China Electronic-Grade High-Purity Silicon Chemicals 2026 — Closing the Loop on TEOS, HMDS and Polysilicon Precursors

Manufacturing Research Institute | 2026-06-19

Chapter 1　The Industry Landscape and Definition of Silicon-Based Electronic Chemicals

1.1 From Sand to Silicon Chip: Silicon's Industrial Journey

Silicon is the second most abundant element in Earth's crust, accounting for approximately 27.7% of the crust by mass. Yet silicon in nature exists as silicon dioxide, silicates, and related mineral forms — and the chasm between these natural states and the "electronic-grade" purity required by semiconductor chips is not a matter of a few orders of magnitude, but dozens. That gap is the entire reason the silicon-based electronic chemicals supply chain exists.

Transforming an ordinary grain of quartz sand into a complex integrated circuit chip requires at least seven critical silicon-material conversion steps in the industrial process:

The first is metallurgical silicon smelting (high-silicon iron, carbothermal reduction), in which carbon reduces high-purity quartz sand in an electric arc furnace, converting SiO₂ to metallurgical-grade silicon at approximately 98–99% purity — the raw-material starting point for all downstream processes. The second is trichlorosilane (SiHCl₃, TCS) synthesis, in which metallurgical silicon powder reacts with hydrogen chloride gas in a fluidised-bed reactor to produce TCS vapour. The third is reduction and purification via the modified Siemens process or a fluid-bed reactor (FBR), where H₂ reduces TCS and deposits high-purity polysilicon (9N–11N) on heated silicon rods — the most technically intensive step in the entire purification pathway. The fourth is monocrystalline silicon ingot growth, using the Czochralski (CZ) or Float-Zone (FZ) method to melt the polysilicon and pull it into a high-integrity single-crystal silicon rod. The fifth is wafer slicing, grinding, and polishing, in which the ingot is cut into wafers a few hundred micrometres thick and subjected to multiple rounds of chemical-mechanical polishing to achieve a mirror surface with roughness Ra < 0.1 nm. The sixth is the chemical vapour deposition (CVD) step in chip fabrication, where liquid-phase precursors such as tetraethyl orthosilicate (TEOS) are vaporised and delivered by an inert carrier gas (N₂ or He) into the CVD chamber, thermally decomposing at low pressure (200 Pa–10 kPa) and temperatures of 350–750 °C to deposit dense SiO₂ films with nanometre-level precision on the wafer surface, forming critical dielectric layers in the multi-level chip structure. The seventh is wafer surface treatment before the lithography step, in which hexamethyldisilazane (HMDS) is used to hydrophobically modify the silicon wafer surface, providing a stable molecular-layer foundation for photoresist coating, ensuring lithography precision and yield.

Each of these seven steps corresponds to an independent chemical category; each is a technically intensive domain where chemical engineering and materials science intersect deeply; and in each, a vast quality chasm separates industrial-grade from electronic-grade products — this is the core theme of the present research report: silicon-based electronic chemicals.

It is worth emphasising that these seven steps are not carried out by any single company. They are distributed across a highly specialised supply-chain network spanning dozens of countries and hundreds of specialist firms. It is precisely this deep division of labour that makes each category an independently sized market, and makes achieving domestic substitution breakthroughs in any given category both important and complex.

1.2 Definitions and Boundaries of the Eight Major Categories of Electronic-Grade Silicon-Based Chemicals

As used in this report, "electronic-grade silicon-based chemicals" refers primarily to liquid and solid precursors, deliberately distinguished from the electronic specialty gases segment (gas-phase products such as silane SiH₄, trichlorosilane SiHCl₃, and silicon tetrafluoride SiF₄, which are covered in a separate thematic report published by this institute). The eight major liquid/solid categories are:

(I) Tetraethyl Orthosilicate (TEOS)

Molecular formula Si(OC₂H₅)₄, molecular weight 208.33, a colourless transparent liquid with a boiling point of 168.8 °C, stable at room temperature but slowly hydrolysed by water to SiO₂ and ethanol. TEOS is the most important liquid silicon-source precursor in CVD processes. Inside a wafer fabrication plant it is atomised into vapour and delivered by an inert carrier gas into the CVD chamber, where it thermally decomposes under low pressure (200 Pa–10 kPa) at 350–750 °C, depositing dense SiO₂ films on the wafer surface with nanometre-level precision. These SiO₂ films serve multiple critical process nodes: interlayer dielectric (ILD), shallow-trench isolation (STI) fill, and passivation layers.

Impurity requirements for electronic-grade TEOS are extremely strict: metallic ions (Fe, Ni, Cr, Cu, Na, K, Al, etc.) must be controlled to ppb (parts per billion) or even ppt (parts per trillion) levels; moisture content must be below 20 ppm; particle size must pass 0.1 µm ultrafiltration; the appearance must be colourless, transparent, and free of suspended solids. Industrial-grade TEOS has far lower purity requirements and belongs to an entirely different price tier and market segment.

The global high-purity TEOS market was approximately USD 265.6 million in 2025 (another dataset cites the CVD-dedicated TEOS market alone at ~USD 158.5 million). The Asia-Pacific region accounts for roughly 84% of global consumption, with China, Taiwan, and South Korea being the most important markets. The world's top-six TEOS suppliers (Evonik, Entegris, Wacker Chemie, Dow, Soulbrain, and Nantong Chruker) collectively hold approximately 75% market share.

(II) Hexamethyldisilazane (HMDS)

Molecular formula (CH₃)₃Si-NH-Si(CH₃)₃, molecular weight 161.39, a colourless liquid with a boiling point of 125–127 °C and a strong ammonia-like odour. Highly hygroscopic — once exposed to moisture in air it decomposes into trimethylsilanol and ammonia, so storage and transport must maintain strict moisture exclusion.

