of AI being mere formality or only for short-term effects, when designing and developing the sNOVA system, Macronix pays great attention to practical application effects. In addition to making engineers willing to use this system due to the simplicity of the operation interface and workflow, the powerful effects produced after sNOVA's intervention in production also make this system naturally integrated into the entire production process. Every IC produced by Macronix today is intervened and contributed significantly by AI. The sNOVA system is the core system in Macronix's wafer manufacturing and also Macronix's key manufacturing competence.
Ensuring that AI technology contributes to product quality is Macronix’s essential spirit when developing the sNOVA system, and it is also a key factor for the success of any company that is dedicated to developing an AI system. The relationship between the sNOVA system and product quality lies in the fact that it has been determined as the key issue in the production process in the beginning, while all aspects such as timely feedback on the results after the use of the production line must be considered, preventing AI development become a mere formality or just a short-term result. This is also the most important feature that distinguishes the sNOVA system from other AI systems.
In addition, a company must invest considerable manpower and material resources in the long term to develop AI. If the system cannot continue to contribute, the effectiveness will be greatly reduced, which indirectly affects a company’s willingness to persist on investing in AI systems. But the difficulty lies in that the AI system is not identical to other information systems, it cannot be completed once and for all. The system must be adjusted in a timely manner in accordance with machine conditions and process changes, otherwise AI will not be able to continue to create value.
Deep integration with workflow and environment
In response to the needs of different environments, "customization" is a necessary task for the development of AI. However, sNOVA's customization is not limited to the user-friendly that traditional information systems focus on, but requires a deeper integration with the workflows and work environment so that AI can be naturally integrated into daily work and become a part of life.
Like the traditional defect images, it is necessary to take pictures by the machines and then carry out human identification. When abnormal images are found, the goods must be treated specially, and a quality abnormality list must be submitted to the engineer to handle the abnormal goods. The sNOVA defect image recognition system can not only automatically judge the image, but also naturally integrate it into the process: after the machine takes a picture, it will automatically start the recognition, identify the abnormal situation, automatically drive the special goods system to process, and automatically issue a quality abnormality list and notify the engineer.
The sNOVA system has been developed for more than 30 years, integrating with various fields in the fab, and has established a natural way to operate AI. Engineers use sNOVA to analyze and measure, different departments use sNOVA to communicate, supervisors use sNOVA to review projects and make decisions, new employees use sNOVA to undergo educational training. As sNOVA becomes a part of engineers' work, it thoroughly realizes the essence of AI data management. The integration requires sNOVA to not only be equipped with general AI system technology, but also covered skills such as ERP, MES, IoT, Tool Automation, etc.
Supporting texts and citations to establish the dates, location, and importance of the achievement: Minimum of five (5), but as many as needed to support the milestone, such as patents, contemporary newspaper articles, journal articles, or chapters in scholarly books. 'Scholarly' is defined as peer-reviewed, with references, and published. You must supply the texts or excerpts themselves, not just the references. At least one of the references must be from a scholarly book or journal article. All supporting materials must be in English, or accompanied by an English translation.
1) C.-F. Chien, S.-C. Hsu*, and J.-F. Deng, “A cutting algorithm for optimizing the wafer exposure pattern,” IEEE Trans. Semiconductor Manuf., vol. 14, no. 2, pp. 157-162, May 2001. (Submitted: Aug. 5, 2000, *: employee of the Macronix Int. Co.)
The semiconductor manufacturing industry competes by increasing yield and lowering die costs, thereby taking advantage of significant capital investments. Many studies focus on defect reduction to improve yield rate. However, the problem of optimizing wafer exposure patterns has received little attention. In this paper, given the specific patterning constraints, we develop a two-dimensional (2-D) cutting algorithm to maximize the gross die yields of the eight-inch wafer and larger circular wafers. The empirical results that we implemented in a wafer fabrication factory in Taiwan validate the practical viability of this approach. Similar approaches can readily be applied to other wafer patterning.
