ShowDaily Day1 - Flipbook - Page 12
sponsored by
Semiconductor
continued from page 10
refrigerant emissions from chillers
and other cooling equipment.
Refrigerants are normally contained
in a closed refrigeration circuit, but
losses can occur through leakage,
servicing and end-of-life handling. “It is quite different from
process-gas emissions, but it is
still part of the Scope 1 picture,”
Jones said. “If refrigerant has to be
topped up repeatedly, that tells you
material is being lost somewhere.
The way the accounting works
today, any refrigerant replaced
during these top-offs is assumed
to have been released. This fails
to account for the effects of mitigation efforts such as containment,
leak reduction and recovery during
servicing. These can significantly
reduce the material released and
deserve more emphasis.” HFC
phase-down policies, including the
Kigali Amendment and the U.S.
AIM Act, are also accelerating
the move toward refrigerants with
lower-GWP.
Beyond greenhouse gases
Jones cautioned against tunnel
vision on greenhouse gases alone.
“We mustn’t forget non-greenhouse-gas emissions,” he said.
“NOx and particulates are highly
regulated because they can affect
the local environment and community quickly. And remember what
a fab is set up for — safe operation,
throughout the system, whether
that’s the tool, the vacuum pump,
the abatement, the ductwork or the
wastewater plant. Greenhouse gases
aren’t typically the safety driver;
flammable, toxic and corrosive
gases are.”
Energy is the other piece Jones
wants the industry to weigh more
heavily. Scope 2 emissions industry-wide run roughly twice the
process chemistry side, he said, and
globally, fabs consume on the order of
200 to 220 terawatt-hours of energy
a year. The subfab components that
Edwards and its competitors sell to
device manufacturers account for
roughly 25% of a fab’s total energy
use. “Renewable energy helps cut the
emissions from that consumption,
but it doesn’t reduce the consumption
itself,” he said.
A system-wide scorecard
The throughline of Jones’s presentation, and of the case studies in
it — covering cryogenic cooling and
cryopump support, turbomolecular
and dry pump performance, predictive maintenance, and equipment
retrofit versus replacement — is that
no single technology or metric tells
the whole story.
“The question isn’t whether one
technology performs efficiently in
isolation,” Jones said. “It’s whether
the overall support system enables
reliable, scalable, lower-impact manufacturing. The industry needs to move
from individual metrics toward system-wide performance assessment.”
Jones will present the full framework
today from 1:37 to 1:54 p.m. in the
Sustainability Pavilion (South Hall,
Lower Level, Room 1). Edwards is at
Booth 633 in the South Hall.
Taking a Holistic Approach to Co-Packaged
Optics Manufacturing
THOMAS ACHLEITNER, B u s i n e s s D e v e l o p m e n t M a n a g e r , E V G r o u p
Moving co-packaged optics into high-volume manufacturing will take
expertise across bonding, nanoimprint lithography and heterogeneous
integration, not advances in any single process.
The rapid expansion of artificial intelligence (AI) is changing not only the
amount of computing power required
in data centers, but also the way that
computing resources must communicate. As AI systems scale from
individual processors to increasingly
12 | Tuesday, October 13
large clusters of accelerators, moving
enormous amounts of data between
those devices is becoming a critical
system-level challenge. Conventional
electrical interconnects face
increasing limitations in bandwidth,
reach, signal integrity and power
efficiency as data rates rise, driving
the need for optical interconnects as a
means of moving data at higher bandwidths more efficiently over longer
distances.
Co-packaged optics (CPO)
addresses these
Continued on page 14
www.semiconductordigest.com