HMDS plays the role of a "photoresist adhesion primer" in semiconductor manufacturing. Before the lithography step begins, the wafer undergoes a dehydration bake (typically 150–200 °C for several minutes) to remove physically adsorbed water from the surface; HMDS vapour is then introduced into a sealed chamber to treat the wafer surface. HMDS reacts with surface silanol groups (Si-OH) via silanisation:

(CH₃)₃Si-NH-Si(CH₃)₃ + 2 Si-OH → 2 Si-O-Si(CH₃)₃ + NH₃

The reaction forms a dense trimethylsiloxy self-assembled monolayer (SAM) on the wafer surface, changing its character from strongly hydrophilic to hydrophobic, raising the contact angle from <10° to above 70°. This surface modification substantially reduces surface polarity, greatly improving photoresist adhesion and reducing edge defects and pattern collapse — a key yield-management step in the lithography process.

The global HMDS market was approximately USD 178 million in 2025, rising to ~USD 190.5 million in 2026. Semiconductor applications account for roughly 46% of consumption, with the remainder going to silicone rubber processing aids and agrochemical intermediates. Electronic-grade HMDS is growing at approximately 8.4% CAGR, above the overall market, reflecting rapid purity-specification upgrades driven by the semiconductor industry.

(III) Octamethylcyclotetrasiloxane (D4)

Molecular formula [(CH₃)₂SiO]₄, molecular weight 296.62, a colourless transparent liquid with a boiling point of 175 °C, and one of the most important intermediates in the silicone industry. D4 is the core component of the hydrolysis-condensation products of dimethyldichlorosilane (DMDCS); ring-opening polymerisation yields polydimethylsiloxane (PDMS), which is further processed into silicone gels, silicone rubber, silicone oil, and functional silicone materials.

In electronics, D4-derived silicone gels and rubbers are prized for their excellent electrical insulation (volume resistivity >10¹⁵ Ω·cm), broad thermal stability (−50 °C to +200 °C, with some grades up to +250 °C), high optical transparency (visible-light transmittance >90%), and low ionic contamination. They are widely used in semiconductor packaging (power-module encapsulation, automotive chip potting), LED packaging (primary-optical silicone gel, lens-moulding rubber), MEMS device protective layers, and optical element protective coatings.

The global D4 market was approximately USD 1.34 billion in 2024, with Asia-Pacific consuming 58% of the total and China accounting for more than 45% of global production capacity. Electronic-use D4 consumption grew approximately 25% in 2022–2023 (from 16,000 tonnes to 20,000 tonnes), making it one of the fastest-growing downstream application segments.

(IV) Silane Coupling Agents

General formula X₃Si-R-Y, where X is a hydrolysable group reactive towards inorganics (methoxy -OCH₃, ethoxy -OC₂H₅, chloro -Cl, etc.), Y is a functional group reactive towards organics (amino -NH₂, epoxy, methacryloyloxy, mercapto -SH, etc.), and R is an organic spacer connecting Si and Y (typically a three-carbon propyl chain). This bifunctional "one end to inorganic, one end to organic" molecular architecture gives silane coupling agents their unique ability to act as molecular bridges at inorganic-organic interfaces.

The electronics and new-energy applications for silane coupling agents are extremely broad. In photovoltaics, grades KH-560 (epoxy) and KH-570 (methacrylate) are used to modify EVA/POE encapsulant films, improving interfacial bonding between the encapsulant and the glass cover and silicon cells, ensuring 25-year module reliability. In lithium-ion battery separators, KH-570 is used to couple nano-Al₂O₃/boehmite coatings to PE/PP base films. In chip packaging, epoxy- or mercapto-functional coupling agents improve adhesion between the die and epoxy moulding compound (EMC). In optical fibre, KH-550 (amino) improves adhesion of the fibre coating to the quartz glass surface.

The global silane coupling agent market was approximately USD 1.65 billion in 2025. The Chinese market leads global growth at 7.8% CAGR, with photovoltaic applications accounting for roughly 35–40% of global consumption volume.

(V) High-Purity Polysilicon (Electronic-Grade Polysilicon)

Polysilicon consists of multi-crystalline silicon material in which grains of various orientations are assembled; it is deposited via chemical vapour deposition at high temperatures. Solar-grade (SoG) polysilicon at 6N–8N purity is sufficient for photovoltaic cell efficiency; semiconductor-grade (electronic-grade) polysilicon requires purity of 9N–11N (99.9999999%–99.999999999%), meaning that for every 10 billion silicon atoms, the total count of other elements (B, P, As, Fe, Cr, Ni, Cu, Na, K, C, etc.) must not exceed one, with metallic impurities controlled to sub-0.01 ppb, and electroactive dopants B and P below 0.001 ppba.

Used for pulling semiconductor-grade single-crystal silicon ingots (CZ or FZ method), which are then sliced into semiconductor wafers — the initial solid silicon source for chip fabrication. The stringent purity requirements of electronic-grade polysilicon dictate purification processes, equipment investment, and certification cycles far exceeding those of the solar-grade track. The global semiconductor-grade polysilicon market was approximately USD 1.074 billion in 2026, forecast to reach USD 1.576 billion by 2035, CAGR ~4.4%. Global annual demand for semiconductor-grade polysilicon is approximately 33,500 tonnes (2025), just ~2.4% of total polysilicon demand — far smaller in scale than the solar-grade track.

(VI) Electronic-Grade Silicon Dioxide

High-purity SiO₂ exists primarily as ultrafine powder (fumed silica) or colloidal suspension (colloidal silica), and is the core abrasive in semiconductor CMP (chemical-mechanical planarisation) slurries. CMP is the critical global-planarisation step in chip fabrication, using differently formulated slurries for planarisation of tungsten plugs, copper interconnects, and interlayer dielectrics, with ultrapure SiO₂ as the most important inorganic abrasive component. High-purity SiO₂ is also used in chip-packaging filler (compounded with epoxy resin as EMC material) and in optical fibre preform fabrication.

