Tag Archives: letter-leds-business

Harnessing big data


July 28, 2014

Addressing the analytics challenges in supply chain management. 

BY NORD SAMUELSON, CHRISTOPHER POCEK and CHRIS LANMAN, AlixPartners, San Francisco, CA 

A changing workforce and lack of convergence between information technology (IT) and business may be preventing many companies from joining the big-data revolution. Defined as very large sets of data but more commonly used in reference to the rapid increase in amounts of data in recent years, big data will divide companies into two groups in the next decade: those able to benefit from big data’s potential and those unable. Companies that create capabilities for capturing, processing, analyzing, and distributing data in order to make better decisions in real time will likely be able to outperform their competition and respond more quickly to their customers’ needs. The data avalanche is coming from a number of sources, such as enterprise resource planning, orders, shipments, Weblogs, GPS data, radio-frequency identification, mobile devices, and social channels; and there is value to be created in all areas of a business by adopting a data-driven culture.

However, in discussions about big data’s arrival, we sometimes forget to ask how effectively we’re converting the data into value. Too often, huge investments in IT infrastructure coupled with sophisticated analytical and reporting software have delivered little value. Why? We often find it’s because companies are understaffed, or they may lack the analytics talent who know how to build links between the data and the value drivers. There is also a gap between finding insights from data and then applying the insights to create value. That is where the levels of training and experience of a company’s analysts enter the equation.

One area of particular concern is supply chain management (SCM). A company’s SCM organization makes decisions about build plans, stocking locations, inventory levels, and so forth based on the conversion of raw data about demand, sales, and inventory on hand. And when there’s a shortage of analytics talent, SCM is typically one of the first areas affected. Traditionally, analytical innovation happens in two ways: either through an internal-pipeline process of developing junior analysts into senior analysts or by periodically bringing in external experts to seed knowledge. But big data is challenging both approaches.

The internal pipeline is challenged by a workforce marked by shorter tenures. Shorter tenures result
in more generalists in the workforce, often in place of the specialists needed for analytical innovation. For example, younger workers, such as millennials, are significantly less likely to settle into a long career at a company. According to a survey by Future Workplace, 91% of millennials (born in the 1980s and ’90s) expect to stay in a job for less than three years (Meister 2012), meaning that those in analytical roles are usually in the job only long enough to execute established analytics—and not long enough to develop a holistic understanding of how data can be applied to drive business value. As a result, those on the business side and those on the IT side don’t always learn to make the end-to-end connections between raw data and measurable value. The internal-pipeline approach is further challenged by companies themselves: frustrated by high turnover, companies are less likely to invest in developing their people— only to watch the people leave for higher-paying positions.

The second approach—that of periodically bringing in external experts to rebuild a process or implement the latest software package—is also starting to show wear. The evolution cycle of new analytical techniques is rapidly slowing down as big data brings opportunities to better integrate internal and external data sources. Traditionally, companies have been able to implement software solutions or bring in experts to install the latest offering and then profit from that investment for five or seven years. The initial cost was justified by the continued value for years to come. But now, the volume, variety, and velocity of the new data being generated are changing the business landscape by calling for a more rapid cycle of analytical-tool introduction. And that landscape itself usually changes every two or three years. So, as a result, the days of big-bang projects appear to be coming to an end.

What can be done? Companies should look across the entire supply chain—or across any function,
for that matter—and measure the amount of data being generated. Then they should weigh that measurement against the value actually realized. If data volumes are growing more rapidly than the corresponding increase in value, there may be an analytics talent challenge.

Three methods of creating value have proved effective in today’s rapidly changing market.

1. Outsourcing portions of analytic requirements

Companies can approach analytics outsourcing in a variety of ways, ranging from a data prep model—in which a company hires a third party to process raw data to the point where an analyst can consume it— all the way to a fully outsourced model, in which a third party processes and analyzes the data, poten- tially adds other proprietary data, and sends back fully actionable information. The data prep model enables a company to focus a limited pool of analysts on the critical knowledge-capture portion of the process and thereby free up time spent on non-value- added processes. The fully outsourced model enables companies to stay up-to-date on the latest technol- ogies and software without having to make up-front investments to purchase the latest software and technology.

2. Creating central analytics teams

Companies that rely heavily on converting data to knowledge can set up an analytic group focused solely on solving analytical issues across the company. Such companies have adopted analytics
as a core differentiator and encourage analysts to develop the holistic view that facilitates insight. Central analytics groups seem to perform better than embedded groups—and especially when they report through the business side. Of course, maintaining a group dedicated to analytics is an investment that some companies may hesitate to make, but there is tremendous value in having such in-house expertise.

