RequestLink
MICRO
Advertiser and
Product
Information

Buyer's Guide
Buyers Guide

tom
Chip Shots blog

Greatest Hits of 2005
Greatest Hits of 2005

Featured Series
Featured Series


Web Sightings

Media Kit

Comments? Suggestions? Send us your feedback.

 

MicroMagazine.com

Green Manufacturing

Achieving sustainability in the semiconductor manufacturing industry

Walter F. Worth, International Sematech Manufacturing Initiative (ISMI)

From the outset, the semiconductor industry has paid much attention to environment, safety, and health (ESH) issues. Indeed, the industry has made a concerted effort to build a reputation for being progressive, clean, and environmentally friendly while attempting to achieve long-term sustainability. This article addresses the industry’s progress on the road to sustainability.

Sustainability is defined as the ability of an endeavor “to be maintained at length without interruption, weakening, or losing in power or quality.” In the words of the World Commission on Environment and Economy, sustainability is “a condition whereby the needs of the present are met without compromising the ability of future generations to meet their own needs.” The Dow Jones Sustainability Index has another take, defining sustainability as “a business approach that creates long-term shareholder value by embracing opportunities and managing risks deriving from economic, environmental and social developments.”

Besides altruistic motives, there are several practical reasons why sustainable, cost-effective manufacturing is important to the industry. First, product margins are declining at the same time that companies’ market capitalization depends on their financial performance. Second, semiconductor manufacturing facilities use significant quantities of electricity and water. For example, a 2002 survey of member companies by the Semiconductor Industry Association (SIA) indicated that the average fab consumes 175 million kWh of electricity annually, the amount consumed by 13,500 U.S. homes, and 1.6 billion liters of water annually, the amount consumed by 4000 homes. As a result, in some regions the ability to expand an existing site or to locate a grassroots fab at a new site may be limited by the availability of resources such as electricity and water.

To achieve sustainable operations, the industry must make efficient use of natural resources, minimize environmental impact, protect workers’ health and safety, maintain positive community relations, stay in compliance with regulations, and—last, but not least— bring value to its shareholders.

Reducing Utilities Consumption

Water Consumption. For many years the industry has pursued efforts to reduce water consumption with encouraging results. Most fabs now recycle significant portions of their ultrapure water (UPW) by collecting wafer rinsewater at the tools and recycling it back to the UPW feed tank. This procedure has resulted in significant water and chemical savings, since the rinsewater is usually of better quality than the water purchased from a city or private water utility. On-line conductivity and total organic carbon monitors have considerably improved over time and are typically used to protect the UPW system against major process upsets. A UPW flow model developed by the University of Arizona has helped pave the way for the adoption of this technology by allowing the industry to predict the impact and recovery time of potential process upsets before implementing UPW recycling.1

Several International Sematech Manufacturing Initiative (ISMI) studies on water consumption in chemical-mechanical polishing (CMP) tools have led to water conservation measures, by lowering water flows for wafer staging and tool rinsing and by segregating and reusing the water that typically goes to waste during tool idle periods. Considering that CMP constitutes up to 50% of a fab’s total UPW use, water savings using such techniques can be significant.2

In wet benches, improved tank design and optimized rinse processes have also resulted in significant water savings during wafer rinsing. Because of a more-conformal tank design, water no longer bypasses the wafers in the center of a wafer boat, as illustrated in Figure 1. Similarly, the move away from overflow rinsing to quick-dump rinsing and the greater use of sprays have reduced both the time and quantity of water required for wafer rinsing. The move to single-wafer spray processors has further reduced UPW consumption. In addition, fabs have been very creative in reusing “clean” wastewater for cooling-tower makeup, house wet scrubbers, and site irrigation.

Figure 1: Before tank modifications in wet benches, most UPW bypassed the wafers in the center of the wafer boat during overflow rinsing. (COURTESY OF NSF/SRC ENGINEERING RESEARCH CENTER FOR ENVIRONMENTALLY BENIGN SEMICONDUCTOR MANUFACTURING)

Energy Consumption. Similar efforts in the area of energy conservation have significantly lowered fab energy consumption relative to unit area of silicon processed. For example, tracer-gas studies on wet benches, gas panels, cylinder cabinets, and diffusion furnaces have shown that exhaust can be reduced by 30–80% without affecting yields or worker exposure (<25% of the threshold limit value or occupational exposure limit is considered safe).3,4 While implementation in the installed tool base is somewhat problematic because of factory preset low-flow alarms, new tool designs are taking these findings into account. Lowering high-efficiency particulate air (HEPA) filter flow velocities has also had a significant impact on fab energy consumption. Most fabs have now reduced HEPA velocity from 120 to 70 ft/min without affecting laminar flow, particle count, and worker comfort standards. An ISMI member company reported that its fabs saved more than $3.3 million per year after implementing HEPA velocity reductions.

