Chapter 3
Semiconductor Materials
Chapter 3: Semiconductor Materials
3.1 Overview
Every semiconductor chip begins as a raw material. Before a transistor is etched, before a photon of EUV light touches a wafer, a supply chain of extraordinary precision must deliver materials at purities measured in parts-per-trillion. This chapter covers the materials that sit between raw industrial inputs (Chapter 2) and the equipment that processes them (Chapter 4): silicon wafers, compound semiconductors, photoresists, CMP slurries, electronic chemicals, ultra-pure gases, photomasks, and sputtering targets.
What makes this layer strategically important is concentration. The companies that produce these materials are few, the qualification barriers are high (switching a photoresist supplier at a leading-edge fab can take 2-3 years of testing), and the materials themselves are a small fraction of total chip cost but absolutely mission-critical. A single contamination event in a photoresist batch can destroy an entire production lot worth hundreds of millions of dollars. This combination of low cost share, extreme quality requirements, and limited suppliers gives incumbents structural pricing power.
The AI buildout amplifies demand for these materials in two ways. First, more chips means more wafers, more chemicals, more gases. Second, the shift to advanced nodes (3nm, 2nm) and advanced packaging (CoWoS, 3D stacking) requires newer, higher-specification materials that only a handful of suppliers can produce.
3.2 Market Sizing & Growth
The semiconductor materials market encompasses multiple sub-segments:
Silicon wafers: Global market valued at approximately $14.5-15.9B in 2024, projected to reach $20-26B by 2030-2032 depending on the source, at a CAGR of 5-9% 123. The top five manufacturers (Shin-Etsu, SUMCO, GlobalWafers, Siltronic, SK Siltron) control approximately 82-85% of revenue and an even higher share of 300mm capacity 145. 300mm wafers represent roughly 75% of market value and are the standard for advanced logic and memory 1.
Photoresists: Global photoresist and ancillaries market estimated at $4.1-5.5B in 2023-2024, projected to reach $5.3-11B by 2028-2034, at a CAGR of 5-11% depending on whether the estimate includes only advanced lithography resists or all photoresists 67. EUV photoresists are the fastest-growing segment, representing over 25% of advanced photoresist revenues in 2024 and expected to exceed 45% by 2030 7.
Electronic specialty gases: Global market valued at approximately $5.1-16B in 2024-2025 (wide range reflects different scope definitions), projected to grow at 4-11% CAGR through 2032-2035 8910. The big three industrial gas companies (Linde, Air Liquide, Air Products) collectively account for over 35% of the semiconductor specialty gas market 10.
CMP slurries and pads: Approximately $2-3B market, dominated by Entegris (which acquired CMC Materials) and Fujimi 11.
Total semiconductor materials market (wafers + gases + chemicals + photoresists + CMP + targets + photomasks): SEMI estimates the total at roughly $66-70B in 2024 18.