(VII) High-Purity Quartz Sand (Fused Silica Sand)

With SiO₂ content above 99.9% and removal of aluminium, iron, calcium, manganese, and other impurities found in natural quartz, high-purity quartz sand is the key raw material for producing semiconductor-grade quartz crucibles (containers for pulling silicon ingots), quartz tubes (diffusion/oxidation tubes), and quartz boats (wafer carriers) — high-temperature-resistant components throughout chip fabrication — as well as optical-fibre preforms (the quartz glass core of optical fibres). Semiconductor quartz crucibles place extremely demanding requirements on quartz sand purity (SiO₂ ≥ 99.99%, total metallic impurities < 10 ppm). The global high-purity quartz sand market was approximately USD 1.38 billion in 2025, forecast to reach USD 2.74 billion by 2034, CAGR ~7.9%. China's market was approximately USD 199 million in 2024, forecast to reach USD 663 million by 2031, CAGR as high as 20.13%.

(VIII) Silicone Gel / Silicone Rubber

Elastomeric or gel materials formed by crosslinking polysiloxanes, including two-part addition-cure silicone gels and condensation-cure silicone rubbers. In electronics, silicone materials are widely used in power semiconductor module encapsulation and protection (IGBT, SiC MOSFET, etc.), LED chip packaging (primary-optical silicone gel), automotive ECU circuit-board conformal coating, and optical lens protective layers — owing to their excellent electrical insulation, thermal stability (−60 °C to +250 °C), low ionic content, and high optical transparency. Electronic-grade silicone must have extremely low ionic impurities such as Na⁺ and K⁺ (typically < 1 ppm) and must pass semiconductor-grade certification (e.g., MIL-SPEC or AEC-Q standards).

1.3 Clarifying the Boundary with Electronic Specialty Gases

Electronic specialty gases (such as silane SiH₄, trichlorosilane SiHCl₃, silicon tetrafluoride SiF₄, ammonia NH₃, and nitrous oxide N₂O) are gaseous substances handled in high-pressure cylinders or pipeline delivery systems at production, storage, and use sites. Their logistics, safety management, and pricing structures differ fundamentally from liquid chemicals and are covered in a separate thematic report published by this institute.

The eight categories covered in this report are all liquid or solid silicon-based chemicals. Their packaging (liquid drums, solid bags), logistics (liquid tankers or small-batch chemical-dedicated transport), production scale (measured in tonnes, not cylinders), and customer-certification logic differ fundamentally from gaseous specialty chemicals and should be regarded as an independent industrial segment.

Furthermore, silicon wafers themselves are an independent semiconductor-materials category positioned downstream of the "chemicals" segment discussed in this report, and are not within its scope.

1.4 Industry Scale Overview and Growth Drivers (2025–2026)

Drawing on multiple authoritative data sources, estimated global market sizes for all liquid/solid electronic-grade silicon-based chemical categories in 2025–2026 are as follows:

Category	2025–2026 Global Scale	Primary Growth Drivers
High-purity TEOS	~USD 270 million	Wafer fab CVD capacity expansion
Electronic-grade HMDS	~USD 190 million	Advanced lithography process iteration
D4 (electronics applications)	~USD 500–600 million	Semiconductor packaging / LED / automotive electronics
Silane coupling agents	~USD 1.65 billion	PV / lithium battery / optical fibre
Electronic-grade polysilicon	~USD 1.07 billion	AI compute wafer capacity expansion
High-purity quartz sand	~USD 1.38 billion	Semiconductor crucibles / optical fibre
Electronic-grade SiO₂ + silicone	~USD 800 million	CMP + packaging
Total	~USD 5.9 billion	Semiconductors + new energy dual drivers

Three tiers of forces drive growth in the above markets:

First tier — continuous global wafer capacity expansion: TSMC (Taiwan, USA, Japan, Germany), Samsung (South Korea, USA), Intel (USA, Germany), and SMIC/Hua Hong/CXMT/YMTC (mainland China) are all continuously investing in new fabs or capacity expansions. Each additional unit of wafer output directly drives absolute volume growth for each category of electronic chemical.

Second tier — technology node progression: As chip process nodes advance from 28 nm through 7 nm, 5 nm, 3 nm, and 2 nm, shrinking device dimensions require more dielectric deposition layers per unit of wafer area (interlayer dielectric stacks exceeding 10 layers). This means per-wafer consumption of precursors such as TEOS and HMDS does not decline — it increases. Simultaneously, each new node raises the purity specification versus its predecessor, expanding the premium commanded by high-purity products.

Third tier — incremental demand driven by new-energy applications: The rapid volume ramp of photovoltaic modules, lithium-ion batteries, and electric-vehicle electronic systems constitutes an important new source of incremental demand for silane coupling agents, electronic-grade silicone, and high-purity quartz sand, broadening the market boundary of silicon-based electronic chemicals from pure semiconductor to a wider "electronics + new energy" mega-segment.

1.5 Supply-Chain Security: Hidden Strategic Assets and China's "Chemical Chokepoint" Reality

Before surveying the full landscape of the electronic-grade silicon-based chemicals industry, one critically important backdrop must be understood: these chemicals are far more than ordinary industrial inputs. They are hidden strategic assets underpinning the digital infrastructure of modern civilisation — enabling materials for chip fabrication that cannot be bypassed. Without them, even the world's most advanced lithography machines cannot produce conforming chips.

Three levels of meaning as strategic assets

Level one: non-substitutability. In semiconductor manufacturing, every electronic chemical has an irreplaceable chemical role. TEOS is the only silicon source that, as a liquid precursor, can efficiently deposit high-quality SiO₂ at low temperature (<400 °C) in CVD processes; it has been the industry's de facto standard for ILD/STI/passivation applications in mainstream logic processes for decades, with virtually no seamless low-cost alternative. HMDS's hydrophobic surface modification function is equally irreplaceable in mainstream lithography workflows — switching to any other surface treatment agent would require re-qualifying the entire lithography process flow at prohibitive cost. High-purity polysilicon is the sole feedstock for growing semiconductor-grade single-crystal silicon ingots, and its purification pathway (STC-TCS cycling + zone-refining) is non-optional.