Another ISMI study investigated on-site nitrogen generation and the potential for optimizing the use of nitrogen in the fab. It was found that rewheeling compressors, sealing equipment openings, and replacing nitrogen with clean dry air in nonprocess applications can save energy.
Another industry achievement has been the development of more-efficient vacuum pumps. Considering that fabs can have up to 700 vacuum pumps, a facility can reduce its energy consumption significantly by switching to more-efficient pumps when existing ones wear out. As shown in Figure 2, which estimates power usage from pumps by two different suppliers, energy savings greater than 50% per pump are possible. However, since pumps have to be placed close to the tool and cleanroom space may be limited in existing fabs, use of energy-efficient pumps may be limited to new fabs.

Figure 2: Energy reduction resulting from the use of high-efficiency vacuum pumps.

ISMI is also investigating the possibility of developing an idle- or sleep-mode capability for vacuum pumps when tools are on standby. Since tools are idle 20–30% of the time but are turned on continuously, idling vacuum pumps should result in significant energy savings. The tool-to-pump communication interface is a hurdle that is being resolved by developing additional tool software and new tool-pump communications standards. Ancillary benefits of a vacuum-pump idle mode include lower pump-purge nitrogen consumption and reduced cooling-water requirements.

Thermal Management Efforts. Tool thermal management, or heat removal from tools, is another area that is receiving attention. Measurements on actual tools in high-volume manufacturing have shown that up to 75% of the power input into a tool is dissipated to the cleanroom, and only about 15% is removed with process cooling water (PCW).5 As presented in Figure 3, similar observations have been made for an implanter, furnace, etcher, and wet sink. On the other hand, other analyses have shown that heat removal using cooling water is about four times more efficient than radiating the tool’s heat to the cleanroom air and then removing it in the fab’s heating, ventilating, and air conditioning (HVAC) system.6 Figure 4 demonstrates the efficiency of dissipating heat using PCW.

Figure 3: Inefficient systems for removing heat from implanters dissipate large amounts of heat into the cleanroom.
Figure 4: Heat-removal efficiency comparison. Heat removal using PCW is about four times more efficient than radiating the tool’s heat to the cleanroom air and then removing it in the fab’s HVAC system.

Data collected from SIA member companies for 40 U.S. fabs demonstrate that over the past four years (2001–2004), there have been steady, significant decreases in the consumption of feed water, UPW, and electricity of 46, 23, and 31%, respectively, as presented in Figure 5.6 This progress is encouraging and shows that the industry’s concerted efforts in the area of resource conservation are paying off.

Figure 5: Reduction in the consumption of utilities (feed water, UPW, and electricity) from 2001 to 2004. (Original data were collected in metric per cm2 of wafer outs.) (DATA COURTESY OF SIA)

Reducing Chemical Consumption

Photoresist and Photoresist Stripping Chemicals. As in the utilities area, the semiconductor industry has made significant progress in reducing chemical consumption. An outstanding example is the decreased use of photoresist. By prewetting wafers and programming tools to spin the wafers at varying speeds during resist application, fabs have been able to reduce photoresist consumption from 6 to 1 ml per 300-mm wafer. This reduction has resulted in savings of up to $5 million per year for fabs using advanced-technology photoresists.

Another achievement has been the elimination of whole classes of hazardous chemicals from semiconductor manufacturing. Eliminated materials include certain glycol ethers; polychlorinated biphenyls; fully halogenated chlorofluorocarbons; carbon tetrachloride; 1,1,1 trichloroethane; the halons 1211, 1301, and 2402; and hydrobromofluorocarbons.

An example of the industry’s attempt to move away from hazardous chemicals is its replacement of Piranha, a hot sulfuric acid–hydrogen peroxide mixture used for resist stripping, with less-hazardous ozonated DI water. This change not only reduces the risks associated with the use of hazardous chemicals, but also may reduce UPW consumption, since rinsing residual sulfuric acid is quite difficult and requires copious amounts of water.