2) C.-F. Chien, S.-C. Hsu*, and C.-P. Chen*, “Procedure of alignment for optimal wafer exposure pattern,” US Patent 6,368,761 B1, Filed: May 9, 2000, Issued: Apr. 9, 2002. (Assignee: Macronix Int. Co.)
Conventionally, efforts to improve the yield of chips produced on a wafer focused on defect reduction. Another approach is optimizing wafer exposure patterns. The present invention includes a computer-based procedure and apparatus to expose cells on the surface of a wafer so as to maximize the number of dies produced from a wafer. The invention is useful in the exposure of six- and eight-inch wafers, as well as larger wafers.
3) C.-F. Chien, K.-H. Chang, and C.-P. Chen*, “Design of a sampling strategy for measuring and compensating for overlay errors in semiconductor manufacturing,” Int. J. Prod. Res., vol. 41, no. 11, 2547–2561, 2003.
To enhance the resolution and alignment accuracy in semiconductor manufacturing, it is important to measure overlay errors and control them into tolerances by removing assignable causes. A number of related studies have been done to examine the factors causing the overlay errors, propose mathematical models, and develop overlay error control methods. However, the involved sampling strategies received little attention. This study aimed to propose specific designs of sampling patterns effectively to measure and compensate for overlay errors within the limited number of samples in practice. To verify the validity of the proposed approach, the sampling strategies were compared using empirical data from a wafer fabrication facility. The proposed sampling patterns had higher goodness of fit for the overlay model and lower residuals after compensation. This paper concludes with our findings and discussions on further research.
4） C.-F. Chien, K.-H. Chang, C.-P. Chen*, and S. L. Lin, “Overlay error model, sampling strategy and associated equipment for implementation,” US Patent 6,975,974, B2, Filed: Aug. 1, 2001, Issued: Dec. 13, 2005. (Assignee: Macronix Int. Co.)
In manufacturing VLSI circuits, overlay production is a critical step. To obtain a higher resolution and alignment accuracy in the microlithographic process, overlay errors must be measured so that overlay errors can be reduced to a tolerable level. This invention provides an overlay error model and a sampling strategy. Utilizing the overlay model and sampling strategy, a device for measuring overlay errors is also designed.
5） C.-J. Kuo, C.-F. Chien, and J.-D. Chen*, “Manufacturing intelligence to exploit the value of production and tool data to reduce cycle time,” IEEE Trans. Automat. Sci. Eng., vol. 8, no. 1, pp. 103-111, Jan. 2011. (Submitted: June 17, 2009, Best Paper Award of the Year by IEEE RA-S)
Cycle time reduction is crucial for semiconductor wafer fabrication companies to maintain competitive advantages as the semiconductor industry is becoming more dynamic and changing faster. According to Little’s Law, while maintaining the same throughput level, the reduction in Work-in-Process (WIP) will result in cycle time reduction. On one hand, the existing queueing models for predicting the WIP of tool sets in wafer fabrication facilities (fab) have limitations in real settings. On the other hand, little research has been done to predict the WIP of tool sets with tool dedication and waiting time constraints so as to control the corresponding WIP levels of various tool sets to reduce cycle time without affecting throughput. This study aims to fill the gap by proposing a manufacturing intelligence (MI) approach based on neural networks (NNs) to exploit the value of the wealthy production data and tool data for predicting the WIP levels of the tool sets for cycle time reduction. To validate this approach, empirical data were collected and analyzed in a leading semiconductor company. The comparison results have shown the practical viability of this approach. Furthermore, the proposed approach can identify and improve the critical input factors for reducing the WIP to reduce cycle time in a fab.
Note to Practitioners—Time to market and wafer fabrication cycle time is critical for maintaining a competitive advantage for a semiconductor manufacturing company. Although a number of approaches have been developed for cycle time reduction, little research has been done to exploit the value of the data of various toolsets. Focused on real settings, this study provides a novel method of manufacturing intelligence that employed NNs to derive empirical rules from huge production and tool data that are automatically collected in the semiconductor manufacturing process for managing WIP levels for corresponding toolsets. This approach was validated in an Integrated Device Manufacturer (IDM) focused on a niche market in Taiwan and has been implemented online.