3. Partnering with academic or not-for-profit institutions

Academic and nonprofit organizations are often-overlooked resources. For instance, the brand-new Center for Supply Chain Management at the University of Pittsburgh intends to provide student and faculty interactions with industry representatives who will promote experience-based learning activities within the university’s supply chain management courses. To improve the center’s effectiveness, the university plans to create a Supply Chain Management Industry Council composed of member companies dedicated to SCM. The council members, along with tenured faculty specializing in teaching SCM, will foster interest and excellence in SCM and analysis. Other institutions offer training, certifications, and conferences that encourage and enable analysts to further develop and share ideas. The Institute for Operations Research and the Management Sciences recently introduced the Certified Analytics Profes- sional certification to give companies an option for developing their people without having to make hefty investments in training organizations.

Big data is fundamentally transforming the way business operates. It is enabling management to track the previously untrackable, forecast the previ- ously unpredictable, and understand interactions between suppliers and customers—all of it with unprecedented clarity. And winning organizations will invest in the necessary infrastructure and people to harness the transformative power of data.

By Mike Rosa, Applied Materials

In 2004/2005, shipments of 300mm wafer fab equipment (WFE) began to outpace that of 200mm platforms.  As the “baton” in the node-scaling race appeared to pass from 200mm to 300mm, it was clear that device manufacturers were transitioning to higher-volume, more cost-effective 300mm toolsets for cost efficiencies of the production of advanced memory and microprocessor devices.  Tool suppliers enabled the transition with the availability of the comprehensive 300mm toolset and began a new 300mm technology race, and leaving the major OEMs to focus on service and spares for the now legacy 200mm toolsets.  With advanced device designs fully transitioned to 300mm, many IDMs and foundries were left with growing excess capacity on their 200mm production lines.

Surprisingly, new life and attention has been refocused on the 200mm tool sets and available capacity as two phenomena are driving new requirement and economics.

First, in 2006, a MEMS (Micro-Electro-Mechanical Systems)-based accelerometer became a game changer when introduced into Nintendo’s next-generation Wii motion controller.  This was the first significant and novel use of a MEMS device for motion tracking in a high-volume consumer application.  Next, in 2007, when Apple Inc. first introduced the iPhone to the world, it came to light that MEMS devices were enabling a number of its advanced motion-based features.

Later, it would be noted that more than 75% of the semiconductor device content in the iPhone was sourced from 200mm wafer starts.  The devices manufactured on 200mm wafers spanned a wide variety of applications that included not only MEMS applications (motion, audio, RF, etc.) but also CIS (CMOS Image Sensor), communications, power management and analog devices.

Sold in the hundreds of millions per year, first the iPhone and then the multitude of other smart phones, tablet PCs, and related digital devices, that followed, drove the adoption of the emerging “More-than-Moore” class of devices (which were first pioneered  on 150mm wafers at the time) onto 200mm wafers.  These high-volume consumer applications gave rise to a resurgence in both new and used of 200mm equipment. This sudden requirement for new sourcing of “legacy” 200mm toolsets placed considerable strain on a supply  chain that then focused almost exclusively on 300mm; tool vendors struggled in  refurbishment, upgrade, and production of matching tools and processes that performed outside the requirements of traditional semiconductor applications (see Figure 1).

200mm equipment market gaiting new lease on life

200mm equipment market gaiting new lease on life

Some of these additional requirements — including new and thicker films (>20µm), advanced DRIE (Deep-Reactive-Ion-Etch) capabilities capable of delivering aspect ratios approaching 100:1, and new process capabilities like HFv (Hydrofluoric Acid vapor) release etch and Wafer Bonding — resulted in OEMs needing to restart 200mm tool development.  In some cases, OEMS needed to expand their product portfolios to support the growing needs of customers producing devices in the rapidly expanding “More-than-Moore” device segment.

Fast forward to 2014 —what a difference approximately seven years has made to the industry segment and more specifically the number of opportunities in the 200mm WFE market for the new class of devices.

The surge in mobile device applications and more recently wearable technologies, has meant that device manufacturers are increasingly  under  pressure to produce cheaper, smaller, more capable and more power efficient devices most economically and efficiently — and this remains optimally on legacy 200mm toolsets.  Combining this with the materials and production challenges presented by ultra-high volume applications spelled out in the ‘Trillion Sensor Vision’ and the now looming IoT (Internet-of-Things) (see Figure 2), and it becomes clear that OEMs who continue to support and develop solutions for the 200mm WFE market  have both significant challenges and potential rewards.