3.3 Supply Chain Flowchart
RAW SILICON (Polysilicon)
|
v
CRYSTAL GROWTH (Czochralski / Float Zone)
|
v
SILICON INGOT
|
v
WAFER SLICING, LAPPING, POLISHING
| Shin-Etsu, SUMCO, GlobalWafers, Siltronic, SK Siltron
|
v
BARE SILICON WAFER ------------------------------------------+
| |
v v
EPITAXIAL GROWTH (optional) COMPOUND SEMI WAFERS
| Shin-Etsu, SUMCO, Soitec (SOI) SiC: Wolfspeed, Coherent, ON Semi
| GaN: Wolfspeed, Qorvo, MACOM
v InP: Sivers Photonics, IQE
PROCESSED WAFER GaAs: WIN Semi, IQE
|
|---> PHOTORESIST APPLICATION
| JSR, TOK, Shin-Etsu Chemical, Fujifilm, Dongjin Semichem
|
|---> ETCHING & DEPOSITION GASES
| Linde, Air Liquide, Air Products, Taiyo Nippon Sanso
| SK Specialty, Showa Denko, Mitsui Chemicals
|
|---> CMP SLURRIES & POLISHING
| Entegris (CMC Materials), Fujimi, DuPont, JSR
|
|---> ULTRA-PURE CHEMICALS & FILTRATION
| Entegris, Fujifilm, Merck KGaA, Soulbrain, ENF Technology
|
|---> SPUTTERING TARGETS & DEPOSITION MATERIALS
| Materion, JX Nippon Mining, Tosoh, Heraeus
|
+---> PHOTOMASKS
Photronics, Toppan, DNP, HOYA (blanks), Shin-Etsu (blanks)
|
v
TO SEMICONDUCTOR EQUIPMENT (Chapter 4) & FOUNDRIES (Chapter 7)
3.4 Key Companies
3.4.1 Silicon Wafers
| Company | Ticker | Exchange | Approx. Mkt Cap | Role | Key Metric |
|---|---|---|---|---|---|
| Shin-Etsu Chemical | 4063 | TSE | ~$93.0B | Largest silicon wafer producer globally; also photomask blanks, photoresists, PVC | ~18% wafer market share; invested JPY 150B ($1B) in 2025 for 2nm/3nm capacity 5 |
| SUMCO | 3436 | TSE | ~$8.7B | Pure-play silicon wafer manufacturer, #2 globally | ~17% market share; invested JPY 229B ($2.05B) in 300mm capacity expansion 12 |
| GlobalWafers | 6488 | TWSE | ~$8.0B | #3 globally; Taiwan-based; acquired Siltronic bid failed | $5B Texas 300mm plant targeting 1.2M wafers/yr by 2027; received $406M US CHIPS Act subsidy (Dec 2024) 35 |
| Siltronic AG | WAF | XETRA | ~$3.3B | European leader; majority owned by Wacker Chemie | Focus on ultra-flat, defect-minimized substrates for advanced nodes |
| SK Siltron | (SK Group subsidiary) | Private | Private sub. | Korean wafer maker; acquired DuPont SiC wafer business | 1.2M+ 300mm wafers/month capacity; SiC wafer capabilities via DuPont acquisition |
| Soitec | SOI | Euronext Paris | ~$6.1B | SOI (silicon-on-insulator) wafer specialist; Smart Cut technology | Dominates SOI market; 18 patents filed in 2024-2025 for defect reduction 5 |
| National Silicon Industry Group (NSIG) | 688126 | Shanghai SSE STAR | ~$10.5B | Largest Chinese silicon wafer maker | Focused on mature-grade wafers; pricing 10-15% below Japanese peers 5 |
3.4.2 Compound Semiconductors (SiC, GaN, InP, GaAs)
| Company | Ticker | Exchange | Approx. Mkt Cap | Role | Key Metric |
|---|---|---|---|---|---|
| Wolfspeed | WOLF | NYSE | ~$2.0B | Leading SiC wafer and device manufacturer | Mohawk Valley 200mm SiC fab; transitioning from 150mm to 200mm SiC wafers |
| Coherent (fmr. II-VI) | COHR | NYSE | ~$65.6B | SiC substrates + photonics; key Sivers competitor in InP | NVIDIA invested $2B in early 2026; multi-billion purchase agreement 13 |
| ON Semiconductor | ON | NASDAQ | ~$40.4B | SiC power devices and substrates for EV/industrial | Growing SiC revenue; vertically integrated from substrate to module |
