# The Most Expensive Material Inside a Chip Factory

An episode of [The Hardware Podcast with Fexingo: Chips, Devices, and Electronics Engineering Conversations](https://invisiblenetwork.ai/podcast/1896799982), 10 Jun 2026, 9:06.

Audio: https://audio.fexingo.com/business/the-hardware-podcast/episode-0042.mp3

Episode 42 of The Hardware Podcast dives into a material most people have never heard of but is absolutely critical to every chip ever made: photoresist. Lucas and Luna explore why a single gallon of the high-end photoresist used in extreme ultraviolet lithography can cost over $2,000 — more than gold by weight. They break down the chemistry, the purity requirements that make it so expensive, and the handful of Japanese and German companies that dominate the supply. This episode also touches on the geopolitical stakes, as Taiwan and the US scramble to build domestic photoresist production. If you've ever wondered what liquid actually turns into the circuits on a chip, this is the episode. #Photoresist #EUVLithography #ChipManufacturing #SemiconductorSupplyChain #TokyoOhka #JSR #SumitomoChemical #Fujifilm #ChemicallyAmplifiedResist #Methanol #IntegratedCircuits #ChipFab #TSMC #ASML #Technology #Hardware #FexingoBusiness #BusinessPodcast Keep every episode free: buymeacoffee.com/fexingo

