The Challenges Faced by Manufacturers Today

Critical Manufacturing recently published e-book by Tech-Clarity. It explores the current challenges faced by semiconductor manufacturers and how their MES needs to evolve in order to alleviate these challenges.

If the pandemic taught the global manufacturing fraternity anything, it is the need to have agile and highly resilient operations, where market dynamics are quickly adapted to and value generation is sustained through an integrated supply chain and an informed manufacturing operation. For Semiconductor manufacturers, there is an evident need to upgrade their processes to incorporate advanced fab capabilities and to deliver value with a diverse product mix catering to several customers/industries, to utilize data better and move towards a more platform-based approach when it comes to Industry 4.0 and IT systems within their operations.

Typically, semiconductor fabs have an array of IT applications covering WIP, track & trace (MES), scheduling, maintenance, reporting and analytics. While this approach might have worked in the past, the high demand from the market and added complexities in processes, coupled with the need to have a larger product mix, have subjected the industry to a great deal of pressure.

Increasing the product mix also increases the amount of data generated. Coupled with executing new processes, it can become impossible to contextualize the data as it moves through different applications and iterations. Without contextualization, informed decision making is challenging.

High Demand equals High Pressure

Looking beyond electronics and telecom, nearly every industry segment in their pursuit of Industry 4.0 is in need of semiconductors. Chip demand is booming and so is the need to focus on design wins, deliver more through existing capabilities while adding new ones, add new processes and products. While the need for faster innovation and delivering products at a lower cost isn’t new, the need to have a larger product portfolio is.

In order to ensure these new products added to the portfolio are designed, developed and manufactured in an optimal manner, data needs to be converted into actionable insights and information. The industry is growing at a strong pace of almost 30% in terms of year over year revenue, and this high demand translates to high pressure to deliver more, faster.

Semiconductor Industry revenues and percentage growth 1996-2021(Source: WSTS and Tech-Clarity)

Urgent Innovation

The product portfolio across fabs has increasingly become more diverse over the last decade or so, with fabs offering both old and new technologies with products ranging from simple to highly complex. ‘Smart everything’ means products need more and more chips and they need to be extremely reliable especially when human life depends on its performance. Sustainability also plays a major role with customers demanding lower footprint and reduced power specs.

Innovation in the industry needs to expand beyond the fab and into the value chains with tighter collaboration in the ecosystem, composed of research labs, customers and foundries. Managing experiments and non-production wafers is becoming increasingly important to deliver first time right results and faster TTM and TTV.

More Products More Challenges

A higher product mix means more recipes, reticles and mask sets to manage. Companies oftentimes mix experiment and non-production wafers into production lots as well to meet the need for faster experimentation, validation and roll out of new products. This further complicates tracking and sequencing.

Most semi companies are adding capacities, but a higher product mix tends to effect capacity utilization negatively. Ensuring equipment qualification, determining the qualification state and the deriving the correct meaning of qualifications becomes extremely important while planning several complex process steps for a given lot.

 Having invested millions in equipment, our challenge is to get ROI on it. We drive utilization and uptime of these machines. We also ensure, where we can, commonality in the projects and products we accept.”

Dr. Frank van de Scheur, Head of MEMS and Micro Devices, Philips (quote from the e-book)
The Processing Problem

Complex products need special treatments to deliver requisite characteristics and yields, which means dealing with chamber and tool related complexities, creating highly complicated recipe models for mixed lots on same equipment and ensuring dwell time isn’t increased while the right product is processed through the right equipment.

Streamlining this process flow with different tooling needs for the same lot across multiple recipes and containing products being made per multiple design specs is posing a major challenge for modern fabs, especially when the time required to achieve perfect execution is shorter than ever.

Improvement with Agility

Semiconductor companies collectively feel that they aren’t doing as good as they should be when it comes to Continuous Improvement efforts. The graphic below portrays just that. Perhaps one of the reasons for this sentiment is the inability of being able to leverage all process data and convert it into actionable information through proper and real-time contextualization.

Portion of semiconductor vs. all respondents who rated themselves as excellent at continuous improvement (Source: “The Manufacturing Data Challenge,
© 2020, Tech-Clarity, Inc.)

Agility too has gained prominence in the strategy mix alongside CI. There were supply chain disruptions and rapid decline in demand at the beginning of the pandemic from some industries, coupled with a massive surge of demand from others. The need to pivot chip production from one industry and product type/s to other has become a major challenge and requirement for any modern fab and the ongoing digital transformation initiatives. 

High Quality, Higher Yield

As the potential impact of product quality on user experience and some cases even their wellbeing increases, so does the need to adhere to specs and deliver products with ‘zero defects’. As the pressure on faster NPI, cost and quality mount, they are coupled with high penalties for product failure. From a process perspective, ensuring carriers and measurement tools remain free from cross-contamination is extremely critical and challenging with complex flows and repetitive processing across chamber/tools and contamination zones.

High yield too becomes a challenge when faced with complex products and process/material flows. Semiconductor companies need their processes, people and equipment to contribute towards not just maintaining yield levels but increasing them, even when faced with complexities pertaining to product mix, process and quality. Their IT infrastructure and the speed with which it adapts to and enables changes and compliance plays a major role in boosting yield to the desired levels.

Workforce Woes

Countries like China, which dominate the semiconductor manufacturing arena, report a shortage of around 400,000 skilled workers.

Like China, the US semiconductor workforce is at risk, with nearly of the 40% personnel above the age of 50 years. These personnel know the process intimately and their way around process and tooling related challenges. MES administrators too belong to this age category, and may soon retire, making the legacy MES maintenance and upkeep a major challenge, with a lack of similar expertise around to maintain/modify the current application.

Younger recruits want better tools, guidance and collaborative applications from an IT perspective. They expect to use technologies like AR/VR to perform better and learn faster.

An MES, which orchestrates the overall semi process, becomes extremely important for companies facing knowledge erosion and coping with the challenges of preserving tribal knowledge while training new and comparatively lesser-skilled staff.

Better IT Now

Managing complex recipes with multiple tolling requirements and material flows to enable a larger product mix poses a massive challenge from a data perspective. Handling master data such that data analytics requirements are met and real-time decision support is achieved becomes challenging when large amounts of data being generated from process and IoT sensors either isn’t captured or worse yet, captured but can’t be contextualized due to the lack of integration between point solutions, thereby creating siloed data and delayed actions on the shop-floor.

Legacy MES needs to evolve from a traceability solution to a full-fledged orchestration, control and compliance platform, which enables real-time data capture, edge computing and unleashes the power of AI and advanced analytics within the process, while integrating across the enterprise and extended value chain to create better collaboration.

Companies at this point have two choices: either recruit an army of IT experts and data scientists to keep legacy applications alive and deal with the data requirements of process engineers and designers, or invest in the right IT infrastructure, starting preferably with a reimagined MES. A modern MES lays the foundation for Industry 4.0, a more efficient operation, better quality, higher yield, lower costs and above all delivers the agility and resilience most desired in today’s uncertain times.