| Infineon Technologies | IFX | XETRA | ~$89.0B | SiC and GaN power semiconductors | Largest power semiconductor company globally |
| STMicroelectronics | STM | Euronext Paris / NYSE | ~$52.6B | SiC for automotive (Tesla supplier) | Major SiC revenue from automotive power electronics |
| IQE | IQE | LSE AIM | ~$580M | Compound semiconductor epitaxial wafer specialist | Serves RF, photonics, and power electronics markets. Key InP supplier for CPO laser ecosystem. |
| AXT Inc. | AXTI | NASDAQ | ~$7.6B | InP, GaAs, and Ge substrate manufacturer | One of the few InP substrate suppliers globally. InP substrates are critical for the EML lasers described in Chapter 11. |
| Sumitomo Electric | 5802 | TSE | ~$59.0B | InP epitaxial wafers, optical fiber, power cables | Cross-layer company: InP for photonics (Chapter 11), fiber for interconnects (Chapter 12), power cables. Vertically integrated InP supply chain. |
| WIN Semiconductors | 3105 | TWSE | ~$5.4B | Pure-play GaAs/GaN foundry | Manufacturing partner for Sivers Semiconductors’ DFB lasers 14 |
| MACOM Technology | MTSI | NASDAQ | ~$27.5B | GaN-on-Si for telecom, defense, data center | GaN technology for high-frequency applications |
| Qorvo | QRVO | NASDAQ | ~$7.0B | GaN RF devices for defense and 5G | Leading GaN RF supplier |
| Rohm | 6963 | TSE | ~$6.0B | SiC power devices | Japanese SiC leader; vertically integrated |
3.4.3 Photoresists & Lithography Chemicals
| Company | Ticker | Exchange | Approx. Mkt Cap | Role | Key Metric |
|---|---|---|---|---|---|
| JSR Corporation | Private | Taken private (JIC) | ~$6.3B (acquisition price) | Leading photoresist maker; EUV resist pioneer | Taken private by Japan Investment Corp in 2023 for ~$6.3B; ~91% Japanese market share (with TOK) 615 |
| Tokyo Ohka Kogyo (TOK) | 4186 | TSE | ~$7.7B | #1 or #2 photoresist maker globally | Investing ¥20B ($130M) in new South Korea photoresist plant for 2030; ¥12B in Korean high-purity chemicals facility 15 |
| Shin-Etsu Chemical | 4063 | TSE | ~$93.0B | Photoresists (plus wafers, PVC, photomask blanks) | Diversified across multiple semiconductor materials |
| Fujifilm Holdings | 4901 | TSE | ~$24.0B | Photoresists; acquired CMC Materials’ KMG process chemicals from Entegris | Expanding in EUV resists; acquired Entegris’ KMG unit for ~$700M 11 |
| Dongjin Semichem | 005290 | KRX | ~$2.0B | South Korean photoresist producer | Growing share in ArF and KrF resists; one of top five globally 7 |
| Adeka | 4401 | TSE | ~$3.0B | Investing in metal oxide resists for high-NA EUV | Investing ¥3.2B in new photoresist materials for next-gen EUV 15 |
| Brewer Science | Private | Private | Private | Anti-reflective coatings, specialty lithography materials (US) | Niche but important for advanced lithography stacks |
Critical note on photoresist concentration: Japanese producers (JSR, TOK, Shin-Etsu, Fujifilm) together account for approximately 91% of the global photoresist market 15. For EUV photoresists specifically, concentration is even higher. This is one of the most extreme single-country dependencies in the entire semiconductor supply chain. A major earthquake affecting Shin-Etsu’s Shirakawa facility or TOK’s Koriyama plant could halt leading-edge chip production globally.