Level two: concealment and strategic deceptiveness. Compared with chip-manufacturing equipment (lithography machines, etch tools), electronic chemicals have an extremely low public profile. Yet their strategic importance to chip fabrication is no less than that of the equipment. A fab can spend billions of dollars on the world's most advanced lithography machines, but without electronic chemicals that meet specification, the machines cannot produce conforming chips. More importantly, managing controls on electronic chemicals is far harder than managing controls on equipment — equipment consists of large, visible, easily traceable physical assets, while high-purity chemicals are liquid bulk goods that can move through multiple trade channels and whose end use is difficult to trace once processed. This concealment makes electronic chemicals a potential "hidden strategic lever" in great-power competition.

Level three: continuous consumption. Chip-manufacturing equipment, once purchased, can be used for many years (typical depreciation period 10–15 years), whereas electronic chemicals are continuously consumed — every wafer produced requires a fixed volume of TEOS, HMDS, CMP abrasives, etc., and these cannot be stored long-term or recycled in the production cycle (some CMP waste liquid can be partially recovered, but new material must still be replenished). This means that managing the electronic chemicals supply chain can effectively constrain chip-manufacturing capability — without disrupting the hardware itself — simply by cutting off supply.

China's electronic-chemicals supply-chain exposure

Based on 2025–2026 actual data, China still has significant import dependence in key categories of electronic-grade silicon-based chemicals (reference values by value for import-dependence ratios):

Semiconductor-grade polysilicon (11N): import dependence ~50%, primarily from Japan (Tokuyama), Germany (Wacker), and the United States (Hemlock)
High-purity HMDS: import dependence ~70%, primarily from the USA (Entegris, Gelest), Germany (Merck/Versum), and Japan (Shin-Etsu, Tokyo Chemical Industry)
Advanced-node TEOS (qualified for 7 nm and below): import dependence ~95%, primarily from Japan (Evonik Japan), South Korea (Soulbrain), and the USA (Entegris)
High-purity quartz sand (top crucible grade): import dependence ~40%, primarily from the USA (Spruce Pine/Sibelco/Covia) and Norway (Elkem Quartz)

These figures reveal not only market opportunity (domestic-substitution space) but also a strategic risk map (likelihood and scope of potential supply disruption). Under the current geopolitical configuration, the US–Japan–Europe export-control coordination mechanism (the ECRA framework, Dutch export-licence management) is tightening progressively. Although direct controls on electronic chemicals are currently weaker than those on semiconductor equipment, the possibility of "adding high-purity HMDS and specialty TEOS precursors to control lists" is a medium-term risk scenario that China's industrial policymakers must take seriously.

Strategic reserves and multi-vendor sourcing strategies

To address the above supply-chain vulnerabilities, major chip manufacturers and chemical-procurement decision-makers have adopted the following countermeasures:

First, a "dual/triple sourcing" strategy: for critical chemicals, fabs typically require two to three independent suppliers on their qualified vendor lists (QVL/AVL) so that, if the primary supplier faces supply-disruption risk, they can switch rapidly to maintain production continuity.

Second, "strategic inventory" strategies: for certain high-risk chemicals (high import dependence, single-source supply, potential export-control exposure), fabs and chemical distributors have established strategic stockpiles equivalent to three to six months of consumption to absorb short-term supply disruptions.

Third, "accelerated domestic-substitution qualification" strategies: at the policy level, national capital including the Big Fund Phase III has entered the market and is accelerating the otherwise lengthy qualification process for domestic-substitute materials to enter fab qualification programmes through policy incentives such as subsidies and technical coordination.

Fourth, "geographic diversification" strategies: where possible, overseas suppliers in any single category should come from different countries, avoiding geographic concentration of risk (for example, not relying simultaneously on a single Japanese or American supplier, but introducing Korean suppliers to spread risk).

These countermeasures explain, from a supply-chain structure perspective, why fabs sometimes cooperate actively with domestic substitution: for the fab, adding a domestic supplier to the QVL is essentially building "supply-disruption insurance" into the supply chain — not merely a patriotic gesture or cost-saving measure.

Chapter 2　The Global Landscape and China's Position

2.1 A Tripolar Global Supply Structure

The current global supply of electronic-grade silicon-based chemicals has formed a tripolar competitive structure centred on Japan, Europe/USA, and China, but the hierarchy among the three poles is sharply unequal, and the relative strengths differ markedly across product categories.

Japan: the deepest technological moat

Japan's technological accumulation in electronic-grade silicon-based chemicals is the most historically rooted and moat-deepest of the three poles. Companies such as Shin-Etsu Chemical, Tokuyama, and Mitsubishi Materials have cultivated semiconductor-grade polysilicon, high-purity TEOS, and photolithography materials for decades, amassing irreproducible process know-how.

Shin-Etsu alone controls approximately 30% of the global high-purity silicon market and has recently announced an additional investment of approximately USD 700 million (~100 billion JPY) in Japan to expand high-purity silicon-materials capacity, explicitly targeting top-tier fabs operating at 3 nm and below — notably aligned with TSMC's ongoing expansion at its Kumamoto plant in Japan.

Tokuyama's Shunan plant is Japan's most important high-purity polysilicon production base, with annual capacity of 8,000–12,000 MT, essentially all electronic-grade (EG-grade), and qualified at all major fabs worldwide. In Q1 2024, Tokuyama completed an approximately JPY 30 billion expansion at this plant, dedicated to next-generation high-purity polysilicon purification units designed for sub-3 nm process nodes. In July 2025, Tokuyama further announced a joint venture with Korea's OCI to launch a 10,000-tonne/year polysilicon plant at the Samalaju Industrial Park in Sarawak, Malaysia, aiming to build a lower-cost offshore polysilicon supply-chain node in Southeast Asia to reduce dependence on Japan's high labour costs.