## Transcript

Lucas: So you know how we talk a lot about the machine that makes chips — the EUV lithography tool from ASML, which costs like $150 million a pop.
Luna: Right. The giant thing that projects tiny patterns onto silicon wafers using extreme ultraviolet light.
Lucas: Exactly. But today I want to talk about the liquid that actually captures that pattern. The stuff that gets coated onto the wafer before the light hits it. It's called photoresist.
Luna: Photoresist — yeah, it's basically a light-sensitive polymer that hardens or softens when exposed to UV, and then you etch away the unwanted parts to leave the circuit pattern.
Lucas: Right. And it's one of those things that sounds like a commodity chemical. But the photoresist used in modern EUV lithography is anything but commodity. A single gallon of the high-end stuff can cost north of $2,000.
Luna: Wait, two thousand dollars for a gallon? That's more expensive than gold by weight.
Lucas: It is. Gold is about $2,000 per troy ounce, which is 31 grams. A gallon of water is about 3,800 grams. So a gallon of gold would be around $240,000. But photoresist at $2,000 per gallon is still incredibly expensive for what is essentially a specialty paint.
Luna: So what drives that cost? Is it the raw materials or the manufacturing process?
Lucas: Both, actually. The base polymer itself isn't that exotic — it's typically a resin derived from things like polystyrene or novolac. But the purity requirements are insane. A chip fab can't have even a single particle of dust larger than about 20 nanometers in the photoresist, because that particle would create a defect on the wafer, and that defect could kill the entire batch of chips.
Luna: So they filter it to an extreme degree.
Lucas: They do. The filtration process uses membranes with pores about 10 nanometers wide. And they have to filter out not just particles but also metal ions — things like iron, copper, nickel — down to parts per trillion levels. If there's too much metal, it can contaminate the transistors and ruin their electrical properties.
Luna: So the purification alone drives up the cost significantly.
Lucas: Absolutely. But there's another layer. For EUV lithography, the photoresist needs to be chemically amplified. That means it contains a photoacid generator — a molecule that, when hit by extreme ultraviolet photons, produces a strong acid that catalyzes a chemical reaction in the polymer, making it soluble or insoluble in developer solution.
Luna: And that photoacid generator is a custom, complex molecule.
Lucas: Exactly. It's often a sulfonium or iodonium salt with a carefully designed structure. Synthesizing those with the required purity is a specialty chemistry business. And the EUV wavelength is only 13.5 nanometers, which is absorbed by almost everything, so the resist formulation has to be optimized for that wavelength — it's not off-the-shelf.
Luna: So who actually makes this stuff?
Lucas: That's the fascinating part. The photoresist market is dominated by a handful of Japanese and German companies. Tokyo Ohka Kogyo, JSR, Shin-Etsu, Fujifilm Electronic Materials — those are the big names. Together they control something like 90% of the global market. And for EUV specifically, the concentration is even higher.
Luna: Right, because the barrier to entry is huge — you need the synthesis capability, the ultra-pure filtration, and the testing infrastructure to qualify your resist for a fab like TSMC's.
Lucas: Exactly. And qualification is a multi-year process. A fab doesn't just switch resists. They test it on thousands of wafers, measure defect rates, line-width roughness, sensitivity. So once a resist is qualified, it's locked in for years. That gives the incumbents enormous pricing power.
Luna: So the cost of photoresist is a combination of R&D, purification, and market concentration.
Lucas: And it's also a matter of yield. Even a tiny defect in the resist can cause a chip to fail. So fabs will pay a premium for consistency. They'd rather spend $2,000 a gallon on a resist they trust than $500 on something that introduces variability.
Luna: That makes sense. But is there any push to bring photoresist manufacturing onshore? I mean, given the supply chain concerns with Taiwan and Japan.
Lucas: There is. The US Department of Defense has funded efforts to build domestic photoresist production. Companies like Brewer Science in Missouri and DuPont are working on it. But it's a slow process. And it's not just the resist itself — the developers, the rinse solutions, the anti-reflective coatings — all those are part of the ecosystem. So even if you make the resist, you might still depend on Japanese suppliers for other chemicals.
Luna: It's like the entire semiconductor supply chain is interconnected at the chemical level.
Lucas: It really is. And that's one reason why chip manufacturing is so concentrated. It's not just the fab equipment — it's the materials, the gases, the chemicals. All of them require immense expertise and capital to produce reliably.
Luna: Hey, speaking of that expertise and capital — this show is possible because of a small group of listeners who chip in monthly through buy me a coffee dot com slash fexingo. It keeps us ad-free and lets us dig into topics like this without chasing clicks.
Lucas: Yeah, it's a pretty direct relationship. A few people covering the cost, and we get to make the show we want. If that model works for you, it's at buy me a coffee dot com slash fexingo. No pressure at all.
Luna: All right. Back to the chemistry. So you mentioned chemically amplified resists. How does that actually work at the molecular level?
Lucas: So imagine you have a polymer that's initially insoluble in the developer. The photoacid generator is mixed in. When EUV light hits it, it creates a strong acid. That acid then diffuses through the polymer film during a post-exposure bake, catalyzing a reaction that changes the solubility. The key is that one photon can generate one acid molecule, but that one acid can catalyze hundreds or even thousands of reactions. That's the amplification — it makes the resist sensitive enough that you don't need an enormous dose of EUV light.
Luna: And EUV light is incredibly expensive to produce. So the more sensitive the resist, the faster the wafer throughput.
Lucas: Exactly. ASML's EUV tools can output about 150 to 200 wafers per hour. If you could double the resist sensitivity, you could theoretically increase throughput — or reduce the laser power needed. That's a huge economic lever.
Luna: So there's continuous R&D to make resists more sensitive while maintaining low defect rates.
Lucas: And also to reduce line-width roughness. As features shrink below 10 nanometers, any roughness in the resist pattern gets transferred to the underlying silicon, which affects transistor performance. So resist chemistry has to balance sensitivity, resolution, and roughness — the so-called RLS trade-off.
Luna: I've heard of that. It's like a triangle where you can optimize two of the three, but not all three simultaneously.
Lucas: Exactly right. And that's the frontier. For the next generation of chips — say, the 2-nanometer node and beyond — the industry needs resists that can resolve line widths of 8 nanometers or less, with roughness below 1 nanometer, and sensitivity high enough to keep EUV tool throughput economical. That is a massive chemical engineering challenge.
Luna: And the companies that solve it will own the market for years.
Lucas: That's the bet. JSR is actually being acquired by a Japanese government-backed fund because it's considered a strategic asset. The US is trying to build domestic capability. It's geopolitics at the molecular level.
Luna: So next time someone complains about chip prices, they should remember that part of that cost comes from a gallon of liquid that costs more than gold.
Lucas: Literally. And it's consumed by the liter every day in fabs around the world. It's one of those hidden costs that makes the semiconductor industry so capital intensive.
Luna: Well, I'll never look at a chip the same way again.
Lucas: Good. That means we did our job.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.