For the past few years, semiconductor companies have focused on addressing the chip shortage and directed their energies into increasing supply. But another customer imperative, as well as the health of the planet, now requires increasing attention: the need for greater sustainability. Many end customers are already asking their suppliers, including semiconductor companies, to step up their efforts to reduce greenhouse-gas (GHG) emissions to achieve net-zero carbon emissions along their entire supply chain (see sidebar “Semiconductor companies respond to global warming with ambitious sustainability commitments”). Suppliers are taking heed of these requests, both to retain business and to join in the fight against global warming.
The urgent need to increase sustainability efforts comes just as semiconductor production ramps up to meet growing demand and more sophisticated chip designs are required for leading-edge mobility, computing, and connectivity applications. As production increases, however, so do emissions. Even with major semiconductor companies’ latest commitments, which are more stringent than past measures, the industry is not on track to limit emissions to the extent required under the 2016 Paris Agreement, which aims to restrict the mean rise in global temperature to 1.5°C from preindustrial levels.
Some semiconductor companies have recently set more aspirational emissions-reductions targets, but getting the industry to net zero will require more comprehensive action. This article offers a coherent, industry-wide road map that could be considered by semiconductor device makers seeking to achieve a 1.5°C trajectory by 2030 and net-zero emissions by 2050.
In addition to describing how companies can apply existing levers to reduce emissions, the road map highlights the importance of developing new or improved decarbonization technologies—a lengthy process—and increasing the supply of renewable energy. Our analysis builds on the findings in a related McKinsey article, “Sustainability in semiconductor operations: Toward net-zero production,” which describes the decarbonization levers for semiconductor front-end fabs.
A multifaceted emissions problem
Emissions from semiconductor device makers fall into different categories:
- Scope 1 emissions arise directly from fabs, primarily from process gases with high global warming potential (GWP) that are used during wafer etching, chamber cleaning, and other tasks; they can also come from high-GWP heat-transfer fluids that may leak into the atmosphere when fabs use them in chillers.
- Scope 2 emissions arise directly from purchased electricity, steam, heating, and cooling equipment; the major sources include production tools and facilities/utilities.
- Scope 3 emissions include all other indirect emissions in a company’s value chain; upstream emissions are those generated by suppliers or their products, while downstream emissions are related to the usage of products containing semiconductors.
By aggregating emissions data from key semiconductor device makers, the percentage of emissions classified as Scope 1, 2, or 3 upstream can be determined (Exhibit 1).
Our research revealed that 35 percent of emissions at a typical semiconductor fab fall into the Scope 1 category, compared with 45 percent for Scope 2, and 20 percent for Scope 3 upstream. This breakdown can vary by fab, however, based on various factors, including the amount of renewable energy used and the extent to which process-gas-abatement systems have been implemented.
This article focuses on reducing Scope 1 and 2 emissions at semiconductor device makers, since these represent the largest share of emissions and are under direct control of semiconductor fabs. In a separate article, we will investigate how the semiconductor industry can reduce Scope 3 emissions.
Modeling of Scope 1 and 2 emissions
The steps that semiconductor companies are now taking will not be enough to get the industry on a 1.5°C trajectory by 2030. In fact, Scope 1 and 2 emissions may rise significantly above current levels as semiconductor production volume increases and the industry moves to advanced nodes with higher expected emissions intensity. To gain more clarity about the extent of the increase, we created three different scenarios to estimate future emissions. The assumptions underlying each scenario are as follows:
- Scenario 1 (conservative). All semiconductor companies continue their current decarbonization efforts and do not pursue more ambitious goals or initiatives, even if they have publicly announced plans to do so.
- Scenario 2 (optimistic). All companies that have announced decarbonization goals deliver on their commitments, while companies that have not announced decarbonization ambitions continue business as usual.
- Scenario 3 (ambitious). All companies undertake the actions needed to achieve 1.5°C trajectory by 2030.
In the conservative scenario, carbon dioxide equivalent (CO2e) would increase from 93 million tons in 2020 to 183 million tons by 2030, reaching 73 million tons for Scope 1 and 110 million tons for Scope 2 (Exhibit 2).
In the more optimistic scenario, where some companies step up their decarbonization efforts as announced, Scope 1 and 2 emissions increase at a slower rate and reach 116 million tons by 2030. In the ambitious scenario, emissions are at 54 million tons that year—lower than the levels recorded in 2020. (For more information on our model, see sidebar “Our methodology for projecting Scope 1 and 2 emissions.”)