Europe/USA: strong chemical capability, polysilicon dominance

The core strengths of the Europe/USA pole are: first, polysilicon manufacturing (Wacker + Hemlock, together holding approximately 75% of global semiconductor-grade polysilicon market share); second, high-end precursors and specialty chemicals (Entegris, Merck/Versum, Air Liquide).

Germany's Wacker Chemie operates the Burghausen polysilicon plant, Europe's most important electronic-grade polysilicon production base, qualified at Intel, TSMC, Samsung, and other top customers, and is one of the most irreplaceable polysilicon sources at present. Wacker's 2025 revenue breakdown shows that the silicones business (including polysilicon and silane products) accounts for approximately 40–45% of total revenue; the high margins from semiconductor-grade polysilicon, protected by qualification barriers, are a key pillar of profitability.

US-based Hemlock Semiconductor (a joint venture between Dow Chemical and Shin-Etsu Chemical, headquartered in Michigan) is North America's largest polysilicon producer and one of the global suppliers with the longest qualification history and most entrenched customer relationships for semiconductor-grade polysilicon. "Hemlock Brand" polysilicon has a decades-long track record supplying Intel, IBM, and TSMC, creating extremely durable customer stickiness.

Entegris (USA) is the world's leading integrated semiconductor-materials solutions provider, with TEOS, HMDS, high-purity solvents, and filtration systems among its globally leading product lines. Its 2022 acquisition of CMC Materials further strengthened its position in CMP slurries and polishing pads.

Merck KGaA's (Germany) semiconductor-materials business (originally Versum Materials, acquired for USD 5.7 billion in 2019) covers HMDS, low-k dielectric precursors (DEMS, etc.), etchants, developers, and other categories, and holds core slots in TSMC's and Intel's advanced-node qualified-supplier lists.

Air Liquide (France), originally a specialty gases company, has through deep expansion into ALD/CVD precursors (its Air Liquide Advanced Materials division) become one of the world's most important CVD/ALD liquid-precursor suppliers. Its Chinese joint venture (Suzhou Jinhong Gas, a JV with Air Liquide) has the capability to supply ultra-high-purity ammonia and TEOS products in the Chinese market.

China: scale leader, high-end laggard

China holds an overwhelming advantage on the "scale" dimension — whether in organosilicon monomers (Hoshine, Xinghuo, Dongyue, etc.), solar-grade polysilicon (Tongwei, GCL, Daqo, Xinte), or silane coupling agents (Momentive, Liansuo, etc.), China is the world's largest producer and consumer.

However, "scale leadership" does not equate to "high-end leadership." In the electronic-grade (semiconductor-grade) segment, Chinese suppliers broadly face the dual constraints of purity-certification thresholds and customer-qualification cycles. In the highest-purity product specifications (11N electronic-grade polysilicon, advanced-node TEOS, semiconductor-grade HMDS), a perceptible technical and market-position gap relative to the world's top-tier suppliers remains.

In concrete figures: Chinese suppliers hold more than 95% market share in industrial/LCD-grade TEOS, but only approximately 40–60% in semiconductor-grade TEOS (mostly at mature nodes; advanced-node share is lower still). In solar-grade polysilicon their share exceeds 85%, but in electronic-grade polysilicon (11N) it is approximately 15–25%. In standard silane coupling agents their share exceeds 85%, but in specialty high-end grades it is approximately 50–60%.

2.2 China's True Position in the Global Supply Chain: Breaking Through the Paradox

A seemingly contradictory yet precisely accurate portrait of China's silicon industry today:

China controls more than 90% of global polysilicon capacity, yet approximately 50% of its electronic-grade polysilicon demand must be met by imports.

The root of this contradiction is that "polysilicon" is not a homogeneous product. Between solar-grade (SoG, 6N–7N) and electronic-grade (EG, 9N–11N) there exists a fundamental technical chasm.

Key parameters for solar-grade polysilicon: B ≤ 0.3 ppba, P ≤ 0.3 ppba, total metals ≤ 1 ppbw, carbon ≤ 0.5 ppma. Meeting these parameters is achievable with the standard modified Siemens process and basic fractional distillation purification. China began to break through on this technology track in the early 2010s and, through cost optimisation and scale expansion, achieved a historic transformation from import dependence to absolute global dominance in the solar-grade polysilicon market.

Electronic-grade polysilicon (11N), however, requires: B ≤ 0.001 ppba, P ≤ 0.001 ppba, total metallic ions ≤ 0.01 ppbw, carbon ≤ 0.1 ppma. These specifications are two to three full orders of magnitude more stringent than solar-grade. Achieving this purity requires, on top of the modified Siemens process, additional high-temperature rectification stages (precision removal of electroactive dopants B and P), float-zone refining (FZR) post-processing, and far higher requirements for process stability, equipment precision, and cleanroom management throughout the entire chain.

A structurally analogous paradox exists in the TEOS space: industrial-grade TEOS (metallic impurities in the ppm range) has long been produced at large scale in China; but semiconductor-grade TEOS (metallic impurities controlled to ppb–ppt) remains in the ramp-up phase domestically. Yoke Technology has made the deepest domestic breakthrough by acquiring Korea's UP Chemical to bring in core process capability, but imports still dominate in the most advanced node TEOS qualifications (7 nm and below).

2.3 China's Market Strategic Importance and Rapid Rise

Despite the gap in high-end categories, China's strategic importance in the global electronic-grade silicon-based chemicals map is rising rapidly:

Market scale: China's key electronic-materials market was estimated at approximately RMB 174.08 billion in 2025 (including semiconductor materials and new-energy chemicals), up approximately 21.1% year-on-year, one of the world's fastest-growing regional markets.

Downstream demand pull: China's wafer-foundry capacity is continuously expanding (SMIC, Hua Hong Semiconductor, CXMT, YMTC, etc.); new energy vehicle production and sales have exceeded 10 million units per year; and annual new solar installations exceed 250 GW (domestically alone), forming a massive domestic demand base for electronic-grade silicon-based chemicals.