Figure 2. The IoT (Internet-of-Things) by most accounts prescribes device volumes as high as 1 Trillion (per year!) by 2024. These device volumes are accompanied by severe reductions in ASP. Maintaining expanded device functionality, a reduced device size and a further reduced cost of fabrication, presents considerable challenge to both device producers and tool OEMs alike.

Figure 2. The IoT (Internet-of-Things) by most accounts prescribes device volumes as high as 1 Trillion (per year!) by 2024. These device volumes are accompanied by severe reductions in ASP. Maintaining expanded device functionality, a reduced device size and a further reduced cost of fabrication, presents considerable challenge to both device producers and tool OEMs alike.

Rising to the challenge presented by the demands of these rapidly growing market segments, Applied Materials is an OEM that has, over the past several years, continued to invest in the R&D of its 200mm portfolio products.  Challenged to deliver new materials and processes (see Figure 3) in support the growing class of 200mm emerging technology applications that have come to include MEMS, CIS, Power Device, Analog, WLP (Wafer Level Packaging), TFB (Thin Film Battery), TSV (through-silicon via), etc., Applied Materials believes that working close to the customer and more collaboratively throughout the supply chain is paramount to success in a technically challenging and price sensitive market. The 200mm ecosystem supporting broadly expanding cost-senstive device classes represent a new fork in the roadmap that has been almost myopically focused on Moore’s Law evolution.

deliver substantially re-engineered 200mm toolsets to produce advanced materials and processes needed to support the next generation of “More-than-Moore” devices. Source: Applied Materials

deliver substantially re-engineered 200mm toolsets to produce advanced materials and processes needed to support the next generation of “More-than-Moore” devices. Source: Applied Materials

Learn more about how this dynamic market is changing at the session on “Secondary Equipment for Mobile & Diversified Applications” at SEMICON West 2014 in San Francisco, Calif on July 8-10.

By Debra Vogler, SEMI

The introduction of new materials, such as III-Vs, into high-volume manufacturing of semiconductors, likely will occur sometime around the 7nm and/or 5nm nodes. III-V’s introduction, along with the potential transition to 450mm wafers, and the increasing expansion of global regulatory requirements, will heighten environmental, health and safety (EHS) concerns that must be addressed as the industry goes forward. The Sustainable Manufacturing Forum to be held in conjunction with SEMICON West 2014, will feature experts in the manufacture of semiconductors, microelectronics, nanoelectronics, photovoltaics, and other high-tech products.

One of the Sustainable Manufacturing Forum speakers, Richard Hill, Technology Infrastructure manager at SEMATECH, will discuss how the addition of III-V materials into the high-volume manufacture of semiconductors will bring sustainability issues to the forefront, primarily driven by the toxicity of arsenic that is used in much greater quantities in III-V production. Challenges include wastewater treatment, toxic gas detection control and abatement, and the need for robust protocols to ensure operator and maintenance personnel safety. Hill will speak at the Next Generation Eco Fab session on July 9 at SEMICON West.

SEMATECH recently completed a joint study of III-V EHS challenges with the College of Nanoscale Science and Engineering (SUNY CNSE). The assessment consisted of running 300mm wafers through a representative 5nm III-V process flow (Figure 1). (Many semiconductor industry experts agree that III-V materials will enter the process flows in high volumes at 5nm.) Among the processes that will pose the greatest challenges with respect to III-V materials are MOCVD, CMP, wet etch/clean, dry etch, and film deposition. The project was heavily focused on understanding the levels of arsenic that would be present in wastewater, as well as loading of other III-V materials. The impact of III-V outgassing that could occur during processing and the amounts of gases that could be released when a tool is opened for maintenance were of particular interest in the project.

Figure 1. Example 5nm III-V flow: key ESH challenges. SOURCE: SEMATECH

Figure 1. Example 5nm III-V flow: key ESH challenges. SOURCE: SEMATECH

Among the high-level challenges associated with wet etch are the potential for arsine and phosphine outgassing (during processing).

“Wet etch tools are designed to have a controlled environment,” said Hill, “but they are not like high-vacuum systems that are designed to contain toxic gases.” Hill told SEMI that if the exhaust system fails during the processing of a wafer, it is critical to know the risks and ensure mitigation. The SEMATECH/CNSE project looked at a range of different chemistries and identified those that are low risk for arsine and phosphine generation (and therefore, a low risk of outgassing) and those that had a high risk of outgassing. The low risk chemistries are, naturally, the ones that the industry should try to design into a III-V flow.