As illustrated in Figure 6, the quantity of Resource Conservation Recovery Act (hazardous) waste, total waste, and volatile organic compounds per unit of silicon processed decreased from 2001 to 2004 by 33%, 56%, and 38%, respectively, according to SIA data for 40 member fabs. That significant achievement highlights the industry’s diligent efforts to reduce the consumption of chemicals.

Figure 6: Reduction of hazardous waste, total waste, and volatile organic compounds from 2001 to 2004. (Original data were collected in metric per cm2 of wafer outs.) (DATA COURTESY OF SIA)

Lowering PFC Emissions. Reducing perfluorocompound (PFC) emissions to prevent global warming is another example of the industry’s efforts to meet global environmental standards and priorities. Through a combination of process optimizations, the use of alternative gases, and the implementation of abatement measures, the industry has been able to reduce total PFC emissions significantly, despite continuing growth.

While utilization of the original chamber-cleaning gas C2F6 roughly doubled from 30 to 60% between 1994 and 2002, the introduction of a remote plasma-chamber cleaning process using NF3 has almost eliminated global-warming emissions, since the dissociation of NF3 in the powerful plasma chamber upstream of the deposition chamber is very efficient. Moreover, because alternative gases such as C3F8, C3F8O, C4F8 and NF3/He are drop-in replacements for C2F6, they require little or no new capital equipment, easing their adoption by the industry. As shown in Figure 7, a 95% reduction in emissions is possible by replacing C2F6 with NF3 to perform in situ chamber cleans.

Figure 7: Global-warming emissions have been reduced thanks to the use of alternative chamber-cleaning gases. (Baseline process used 600 std cm3/min C2F6 with 600 std cm3/min O2.)

The semiconductor industry’s equipment and device manufacturers have developed an array of abatement options, including plasma, combustion, and packed-bed catalytic devices, which can achieve destruction/removal efficiencies of up to 95–99%.

In lieu of regulations, U.S. semiconductor manufacturers have entered into a voluntary memorandum of understanding (MOU) with the U.S. Environmental Protection Agency (EPA) to make a concerted, industrywide effort to reduce PFC emissions. As part of the MOU, which is in its second five-year term, the signatories track their emissions annually and report the data to EPA. Figure 8 plots annual emissions expressed in million metric tons of carbon equivalents (MMTCE) from 1995, when the MOU was adopted, to 2004, the latest year for which data are available.

Figure 8: PFC emissions reported by MOU signatories from 1995 to 2004.

Globally, the World Semiconductor Council (WSC), which is composed of the semiconductor industry associations in the United States, Europe, Japan, Taiwan, and South Korea, has set an absolute emissions reduction goal for each national association. For the United States, the PFC emissions goal is to be 10% below the 1995 level by 2010. As shown in Figure 8, emissions in the U.S. peaked in 1999 and were already below the 2010 goal in 2004.

PFC emissions for all WSC members are trending downward, as illustrated in Figure 9. If steps were not taken, PFC emissions would increase exponentially in line with the historical 15% compound growth rate of the IC industry. Thus, the 10% reduction goal from 1995 levels represents a roughly 90% reduction from what the emissions levels would be without control.

Figure 9: Normalized emissions data for WSC members, including future projections. (DATA COURTESY OF SIA)

International Regulatory Trends

Another way the semiconductor industry is attempting to achieve sustainability is to anticipate and prepare for new regulatory trends worldwide. In the United States, Asia, and Europe, governments are seeking to regulate whole classes of chemicals of concern, such as carcinogenic, mutagenic, reproductive hazards (CMRs); persistent organic pollutants (POPs); and persistent bioaccumulative toxic compounds (PBTs). Goods producers are being forced to take responsibility for the full life cycle of their products (i.e., from R&D through manufacturing and marketing to use and final disposal). In addition, there is increased emphasis on the public’s right to know and a trend toward requiring more documentation and verification of the hazards and risks associated with chemicals.

In many of these areas, the European Union (EU) has taken the lead. A good example is the EU’s upcoming Registration, Evaluation, and Authorization of Chemicals (REACH) regulation, which requires registration of all substances manufactured or imported into Europe in quantities greater than 1 ton/year; evaluation (including information collection and testing) of substances suspected of causing risk to human health or the environment; and authorization of substances that fall into categories of concern, including CMRs, POPs, PBTs, and endocrine disrupters.