Policy support: The Big Fund Phase III (registered capital RMB 344 billion) explicitly lists semiconductor materials (including electronic chemicals) as a priority support direction; policy-level capital injections are accelerating technology breakthroughs and qualification progress.

Urgency of domestic substitution: Escalating US-and-ally export controls have significantly raised wafer fabs' willingness to source critical chemicals domestically. Even where quality gaps exist, the strategic premium of "supply security" is driving fabs to more actively qualify domestic chemical suppliers.

2.4 Geopolitics Reshaping the Global Supply Chain

Since 2022, the USA, Japan, the Netherlands, the UK, and other countries have progressively widened semiconductor-related export-control scope — from chip-design software (EDA) to advanced manufacturing equipment (lithography tools, ALD equipment, etc.) to certain high-end materials, with the control perimeter continuously expanding. Although most electronic chemicals have not yet been directly listed, the threat has profoundly affected Chinese fabs' procurement strategies, accelerating adoption of "dual-vendor" (domestic + import qualification in parallel) or even "domestic-only" secure-sourcing strategies.

This geopolitical pressure represents a historic opportunity for China's electronic-chemicals companies: even where product performance in some categories has not yet fully matched imports, downstream fabs are inclined to accept a degree of quality tolerance, providing domestic suppliers with market space to break through qualification at mature process nodes first, accumulating process data and qualification experience that will build the technical readiness to move up to more advanced nodes.

From a global supply-chain perspective, the geopolitics-driven supply-chain bifurcation is producing signs of "parallel markets" in the global electronic-chemicals sector: a "Western semiconductor-standard supply chain" centred on TSMC/Samsung/Intel, and a "Chinese semiconductor procurement ecosystem" centred on SMIC/Hua Hong/CXMT/YMTC, are gradually forming two not-fully-overlapping vendor qualification systems for chemical suppliers. This bifurcation trend means domestic electronic-chemicals companies have the opportunity to build an independently credible quality track record in the domestic market and gradually form a sustainable localised supply ecosystem.

2.5 Intra-Asia Competition and Cooperation: China-Japan-Korea-Taiwan Dynamics

Even within Asia, the industrial-chain interplay among Japan, South Korea, Taiwan, and mainland China is continuously deepening.

Japanese companies (Shin-Etsu / Tokuyama / JSR / Sumitomo Chemical) maintain an extremely deep technical moat in semiconductor-grade polysilicon and advanced photolithography materials — the most difficult supply sources for China to rapidly replace. Korean companies (SK Materials, DuPont Korea, Soulbrain) compete directly with Chinese companies in TEOS and certain precursors, and the Korean government, motivated by its own semiconductor industry protection, has a relatively open posture toward technology transfer. Taiwan is not a major chemical producer, but TSMC's quality-certification standards effectively set the global highest-specification benchmark for electronic chemicals — materials that pass TSMC qualification carry the world's highest supply credentials. This "TSMC qualification effect" is something domestic Chinese enterprises should deeply understand.

Further unpacking the roles of these four Asian economies:

Japan, as the historical birthplace of high-purity chemical technology, possesses tacit-knowledge barriers formed over decades of process accumulation in organic synthesis and ultrapure chemical purification — the hardest-to-imitate competitive advantage. Japan began extending export controls to advanced semiconductor manufacturing equipment in April 2023 and expanded further into certain chemical precursors in 2024, signalling a strategic tightening trend. Yet Japanese chemical companies have extensive local production in China (Shin-Etsu, JSR, Sumitomo Chemical all have Chinese production bases), and this deep commercial coupling forces the Japanese government to maintain relative restraint in the intensity of controls.

South Korea is one of mainland China's most important sources of technology partnerships in electronic chemicals in the Asia-Pacific region; Yoke Technology's acquisition of UP Chemical (South Korea) is the most successful example to date. Korean specialty-chemical companies (SK Materials, Soulbrain, Oci Materials, etc.) have maintained a degree of openness to transactions with Chinese capital, though as the China-US tech rivalry deepens and Korean government security considerations increase, this window is narrowing. Another distinctive feature of Korean companies is that many are already deeply involved in supplying materials to mainland Chinese fabs (SMIC, SK Hynix's China plants, etc.) and therefore have strong business incentive to maintain open engagement with the Chinese market.

Taiwan's role is "standard-setter" rather than "producer": TSMC, with the world's most advanced process nodes and most stringent quality requirements, objectively sets the highest-specification quality benchmark for electronic chemicals globally. TSMC's procurement team's technical specifications and qualification processes, formed in the course of evaluating suppliers worldwide, are typically treated by the industry as the "highest industry standard." For mainland Chinese chemical companies seeking to expand into the global fab market, TSMC qualification is one of the ultimate objectives — though enormous political obstacles exist in the current environment.

Mainland China is forming a relatively independent electronic-chemicals quality-certification ecosystem: a qualification system centred on SMIC, CXMT, and YMTC that, in certain categories, still trails the global top level (TSMC standards) — but as Chinese fab process nodes continuously evolve (SMIC 14 nm, N+1/N+2; YMTC 192-layer NAND, etc.), the requirements placed on supporting materials are constantly upgrading toward international standards, and this process is itself the core driver of continuous quality improvement for domestic electronic chemicals.

2.6 Evolution of Global Capacity Landscape: The 2025–2030 Map Redrawing

Supply-chain pressure from TSMC's geographic expansion

TSMC has implemented the largest geographic diversification strategy in its history over the past five years: Arizona, USA (N4/N2 fabs, in mass production from 2025, N2 expansion to 2028); Kumamoto, Japan (N28/N16, in mass production from 2024, Kumamoto Phase II N6 under construction); Dresden, Germany (specialty processes for automotive/industrial, production expected from 2027). This multi-site strategy places new pressure on the geographic layout of the supporting electronic-chemicals supply chain.