3.4.4 CMP, Ultra-Pure Chemicals & Contamination Control
| Company | Ticker | Exchange | Approx. Mkt Cap | Role | Key Metric |
|---|---|---|---|---|---|
| Entegris | ENTG | NASDAQ | ~$22.7B | Dominant in filtration, fluid handling, CMP slurries (via CMC Materials acquisition), contamination control | Acquired CMC Materials for $6.5B (2022) 11; sole-source for many fab contamination control systems |
| Merck KGaA | MRK | XETRA | {{MRK.market_cap}} | Electronic materials division: high-purity solvents, specialty chemicals, thin-film materials | Investing $300-500M in India specialty chemicals plant 10 |
| DuPont (Electronics & Industrial) | DD | NYSE | ~$20.4B | CMP slurries, photoresists, advanced materials | Significant semiconductor materials portfolio |
| Fujimi Incorporated | 5384 | TSE | ~$2.0B | CMP slurries and polishing compounds | Japanese specialist; serves major foundries |
| Soulbrain | 357780 | KRX | ~$3.0B | High-purity electronic-grade chemicals (etchants, cleaning agents) | South Korean specialist for cleaning/etching steps |
| ENF Technology | 102710 | KRX | ~$500M | High-purity etchants for semiconductor manufacturing | Korean specialist; growing share in wet etch chemicals |
3.4.5 Electronic & Specialty Gases
| Company | Ticker | Exchange | Approx. Mkt Cap | Role | Key Metric |
|---|---|---|---|---|---|
| Linde | LIN | NASDAQ | ~$220B | Largest industrial gas company; ultra-high-purity N₂, Ar, specialty gases for fabs | >35% semiconductor specialty gas market share (with Air Liquide, Air Products) 10; expanding supply to Samsung Pyeongtaek 16 |
| Air Liquide | AI | Euronext Paris | ~$100B | #2 industrial gas; electronics division serves major fabs | Investing $250M+ in US fab gas supply facility 9 |
| Air Products & Chemicals | APD | NYSE | ~$65.0B | #3 industrial gas; hydrogen and specialty gas supply | Partnership with semiconductor manufacturers for advanced gas delivery systems 8 |
| Taiyo Nippon Sanso | 4091 | TSE | ~$5.4B | Japanese industrial gas; electronics specialty gases | Modular specialty gas plants near Asian fabs |
| SK Specialty | Private | Private (SK Group) | Private | Electronic specialty gases for semiconductors and displays | Sold 85% stake to Hahn & Company for $1.86B (Dec 2024) 10 |
| Showa Denko (now Resonac) | 4004 | TSE | ~$8.0B | Specialty gases, high-purity chemicals, SiC epitaxial wafers | Rebranded to Resonac; diversified semiconductor materials |
3.4.6 Photomasks
| Company | Ticker | Exchange | Approx. Mkt Cap | Role | Key Metric |
|---|---|---|---|---|---|
| Photronics | PLAB | NASDAQ | ~$3.0B | One of two major merchant photomask manufacturers globally | Serves Intel, TSMC, Samsung; US-based |
| Toppan Holdings | 7911 | TSE | ~$9.3B | Major photomask producer (Japanese) | Large share of advanced photomask market |
| Dai Nippon Printing (DNP) | 7912 | TSE | ~$9.1B | Photomask producer plus EUV pellicles | Developing EUV pellicle technology (critical for EUV yield) |
| Mitsui Chemicals | 4183 | TSE | ~$10.0B | Leading commercial EUV pellicle manufacturer (CNT-based) | 70-80% of commercial EUV pellicle market (with Shin-Etsu, S&S Tech). First company to commercialize CNT pellicles for ASML systems. EUV pellicles prevent particle contamination during exposure; without them, EUV yields drop sharply. Strategic partnership with imec. Pellicle market: ~$194M (2024) to ~$551M (2032). |
| HOYA | 7741 | TSE | ~$62.0B | Photomask blanks (glass substrates for EUV masks) | ~75% market share in EUV mask blanks specifically; only vendor with validated blanks for High-NA EUV (sub-2nm). AGC Inc. (5201.T) holds ~59% of overall EUV mask market but Hoya dominates the blank substrate tier. Critical sole-source for next-gen lithography. |
3.4.7 Sputtering Targets & Deposition Materials
| Company | Ticker | Exchange | Approx. Mkt Cap | Role | Key Metric |
|---|---|---|---|---|---|
| Materion | MTRN | NYSE | ~$4.0B | Sputtering targets, advanced materials for thin film deposition | US-based; serves semiconductor and defense |
| JX Nippon Mining & Metals | (ENEOS Holdings sub.) | Private | Sub. of 5020.TSE | Sputtering targets, copper foil, specialty metals | Japanese; critical for copper interconnect deposition |
| Tosoh | 4042 | TSE | ~$5.0B | Sputtering targets (especially ITO for displays), specialty chemicals | Also produces zirconia and specialty ceramics |
| Heraeus | Private | Private | Private | Precious metal sputtering targets, bonding wire | German; private; major supplier of gold and silver targets |
3.5 Bottleneck Analysis
Where does scarcity sit in this layer?