It appears that Europe is interested in regulating not only the content of industrial products but also the way industry manufactures them. While REACH is still being formulated, many other, more-focused restrictions are already in force in Europe. For example, the Restriction of Use of Hazardous Substances has two components: Directive 2002/95/EC restricts the use and bans importation of substances that the EU deems environmentally hazardous. The list includes lead, mercury, cadmium, chromium VI, polybrominated biphenyls, and diphenyls. Directive 2002/96/EC requires manufacturers to recycle products at the end of their lives.

The latest European regulatory initiative is a proposal to ban the manufacture, importation, and use of perfluorooctyl sulfonate (PFOS), a vital ingredient of photoresists and antireflective coatings (ARCs) with no acceptable substitutes. In 2003, the UK concluded that PFOS is a persistent bioaccumulative toxic compound that should be banned. The 25 EU member states have considered the UK’s recommendation, and the European Commission is drafting legislation to implement the ban across Europe. In response, the European SIA has pointed out that the semiconductor industry uses very little PFOS (normally 1% or less in resists and ARCs) and carefully controls PFOS-containing wastes (92% of which are incinerated). Consequently, the draft legislation includes an exemption for photolithography. In fact, the language of the regulation is similar to the EPA’s Significant New Use Rule, which regulates PFOS-containing substances in the United States.

Conclusion

As the semiconductor industry advances to smaller and smaller feature sizes (i.e., 45 and 32 nm), scaling alone will no longer suffice, requiring the introduction of new materials, chemicals, and processes. In fact, the whole periodic table is being scoured for materials that will meet the stringent performance requirements of advanced gate stacks and porous low-k, pore-sealing, barrier, and copper seed interconnect materials. The challenge is to assess the toxicity and environmental hazards of these materials during process development to avoid ESH showstoppers. Early identification of potential hazards is critical so that process and tool designers can be alerted to potential regulatory restrictions or bans, enabling them to consider alternative materials and processes. A hurdle facing ESH professionals is that some materials are either unique to the semiconductor industry or contain proprietary ingredients. In such cases, few toxicity or environmental data exist or are available to make a comprehensive ESH assessment. Nevertheless, ISMI has an active program in this area.

It is fair to say that over the past 5 to 10 years, the IC industry has made very significant progress in reducing water and energy consumption, the use of hazardous chemicals, and the generation of hazardous and nonhazardous waste. Companies have implemented many resource conservation measures that will increase the sustainability of semiconductor manufacturing over the long term. Contrary to a common misperception, environmental improvements usually lower fab operating costs significantly and tangibly. To ensure the success and sustainability of the semiconductor industry, the drive to make IC manufacturing processes even more efficient and environmentally benign must continue.

References

1. F Shadman, “New Processes and Materials for Environmentally Benign Semiconductor Manufacturing” (paper presented at the International Forum on Semiconductor Technology, Antwerp, Belgium, March 7, 2001).

2. R Chiarello, “Water Use Issues and Reduction Strategies for CMP and Post-CMP Cleaning,” Sematech Technology Transfer No. 00124046A-ENG (2000).

3. T Huang, “The Characterization and Impact of an Exhaust Optimization for Two Semiconductor Process Tools,” Sematech Technology Transfer No. 01024084A-TR (2001).

4. T Huang, “The Characterization and Impact of Exhaust Reduction at Various Semiconductor Process Tools,” Sematech Technology Transfer No. 01014068A-ENG (2001).

5. T Huang, “Tool Energy Pareto Study,” Sematech Technology Transfer No. 0094003A-ENG (2000).

6. P Naughton, “Energy Conservation at Equipment: Facts, Needs, and Vision” (paper presented at the Semicon Europa Energy Workshop, Munich, Germany, April 13, 2005).


Walter F. Worth, PhD, is a Sematech Fellow and member of Sematech’s Environment, Safety, and Health Technology Development Group in Austin, TX. He has been employed by the organization since 1993. Previously, he worked for Exxon Research & Engineering, Bechtel, and Brown & Root in the chemical and petrochemical industries. A registered professional engineer in Texas and a member of AICHE, the Electrochemical Society, and SESHA, Worth received a BS from the University of Toronto, Canada, and MS and PhD degrees from the Massachusetts Institute of Technology in Cambridge. All of his degrees are in the area of chemical engineering. (Worth can be reached at 512/356-7199 or walter.worth@sematech.org.)


MicroHome | Search | Current Issue | MicroArchives
Buyers Guide | Media Kit

Questions/comments about MICRO Magazine? E-mail us at cheynman@gmail.com.

© 2007 Tom Cheyney
All rights reserved.