The electronic-chemicals supply network, previously concentrated in Taiwan and South Korea, must now simultaneously build supply capabilities near Arizona, Kumamoto, and Dresden. This is an opportunity for Japanese companies (rapidly positioning around TSMC's Japanese plants) and European companies (attempting to build materials supply around Dresden); for Chinese companies, it represents a "participation threshold": if domestic Chinese electronic-chemicals companies cannot secure a slot in TSMC's global supplier directory, they will find it difficult to share in the incremental market accompanying TSMC's international expansion.

Samsung + SK Hynix: the in-house sourcing trend

Both Samsung Electronics and SK Hynix are actively pursuing partial internalisation of electronic chemicals: Samsung through Samsung SDI and Cheil Industries (merged into Samsung SDI) in certain material directions; SK Hynix through strategic partnerships with SK Materials and POSCO Chemical for stable supply of polysilicon and precursor materials.

This internalisation trend is a potential market-compression signal for external chemical suppliers (including Chinese companies): if Samsung/SK Hynix achieve a high degree of internalisation in a given category (e.g., TEOS or certain ALD precursors), the market-share space available to external suppliers will shrink. But historically, complete internalisation has limitations in terms of technology diversity, cost, and process flexibility; the internalisation ratio in chemicals typically does not exceed 30–40%, leaving substantial space for external suppliers.

The pull from new semiconductor manufacturing nodes in Southeast Asia The construction of new semiconductor manufacturing nodes in emerging Southeast Asian and South Asian locations — TATA Electronics' fab in Gujarat (India), Intel's packaging expansion in Vietnam, Intel's Penang capacity increase and GlobalFoundries' planned new plant in Sg. Penchala (Malaysia) — will generate new regional demand for electronic chemicals in 2027–2030. Compared with the high-barrier markets of Taiwan and South Korea, emerging Southeast Asian fabs are relatively open in supplier selection (many operate mature process nodes with comparatively lower purity requirements), and could become the "first port of call" for Chinese electronic-chemicals companies in overseas markets — accumulating international supply experience and building overseas chemical-supply credentials in lower-barrier Southeast Asian markets before the certification barriers in core European, American, and Japanese markets have been broken.

Chapter 3　Core Technology: Precursor Synthesis, Purification, and ppb-Level Trace Metal Control

3.1 TEOS Synthesis Routes and Purification Technology

Industrial synthesis of TEOS follows two mainstream routes:

Route 1: Silicon Tetrachloride (SiCl₄) Alcoholysis (mainstream)

SiCl₄ + 4 C₂H₅OH → Si(OC₂H₅)₄ + 4 HCl

Silicon tetrachloride reacts with anhydrous ethanol under an inert atmosphere (N₂ protection) at low temperature (−10 °C to +20 °C) to yield TEOS and hydrogen chloride (HCl) as a by-product. Advantages: fast reaction rate (SiCl₄ is highly reactive with alcohols), good TEOS selectivity, relatively simple work-up. Disadvantages: SiCl₄ itself is a hazardous chemical (reacts violently and exothermically with water), and the large volume of HCl gas produced requires comprehensive tail-gas absorption and neutralisation systems. SiCl₄ is a by-product of polysilicon production (modified Siemens process) and organosilane monomer production; it is in abundant supply in China, making this the primary process route for both industrial-grade and electronic-grade TEOS.

Route 2: Sodium Silicate (Na₂SiO₃) Alcoholysis (industrial-grade use)

Na₂SiO₃ + 4 C₂H₅OH + H₂SO₄ → Si(OC₂H₅)₄ + Na₂SO₄ + 3H₂O

This route uses low-cost feedstocks (sodium silicate + ethanol + sulphuric acid) and is suited to large-scale industrial-grade production. However, because sodium silicate feedstock itself contains relatively high concentrations of Na, K, Ca, Fe, and other metallic impurities that enter the TEOS product during esterification to varying degrees, stringent post-processing purification is needed to reach electronic-grade specifications. This route is therefore less commonly used for electronic-grade products.

Electronic-Grade TEOS Purification System

Regardless of the synthesis route, the core technical challenge in preparing electronic-grade TEOS lies in constructing the purification stage, which includes the following key elements:

Rectification and purification system: TEOS (bp 168.8 °C) differs sufficiently in boiling point from common impurities (ethanol bp 78.4 °C, triethyl silicate bp 119.5 °C, mixed methyl–ethyl orthosilicates, etc.) that effective separation by rectification is achievable. Electronic-grade TEOS typically requires a rectification column with 60–100 theoretical plates, using ultrapure metal structured packing inside a packed column, operated entirely under an inert atmosphere (high-purity N₂) to prevent moisture and oxygen from causing TEOS hydrolysis or metal corrosion. Reflux ratio and operating pressure must both be precisely controlled to ensure the collected product cut meets specification.

Trace metallic ion removal: Rectification cannot effectively remove metallic organic-form impurities that are azeotropic with TEOS or have similar chemical properties. Dedicated adsorption-removal units must therefore be installed before and after rectification. Common methods include: (1) specialty ion-exchange resins (must be ultra-pure-grade electronic resins with extremely low metallic leaching); (2) activated-carbon (ultra-pure activated carbon) adsorption — removes certain organic impurities but is relatively non-selective for metallic ions; (3) extractive separation (selective extraction using metal-complexing agents) — suitable for precise control of specific metallic ions. For advanced-process TEOS requiring sub-ppb metallic impurities, multiple combined removal steps are typically required.

Moisture-control system: TEOS hydrolyses and condenses even in the presence of trace moisture at ppm levels, generating SiO₂ particles (silica sol) — one of the primary sources of particulate contamination in the finished product. The entire production process must therefore be carried out in an extremely dry environment: the dewpoint inside the production building must be below −60 °C (equivalent to an absolute humidity of approximately 10 µg/m³); raw-material ethanol must be deeply dehydrated in advance (anhydrous ethanol moisture content < 50 ppm); all piping, vessels, and valves must be made of electropolished 316L stainless steel or PTFE throughout — plastic materials are prohibited (they can leach plasticisers and metal molecules).