Photoresists (SEVERE): Photoresist is a low-cost input per wafer application (estimated at low single-digit percentages of total wafer processing cost), but it is entirely non-substitutable at advanced nodes, and the 91% Japanese concentration is the most concentrated dependency in semiconductor materials 15. EUV photoresists are even more concentrated. Qualification of a new photoresist at a leading-edge fab takes 2-3 years because the resist must be tested across hundreds of process steps. A customer cannot simply switch suppliers when capacity is tight. JSR’s privatization by the Japanese government (via JIC) signals that Tokyo views photoresist supply as a national security asset.
Silicon wafers (MODERATE-HIGH): The top five control 82-85% of the market 14. However, capacity is expanding (Shin-Etsu/SUMCO $1B investment, GlobalWafers $5B Texas plant) and the wafer itself is a more standardized product than photoresist. The bottleneck is tighter at the 300mm leading-edge end, where the top five control even more.
Electronic specialty gases (MODERATE): The big three (Linde, Air Liquide, Air Products) dominate, but they are actively expanding capacity (new plants in US, Taiwan, India) and the product is somewhat more commoditized than photoresists. The real bottleneck is in ultra-high-purity grades (6N = 99.9999%) 17 required for sub-5nm nodes, where purification technology is proprietary.
Contamination control / filtration (MODERATE-HIGH): Entegris has a near-monopoly on certain critical filtration and fluid handling systems for leading-edge fabs. After acquiring CMC Materials ($6.5B) 11 and integrating those CMP capabilities, Entegris controls a uniquely broad slice of the fab consumables chain. Switching costs are extremely high because contamination control systems are qualified as part of the fab’s overall process recipe.
Compound semiconductors (MODERATE for AI buildout): SiC and GaN are critical for power electronics in data centers (power delivery, conversion efficiency), but the AI buildout’s direct demand for compound semis is smaller than for silicon. InP is more directly relevant to the AI buildout via optical interconnects (see Chapter 11).
3.6 Risks
Geographic concentration risk: Japan dominates photoresists (~91%) 15, silicon wafers (Shin-Etsu + SUMCO together >35% share) 15, and several specialty chemical categories. An earthquake, volcanic event, or geopolitical disruption affecting Japanese production facilities could create cascading shortages across the global semiconductor supply chain. The 2011 Tohoku earthquake caused a photoresist shortage that took weeks to resolve, with Tokyo Ohka Kogyo’s Koriyama plant (a significant share of global capacity) shutting down temporarily.
Chinese substitution: China is investing heavily in domestic semiconductor materials. Companies like NSIG (silicon wafers), Shennan Circuit/FastPrint (substrates), and domestic photoresist makers are gaining share at mature nodes. Chinese firms have achieved breakthroughs in i-line and KrF photoresists but remain far behind in ArF immersion and EUV resists 15. The risk is that Chinese substitution compresses margins for incumbents at the mature end while the leading edge remains protected.
Cyclicality: Semiconductor materials demand is cyclical with the broader chip cycle. The 2023-2024 downturn hit wafer and chemical volumes. However, AI-driven demand is somewhat counter-cyclical because hyperscaler capex continued even during the consumer electronics downturn.
Technology disruption: Glass core substrates (being developed by Intel and others) could eventually complement or replace ABF-based substrates for certain applications. New lithography techniques could alter photoresist chemistry requirements. These are long-term risks (5-10 year horizon) rather than near-term threats.
First principles check: Does the concentration make sense? Yes. Semiconductor materials require extreme purity (parts per trillion), years of qualification at customer fabs, massive capital investment in cleanroom-grade production facilities, and deep process knowledge that takes decades to build. These are genuine barriers, not regulatory moats that could be eliminated by policy changes. The concentration is structural.