A potential road map to help the semiconductor industry reach a 1.5°C trajectory by 2030
Based on our modeling, Scope 1 and 2 emissions would amount to 89 million tons of CO2e in 2030, even if companies apply all levers to the fullest extent now possible (Exhibit 3). While this is about half of the 183 million tons projected in the conservative scenario, it would fall short of the 54 million tons needed to remain on track for net zero.
To get on a net-zero trajectory by 2030, the semiconductor industry would benefit from a coordinated effort to apply current strategies in full while simultaneously developing and adopting new technologies. These steps would be required to reach net-zero emissions by 2050—a goal that would call for the industry to reduce Scope 1 and 2 emissions by 95 percent from the 2020 level depicted in the conservative scenario.
For Scope 1 emissions, all semiconductor players would need to double down on their commitment to reduce emissions. This goal would require them to implement existing technologies to a significant extent. Consider the following:
- Process gas. Semiconductor fabs could feasibly install gas-abatement systems that cover 90 percent of tools on average. Processing gas chemistry would have to be optimized to lower GHG usage, such as by replacing nitrogen trifluoride (NF3) and tetraflouromethane (CF4) with fluorine (F2) gas, which has zero global warming potential.
- Heat transfer fluid (HTF). At least 70 percent of HTF would need to be replaced with low GWP options. Semiconductor fabs would also need to reduce chiller leakage.
- Fuel consumption. Semiconductor companies would need to replace the current fuel supply with clean options, such as hydrogen/biomass.
With Scope 2 emissions, all players would need to continue reducing energy consumption per wafer year over year, but they can also go beyond that by increasing their share of renewable energy. Overall, the industry would need to increase its renewable-energy share to a level that is 1.4 times higher than the current share in local grids. In many geographies, this will require ambitious and truly breakthrough approaches, such as importing renewable energy or building renewable-energy-generation plants.
To stay on a net-zero pathway until 2050, every semiconductor company would need to shift toward a more proactive and innovative approach—and that means making a greater investment in decarbonization and showing a willingness to experiment. For instance, the semiconductor industry can learn from other industries by introducing innovative solutions to fabs as they become available. As noted earlier, companies would benefit from investigating new solutions now, given development timelines, and some key players have already done so. Companies could also develop early-stage decarbonization technologies in partnership with leading start-ups and academic labs. Innovation will involve some risk, since not every idea will lead to real solutions that meaningfully reduce emissions.
For Scope 1 emissions, companies could continue to develop and implement alternative gas chemistry. New abatement solutions could also be implemented to enable more efficient gas removal. Simultaneously, companies could explore innovative technologies for recycling and reusing process gas that cannot be eliminated because of technical challenges.
With Scope 2 emissions, companies could increase the use of renewable energy. To do so, the industry would need to continue implementing the innovation solutions to increase supply, such as importing energy to locations with supply constraints.
The benefits of greater collaboration
To implement a road map for net zero, collaboration among semiconductor companies and other players along the value chain may be critical.
When reducing Scope 1 emissions, for instance, companies would benefit by reducing process-gas emissions in fab operations. Achieving this goal might be easier if they work with various specialty suppliers—gas/material, equipment, and abatement—to improve existing solutions and develop innovations, including those related to replacement, abatement, and recycling. Companies could also look for collaborative opportunities along the value chain to identify future gas/material requirements and define a common road map to chart their course forward. For example, chemical and tool suppliers could work jointly with semiconductor fabs to optimize processing recipes to lower emissions. Similarly, abatement tool suppliers could work closely with manufacturing equipment suppliers, fab operations, and chemical suppliers to ensure that they fulfill new abatement requirements.
To reduce Scope 2 emissions, companies would benefit from addressing challenges in sourcing renewable energy, and a joint effort with local electricity consumers could help increase availability.
The semiconductor industry is at a unique juncture: demand for products is soaring to unprecedented heights just as pressure to reduce emissions is sharply increasing. While some companies might be tempted to focus on meeting customer demand, global warming also requires companies to consider new ways of working. Those that hesitate may find that they are unaligned with customers that prioritize sustainability and consumers who want to go green. Both individual and collective actions by semiconductor players can help the entire industry increase its sustainability effort and meet the 1.5°C challenge.
We will provide further details and perspectives on reducing semiconductor emissions in future publications.