Sustainability in semiconductor operations

Climate change is creating many life-threatening disruptions, including extreme weather, rising sea levels, and droughts. Faced with irrefutable evidence of global warming, almost 200 countries have committed to the 2016 Paris Agreement, a treaty that calls for accelerated decarbonization. This agreement is designed to limit the mean rise in temperature to 1.5 degrees Celsius from preindustrial levels to mitigate or prevent some of the most dangerous effects of climate change. While some semiconductor companies have created ambitious targets for reducing their emissions and remaining on a 1.5°C pathway, many others have been less ambitious. The pressure to act may soon increase, however, since businesses across industries are now scrutinizing emissions along their entire supply chain—and in many cases, semiconductor companies will account for a substantial amount of them. Already, some of the semiconductor industry’s most important end customers, including Apple, Google, and Microsoft, have committed to reaching net-zero emissions for their full value chain and set aggressive timelines for achieving their goals. Some semiconductor companies have responded by setting their own emissions goals. For instance, Infineon plans to reduce greenhouse-gas (GHG) emissions by 70 per cent by 2025, compared with its 2019 baseline, and aspires to reach carbon neutrality for emissions directly under its control by the end of 2030. Intel recently committed to net-zero GHG emissions in its global operations by 2040 and has targeted achieving 100 per cent use of renewable electricity as an interim milestone in 2030. Several semiconductor players have also committed to science-based targets, including STMicroelectronics, NXP, and UMC. Over the next few months or years, more semiconductor companies are expected to commit to ambitious and actionable emissions targets.

Achieving substantial emission reductions will require collaboration with peers and suppliers, as well as new technologies, innovative thinking, and the complete engagement of fabs. To help companies move forward, we reviewed the current state of greenhouse-gas emissions within the semiconductor sector and collected best practices for abatement. Our analysis allowed us to identify both short- and long-term solutions along the entire semiconductor value chain. This article focuses on scope 1 and 2 emissions, which are the ones that semiconductor fabs can directly control.

With about 80 percent of semiconductor manufacturing emissions falling into either scope 1 or scope 2 categories, fabs control a large portion of their GHG profile.

As the node size of chips continues to shrink, energy requirements at production facilities are expected to rise significantly.

Scope 1 emissions, which also significantly add to fabs’ GHG emission profile, arise from process gases used during wafer etching, chamber cleaning, and other tasks. These gases, which include PFCs, HFCs, NF3, and N20, have high global-warming potential (GWP); they rise as node size shrinks.2 Scope 1 emissions may also arise from high-GWP heat transfer fluids that may leak into the atmosphere when fabs use them in chillers to control wafer temperature during manufacturing processes.

Additional emissions may come from upstream scope 3 sources, such as suppliers, chemicals and raw materials, or from transportation to customer facilities. These upstream emissions generally account for only about 20 percent of fabs’ GHG profile, however.

Semiconductor companies also generate downstream scope 3 emissions, which are related to use of products containing semiconductors. These vary significantly by use case. For instance, handheld devices with low power consumption during intermittent usage will have much lower emissions than data centers that operate 24/7. As will be discussed in a later article, product design influences scope 3 emissions, giving fabs little control over them during operations.

To help fabs achieve substantial emissions reductions and accelerate decarbonization, we identified three areas that need immediate attention, as well as relevant improvement levers.

Levers for reducing energy consumption are often directly aligned with other operational targets, such as cost reduction, making them easier to achieve. The many options available can be grouped into two major categories. The first group focuses on reducing tool-related energy consumption—for instance, by upgrading and replacing tools with more energy-efficient ones, implementing smart control systems to enable coupling and regulation of facilities and tools. The second group encompasses activities that involve reducing facility-related energy consumption though various measures, such as exclusive sourcing and use of energy from renewable sources, greater energy efficiency of buildings, and replacing existing lighting in fabs with LED fixtures.

To identify the greatest opportunities for decreasing energy consumption, fabs could look at benchmark-based targets and sources of energy loss. They could also review existing levers for energy reduction by tool and facility type. For instance, fabs might discover that they can improve energy consumption in clean rooms by reducing air pressure, increasing humidity, limiting air exchange in unused areas, or eliminating leaks in air-supply lines.

When optimizing process recipes, equipment engineers typically focus on overall equipment effectiveness (OEE) and give little attention to tool-fleet energy consumption. New incentives, such as rewards for creating energy-efficient recipes, might help change this mindset.

mckinsey.com