EMD Electronics, the North American electronics business of Merck KGaA, Darmstadt, Germany, develops materials and process technologies used across semiconductor manufacturing, including thin films, planarization, specialty gases, delivery systems, and advanced process technologies. As AI hardware moves toward smaller nodes, higher transistor densities, advanced packaging, and increasingly demanding power and performance requirements, materials science is becoming a critical part of semiconductor scaling. Pulse 2.0 interviewed EMD Electronics Metallics Business Technology Director Sergei Ivanov to learn more about the growing importance of materials innovation, molybdenum-based metallization, advanced packaging, and collaboration between materials scientists and chip designers.
Sergei Ivanov’s Background
When asked about his background, Ivanov shared:
Chemistry has been the through line of my entire career.
I completed my PhD in Inorganic Chemistry at the Institute of General and Inorganic Chemistry of the Russian Academy of Sciences, and my foundation in Chemical Engineering comes from Mendeleev University of Chemical Technology.
That training instilled in me a very precise way of thinking, understanding materials at the molecular level. How they’re structured, how they behave under real conditions, and critically, how that behavior can be engineered to meet specific performance demands.
What’s shaped my perspective over more than 25 years is the breadth of industries I’ve worked across, from catalysis, lithium-ion batteries, fuel cells, and ultimately semiconductor devices.
Each operates with its own constraints and definition of success. Working across all of them has given me a cross-industry fluency that I think is valuable when you’re trying to develop technology that has to perform at the frontier.
Along the way, I’ve contributed to more than 35 granted U.S. patents and over 25 peer-reviewed publications.
In my role as Metallics Business Technology Director at EMD Electronics, I lead a global R&D team developing novel precursors for thin film deposition in advanced semiconductor devices.
It’s work that sits at the intersection of chemistry and semiconductor technology. That combination is what makes it so exciting.
Why Materials Science Is Becoming Critical To AI Hardware
When asked why materials science is becoming one of the biggest bottlenecks and opportunities in scaling next-generation AI hardware, Ivanov explained:
Because at leading-edge nodes, the industry is no longer limited primarily by lithography; it’s limited by materials physics.
As we push into sub-5 nm scaling and more complex device architectures, the “classic” materials stack starts to hit hard limits: resistivity rises as features shrink, reliability windows tighten, and interfaces become more defect-sensitive.
At that scale, small changes in film quality, impurities, or interfacial chemistry can have outsized impact on performance, yield, and lifetime.
That’s also exactly why it’s such a major opportunity.
The most practical path forward runs through materials innovation: new conductors, new barrier and liner approaches, improved dielectrics, and more precise deposition and etch chemistries.
In other words, materials aren’t just enabling the next step, they are the next step.
And AI is amplifying the urgency.
The demand curve for compute, bandwidth, and energy efficiency is relentless, and materials science is what allows the industry to meet those requirements without trading away yield, reliability, or manufacturability.
How Shrinking Nodes Are Changing Semiconductor Materials
When asked how shrinking nodes and increasing transistor density are changing the role of materials in semiconductor manufacturing, Ivanov said:
At advanced nodes, the margin for error essentially disappears.
When features are only a few nanometers wide, variations of a few atoms in a dielectric, barrier, or metal film can determine whether a device meets spec or fails.
Materials have shifted from being one part of a broader process recipe to being the defining variable in whether the process is viable at all.
As dimensions shrink, traditional conductors face fundamental challenges.
Higher current densities accelerate failure mechanisms like electromigration, while scaling drives resistivity increases that directly impact speed and power.
The industry is also more sensitive to surface chemistry and interface quality than ever before, because a larger fraction of the device is effectively “interface.”
That’s why new materials and new deposition approaches are becoming critical, and why our thin film portfolio, including spin-on dielectrics, organosilanes, and organometallic precursors, is engineered for the precision and repeatability advanced nodes demand.
The Role Of Molybdenum In Chip Performance
When asked how integrated molybdenum solutions can enable tighter geometries and improved chip performance, Ivanov explained:
Molybdenum is one of the most promising developments in semiconductor metallization in years because it addresses a key scaling problem: at nanoscale dimensions, electrical resistance becomes a performance limiter.
Tungsten has been a workhorse metal for decades, but as geometries tighten, its resistance and integration tradeoffs become more constraining.
Molybdenum can deliver lower resistance at those same dimensions, which translates into better performance and improved energy efficiency, both critical for AI workloads.
It also supports tighter geometries by helping reduce RC delay and enabling more favorable scaling paths for interconnect and related structures.
Just as important, adoption needs to be realistic for high-volume manufacturing.
Integrated molybdenum solutions are compelling because they can be designed to fit into existing process ecosystems, so fabs can capture performance benefits without having to reinvent the entire process flow.
Why Advanced Packaging Requires Greater Precision
When asked why advanced packaging requires greater precision at the atomic and molecular levels, Ivanov said:
Advanced packaging has become a major engine of semiconductor progress, especially for AI, because it enables heterogeneous integration: chiplets, 2.5D and 3D architectures, and high-bandwidth memory stacks.
But the moment you start stacking and connecting dies, every interface becomes a critical reliability and performance surface.
Interconnect pitch is tighter, thermal and mechanical stresses are higher, and tolerances are dramatically narrower than in traditional packaging.
That means surface preparation, dielectric deposition, planarization, or CMP, and cleaning have to be executed with exceptional control.
At this level, “small” issues like trace contamination, surface roughness, or imperfect adhesion can become yield-killing defects or long-term reliability risks.
Atomic and molecular precision isn’t a nice-to-have; it’s a prerequisite.
Collaboration Between Materials Scientists And Chip Designers
When asked how collaboration between materials scientists and chip designers is accelerating progress in AI hardware development, Ivanov explained:
The old model, where designers define an architecture and materials teams are asked to make it manufacturable downstream, is increasingly too slow and too risky.
At the leading edge, design intent and materials capability have to evolve together, especially for technologies like atomic layer etching, or ALE, and area-selective deposition, or ASD, where success depends on a shared understanding of chemistry, surfaces, process windows, and device integration.
When materials scientists understand the electrical and architectural targets, and chip designers understand the constraints and opportunities the materials create, teams can converge faster on solutions that are both high-performing and manufacturable.
That reduces iteration cycles, improves yield learning, and shortens time to market.
It’s also where the breakthroughs happen, not in isolation, but at the intersection of design requirements and materials-enabled process innovation.
That’s the collaboration model we emphasize in how we work with customers every day.
Taking A Holistic Approach To Semiconductor Materials
Ivanov concluded by discussing the importance of considering materials and process integration holistically:
One topic I’d add is the value of a holistic approach to materials and process integration.
People sometimes think of materials suppliers in terms of a single product category, but what differentiates EMD Electronics is the breadth of technologies and process expertise we bring across the semiconductor flow, including thin films, planarization, specialty gases, delivery systems, and advanced process technologies.
That breadth matters because many of today’s hardest problems aren’t isolated; they’re system problems spanning multiple steps and multiple interfaces.
As AI-driven demand accelerates and architectures become more complex, integrated collaboration across the flow becomes increasingly important for managing manufacturing complexity and scaling reliably.
If helpful, I’m also happy to discuss where we see the next major inflection points, whether in interconnect, selective processes, or advanced packaging materials.