The joint project also evaluated the III-V loading in wastewater from the wet etch process. “There were measurable quantities of arsenic in the waste stream,” said Hill. Though he added that while the levels weren’t significantly high, some treatment of the waste water would have to be done depending on what’s allowable within local discharge limits and permits. With the industry looking ahead to 5nm and already designing the fabs of the future, Hill believes that these results will be important for specifying wastewater treatment.

The joint SEMATECH/CNSE project also evaluated the wastewater stream from the burn wet scrubber when III-V materials are used in a contact etch (dry) process. The study found measurable arsenic in the wastewater. “Fabs of the future will need wet treatment facilities for arsenic and indium,” Hill told SEMI. “In recent years, concerns about indium have been elevated, and we believe that tighter restrictions on it will be introduced in the future.” Chamber clean is also critical when etching (dry) III-V materials. “If you don’t do the right type of cleaning regimen, you could have next-wafer contamination.” Additionally, without the proper protocol, maintenance personnel could be exposed to arsine or phosphine when the chamber is opened, depending on the process. The cleaning protocol is highly dependent on the type of etch being done, and each type could have different requirements.

For Hill, the key takeaway from the joint evaluation was that, while there are risks when processing III-V materials, there are no showstoppers — solutions can be engineered. “People should take these risks seriously, but they shouldn’t be scared off by them,” said Hill.

Sustainability and the Role of Collaboration and Standards

Steve Moffatt, CTO, Front-end Equipment at Applied Materials (also a speaker at the Next Generation Eco Fab session at the Sustainable Manufacturing Forum at SEMICON West), told SEMI that many established procedures for dealing with arsine and phosphine already exist. He views the efforts by the industry going forward as one of accurately quantifying the size and scope of the problem. “The methods are in place, but the absolute quantities of III-Vs will be substantially higher,” said Moffatt.

Additionally, other emissions (e.g., PFCs) that are well regulated and generally understood, will see an increase in the quantities as a result of more layers being processed for 3D chips. Even the potential transition to 450mm wafers will figure into the industry’s need for a more accurate scope of the EHS challenges involved. The increase in wafer size will naturally lead to larger manufacturing equipment noted Moffatt and that, in turn, will drive increases in energy, water, and process chemical consumption at both the tool and fab levels.

As regulatory pressure increases on a global scale, the situation also becomes more complex. Beyond the use of new materials such as III-Vs and nanomaterials, Moffatt commented that new methods of energetics (i.e., ways of putting energy into a processing system) will require very careful and close assessment of the risk control measures. Another sustainability issue arises from the basic fact that, as opposed to the highly prevalent element of silicon in the earth’s crust, many of the newer materials being used in higher quantities for semiconductor manufacturing (e.g.,Ga, As, etc.) are much less abundant. These exotic materials, of necessity, must be handled in the most efficient of ways.

Going forward, there will be increased regulatory pressure to reduce a fab’s carbon footprint and produce more sustainable products. Moffatt says the industry can expect more pressure to reduce greenhouse gas (GHG) emissions along with adhering to conflict minerals regulations and managing EHS concerns throughout the entire life-cycle of a product (Figure 2). “One company can’t do it on its own, it’s a life-cycle consideration,” said Moffatt. “If we have the right collaboration together, we have a greater probability with the right kinds of standards of bringing good, effective green chemistry solutions to high-value problems.”

Figure 2. Consensus building in multi-stakeholder life-cycle risk assessment of manufacturing technology and products. SOURCE: Applied Materials (used with permission of ITRS)

Figure 2. Consensus building in multi-stakeholder life-cycle risk assessment of manufacturing technology and products. SOURCE: Applied Materials (used with permission of ITRS)

Regarding standards activities on energetics, Moffatt pointed to ongoing collaboration and hazard assessment between SEMI, SEMATECH and other industry groups.

“We will need to continually evaluate the need for additional standards activities — both new and updates — in addition to industry collaboration on “Green” chemistry,” said Moffatt.  “As a starting point, sustainability concerns could be built into the initial assessment of new chemicals and processes, which will begin the discussion and raise awareness of these issues.”

Hill (SEMATECH) and Moffatt (Applied Materials) will be joined by speakers from IMEC, Intel, Samsung, Air Products, and MW Group at the “Next Generation Eco Fab” session of the Sustainable Manufacturing Forum at SEMICON West 2014, July 7-10 in San Francisco, Calif.  For more information, visit: http://www.semiconwest.org.