Semiconductors Moving towards Net Zero Emissions
As semiconductor companies ramp up production to meet surging
demand, greenhouse-gas reduction deserves equal emphasis.
·
Apple, Google, and Microsoft, have
committed to reaching net-zero emissions for their full value chain
·
Infineon plans to reduce
greenhouse-gas (GHG) emissions by 70 percent by 2025, compared with its 2019
baseline, and aspires to reach carbon neutrality for emissions directly under
its control by the end of 2030
·
Several semiconductor players have
also committed to science-based targets, including STMicroelectronics, NXP, and
UMC
·
Scope 2 emissions, which represent
the highest proportion of GHG from semiconductor companies, are linked to the
energy required to run their extensive production facilities
·
Tool fleets containing hundreds of
manufacturing tools, such as lithography equipment, ion implanters, and
high-temperature furnaces
·
Large clean rooms requiring climate
and humidity control with overpressure and particle filtration
·
Extensive subfab facilities for gas
abatement, exhaust pumps, water chillers, and water purification
·
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.
·
High-GWP heat transfer fluids that
may leak into the atmosphere when fabs use them in chillers to control wafer
temperature during manufacturing processes.
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 percent 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 percent 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 (Exhibit 1).1 Scope 2 emissions,
which represent the highest proportion of GHG from semiconductor companies, are
linked to the energy required to run their extensive production facilities. The
sources of these emissions include the following:
·
tool
fleets containing hundreds of manufacturing tools, such as lithography equipment,
ion implanters, and high-temperature furnaces
·
large
clean rooms requiring climate and humidity control with overpressure and particle
filtration
·
extensive
subfab facilities for gas abatement, exhaust pumps, water chillers, and water purification
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. Fabs can also
help decrease GHG emissions during operation by encouraging experts to share their
knowledge of and experience with energy efficiency, using the same tools and mechanisms
they employ when sharing strategies for optimizing OEE. To boost their odds of succeeding,
fabs can include additional stakeholders in their power-consumption efforts, focusing
on external tool suppliers that might be able to modify their equipment to increase
efficiency, or those that offer retrofits or energy-saving options for new tools.
To ensure that sufficient power is always
available, fabs often source their electricity from a combination of on-grid and
off-grid sources. Most off-grid power comes from fab-owned fossil fuel power plants.
Over the short term, fabs can significantly reduce the energy consumption of these
plants by pursuing efficiency improvements or switching to alternative fuels such
as biogas or green hydrogen. They can generate further gains by developing new off-grid
power sources that rely on green technologies, such as photovoltaics, fuel cells,
and battery energy storage systems. But these supplies often only complement, rather
than replace, a fab’s long-standing on-grid sources.
For on-grid power, fabs may be able
to reduce consumption by purchasing renewable electricity from utilities through
green premium energy offerings, although the offerings vary widely by region. In
Europe and the United States, for instance, renewable on-grid sourcing is readily
available and accounts for up to 31 percent of grid energy; in many parts of Asia,
however, renewable on-grid sourcing can be challenging because of limited availability
(Exhibit 2).
Access to renewable energy may be a
major factor as companies decide where they should build new fabs—something that
is becoming more common as they try to increase capacity to alleviate the chip shortage.
Process gases can significantly increase
global warming (Exhibit 3). Their emissions will vary based on a fab’s age and the
sophistication of its abatement technology, but all facilities face some common
challenges. Four levers, which are at different stages of maturity, may help reduce
process-gas related emissions. Because of economic constraints and other issues,
some fabs may not be able to apply some of these levers widely until the technology
improves.
Exhibit
3
Process improvements. Fabs can reduce emissions by adjusting
process parameters, such as temperature and chamber pressure. Process engineers
often overlook this lever and instead focus solely on yield during optimization
efforts, partly because they lack the knowledge and operational experience required
to identify strategies for reducing GHG emissions. Similarly, the suppliers involved
in daily tool operations and maintenance may prioritize cost and uptime targets
over energy savings. If fabs address knowledge gaps and collaborate more closely
with tool suppliers, they may improve emissions—for instance, by simultaneously
optimizing yield and energy consumption during cleaning protocols.
Use of alternative chemistries. Fabs can sometimes lower emissions
by switching to chemicals that have a lower environmental impact. However, they
often encounter obstacles when attempting to use these chemistries; for example,
it can be difficult to get suppliers on board with their plans. In addition, developing
new solutions is both costly and time consuming, as is the process for qualifying
new chemicals on existing processes and tools.
To address these roadblocks, fabs could
emphasize that GHG reduction is a top priority when communicating with suppliers
and note that they want to explore new solutions, including green-chemistry approaches.
Fabs could then work closely with their partners to develop a path forward. For
instance, fabs and suppliers could jointly create road maps for process-gas substitutions
or codevelop alternatives. But even with this push, fabs must temper their short-term
expectations about GHG improvement because few green solutions are now viable alternatives
to current gases. (While fabs have already implemented some major improvements,
such as increased use of NF3, many other shifts, including the replacement of NF3
with F2 or ozone, are still nascent.) Over the long term, fabs could see more gains
if they continue encouraging suppliers to explore new solutions.
Gas abatement. Gas abatement will be the main lever
to address emissions from process gases over the short to midterm, and this will
remain the case until alternative gases with fewer emissions are available, or until
gas recycling is widely adopted. Fabs that want to increase gas abatement can select
from multiple options, including point-of-use (POU) systems for individual production
tools, point-of-area (POA) systems, and central abatement systems. As they roll
out and install new systems, fabs must balance the trade-offs related to cost, impact
on operations, destruction and removal efficiency (DRE), and timing. Further adoption
will also require suppliers to provide innovative technical solutions to do the
following:
·
Address
space constraints, especially in older, 200-millimeter fabs with limited subfab
space (for instance, by providing integrated solutions with a smaller footprint).
·
Keep
by-products, such as nitrogen oxides and carbon monoxide, low while increasing DRE.
·
Prevent
downtime of production tools during installation and maintenance of abatement systems.
·
Enable
regular system qualification, without affecting production, to achieve and report
higher DREs.
Gas recycling. Fabs can capture unutilized process
gases and by-products through various means, such as membrane separation, cryogenic
recovery, adsorption, and desorption. They can then refine them into pure process
gases that can be used again, potentially reducing process-gas emissions. For this
lever to become economically viable, researchers will need to address major challenges
related to the separation of process-gas outflows and purification.
To help the semiconductor industry meet
the critical challenge of reducing GHG emissions, fabs can take four steps to accelerate
their decarbonization efforts:
·
Create
transparency about their scope 1, 2, and 3 (upstream) emissions.
·
Set
near-term and long-term emission targets.
·
Consolidate
existing ideas and estimate their expected costs and impact. This will involve defining
a portfolio of innovative technologies to be developed jointly with external partners
over the next few years.
·
Generate
an abatement cost curve to serve as a road map for short-, mid-, and long-term decarbonization
efforts.
The abatement cost curve can help fabs
identify all potential areas for GHG reduction, as well as associated implementation
costs and savings. Fabs should first implement the levers that would result in a
net cost reduction, and then move to those with net abatement costs that are close
to zero and those that will be costly until regulations change or other advances
make implementation less expensive. Exhibit 4 shows an example abatement cost curve
for a typical fab.
As they embark on their net-zero programs,
semiconductor companies will benefit from internal collaboration among staff from
R&D, operations, supply chain, and other functions. Among other advantages,
such efforts will help ensure that they implement key elements of the decarbonization
program simultaneously. External engagement is also essential, since no single company
can reach its GHG goals without partners, such as industry peers, suppliers, and
customers.