References
[1] The 5th Taiwan Presidential Innovation Award in the individual category, 2022.
In 2022, Miin Wu, chairman of Macronix, won the 5th Taiwan Presidential Innovation Award in the individual category as the first to combine statistics and data mining methods to improve semiconductor process technology, production efficiency and shorten product development time. In an interview in 2022, C.C. Wei, CEO at Taiwan Semiconductor Manufacturing Co. (TSMC), said, "When everyone didn't have much knowledge of artificial intelligence, Chairman Wu had the insight that the production line could run well by integrating production information. Frankly speaking, this kind of thinking and practice was very innovative 30 years ago."
[2] Macronix High Performacne Quad SPI NOR Flash Products Power Xilinx UltraScale (TM) FPGAs for the Aerospace & Defense Market, CISION PR Newswire, 2015.
[3] Vargas-Sierra, S., Tanios, B., González-Luján, J. J., Tilhac, F., Domínguez, M., & Poivey, C. (2019, September). TID Radiation Effects of 1Gb COTS NOR Flash Memories for the ESA JUICE Mission. In 2019 19th European Conference on Radiation and Its Effects on Components and Systems (RADECS) (pp. 1-4). IEEE.
[4] Tanios, B., Kaddour, M., Forgerit, B., Guerre, F. X., & Poivey, C. (2021, September). Single Event Effects Characterization of 55-65nm NOR Flash for Space Applications. In 2021 21th European Conference on Radiation and Its Effects on Components and Systems (RADECS) (pp. 1-4). IEEE.
[5] ibid. TID Characterization of 24-45nm COTS NAND Flash for Space Applications.
[6] ibid. Single Event Effects Characterization of 24-36 nm COTS NAND Flash for Space Applications.
[7] Peng, C. Y., & Chien, C. F. (2003). Data value development to enhance competitive advantage: A retrospective study of EDA systems for semiconductor fabrication. International Journal of Services Technology and Management, 4(4-6), 365-383.
[8] 13th Distinguished Alumni of the College of Science, National Tsing-Hua University - Cheng-yung PENG (graduated from Statistics Group, the Graduate Institute of Applied Mathematics in 1988), 第十三屆理學院傑出校友 － 彭誠湧學長 (應數所統計組1988). Available:https://math-alumni.site.nthu.edu.tw/p/406-1329-142593,r1389.php
[9] "81 Award on Industrial Automation", National Federation of Industries of the Republic of China,1992. Available: https://www.president.gov.tw/NEWS/26696?DeteailNo=總統創新獎
In 1992, Macronix Electronics won the "81 Award on Industrial Automation" award from the National Federation of Industries of the Republic of China. The then Minister of Economic Affairs, Vincent Siew, awarded the "Industrial Automation Excellence Award" to Macronix, commending Macronix for actively introducing the "Computer Integrated Manufacturing System" since its establishment in 1989, integrating manufacturing, marketing and management through computerization.
Supporting materials (supported formats: GIF, JPEG, PNG, PDF, DOC): All supporting materials must be in English, or if not in English, accompanied by an English translation. You must supply the texts or excerpts themselves, not just the references. For documents that are copyright-encumbered, or which you do not have rights to post, email the documents themselves to ieee-history@ieee.org. Please see the Milestone Program Guidelines for more information.
Please email a jpeg or PDF a letter in English, or with English translation, from the site owner(s) giving permission to place IEEE milestone plaque on the property, and a letter (or forwarded email) from the appropriate Section Chair supporting the Milestone application to ieee-history@ieee.org with the subject line "Attention: Milestone Administrator." Note that there are multiple texts of the letter depending on whether an IEEE organizational unit other than the section will be paying for the plaque(s).
Please recommend reviewers by emailing their names and email addresses to ieee-history@ieee.org. Please include the docket number and brief title of your proposal in the subject line of all emails.