Ultra-fine particulate filtration: The final product must be filtered through membranes with pore sizes of 0.1 µm (some advanced-process TEOS requires 0.05 µm) to reduce residual SiO₂ particles and foreign particles to semiconductor cleanliness levels (typically ≥0.1 µm particles < 50 per mL; advanced-process specifications are even more stringent). All filters must be made from ultra-pure PTFE/PVDF and must pass an integrity test before each use.

Packaging and cleanroom management: Qualified electronic-grade TEOS must be filled in a Class 10 (ISO Class 4) or higher cleanroom using dedicated containers compliant with the SEMI C8 standard (PTFE-lined), sealed under nitrogen protection, with complete production batch records maintained for customer inquiry and traceability.

3.2 HMDS Preparation Process and Quality Control

Preparation process

Industrial synthesis of HMDS primarily uses the trimethylchlorosilane ammonolysis route:

2 (CH₃)₃SiCl + NH₃ → (CH₃)₃Si-NH-Si(CH₃)₃ + 2 NH₄Cl

Trimethylchlorosilane (TMS-Cl) reacts with ammonia at low temperature (−20 °C to +5 °C) in an organic solvent (typically a non-polar solvent such as heptane or pentane) to yield HMDS and ammonium chloride (NH₄Cl) as a by-product. NH₄Cl is a solid and is removed by filtration; the organic solvent is recovered by atmospheric distillation; crude HMDS is then purified by rectification to give the high-purity product.

Key technical difficulties of this process route include:

The trimethylchlorosilane (TMS-Cl) feedstock is extremely moisture-sensitive — exposure to humid air causes immediate hydrolysis — so feedstock storage and handling must be strictly anhydrous
The HMDS product itself is similarly highly hygroscopic and decomposes on contact with moisture (generating NH₃ gas); the entire production process (reaction, filtration, rectification, filling) must therefore be carried out under strictly anhydrous conditions (dry N₂ protection)
In the rectification step, the separation of HMDS (bp 125–127 °C) from the solvent (heptane bp 98.4 °C) and possible by-products must be precisely controlled to obtain high-purity product

Electronic-grade HMDS quality specification (typical values)

Parameter	Unit	Specification
Assay (GC)	%	≥99.99
Moisture	ppm	≤10
Fe	ppb	≤0.1
Na	ppb	≤0.1
K	ppb	≤0.1
Cr	ppb	≤0.1
Ni	ppb	≤0.1
Cu	ppb	≤0.1
Colour (APHA)	—	≤5
Particles (≥0.2 µm)	per mL	≤50

HMDS application control in wafer fabrication

HMDS treatment is a standard step in the pre-lithography processing module (TRACK system), with two typical application methods:

Vapour-phase HMDS (VP-HMDS): Liquid HMDS is heated and vaporised; the HMDS vapour treats wafers that have already undergone dehydration baking in a sealed chamber (the HMDS treatment chamber), at treatment temperatures of 90–120 °C for 30–90 seconds. Vapour-phase HMDS distribution is uniform and silanisation efficiency is high; this is the mainstream method in 300 mm wafer fabrication. Purity requirements for HMDS are extremely high — any trace of organic sulphur, organic chlorine, organic amine, and similar impurities can undergo undesirable chemical reactions with the photoresist, causing a rise in lithography defect density.

Spin-on HMDS: Liquid HMDS is applied directly to the wafer surface on a spin-coating stage, suited to certain 200 mm or smaller process lines or low-to-mid-end processes. This method has lower cost but higher HMDS consumption per wafer, and surface treatment uniformity is slightly lower than the vapour-phase method.

One technology-evolution trend in the photolithography industry is that, with the spread of EUV lithography and the introduction of metal-oxide photoresists (MOP), the conventional HMDS treatment process associated with chemically amplified photoresists (CAR) is undergoing adjustment. In some EUV processes, HMDS treatment has been replaced by dedicated bottom anti-reflective coatings (BARC) or chip-adhesion-promoting coatings. However, this transition is occurring mainly at leading-edge advanced nodes (below 7 nm); at mature nodes (28 nm and above), HMDS remains a highly standardised, indispensable process material with no near-term substitute.

3.3 High-Purity Polysilicon Purification Route and Process Breakthroughs

The Modified Siemens Process (MSP) is the current mainstream process for semiconductor-grade polysilicon production worldwide and can be broken down into five main steps:

Step 1: Metallurgical silicon and trichlorosilane synthesis Metallurgical silicon (98–99% purity) powder reacts with anhydrous hydrogen chloride (HCl) gas in a fluid-bed reactor at approximately 300 °C to produce trichlorosilane (TCS, SiHCl₃) along with small amounts of silicon tetrachloride (SiCl₄) and dichlorosilane (DCS, SiH₂Cl₂): Si + 3 HCl → SiHCl₃ + H₂

Step 2: Multi-stage rectification of trichlorosilane The raw TCS feedstock contains various metal chlorides (BCl₃, PCl₃, AsCl₃, FeCl₃, etc.) and organosilane impurities that must be removed by multi-stage rectification columns. In particular, BH₃·SiHCl₃ (boron-trichlorosilane complex) has a boiling point extremely close to TCS (only approximately 0.7 °C difference), making ordinary rectification ineffective; specialist zeolite molecular sieves or extractive distillation methods are required to assist separation — this is one of the most technically challenging boron-removal hurdles.

Step 3: Bell-jar CVD reduction deposition High-purity TCS and high-purity hydrogen (H₂) are mixed in a fixed ratio and fed into a silicon deposition furnace (bell-jar reactor). The reactor is pre-loaded with a number of clean slim silicon rods (~8 mm diameter) as deposition substrates; the rods are resistively heated to approximately 1,100 °C. TCS is reduced by H₂ at high temperature, depositing and growing a high-purity polysilicon layer on the rod surface: SiHCl₃ + H₂ → Si + 3 HCl

As deposition time extends (typically for several days), the slim rods gradually grow into polysilicon rods of 100–150 mm diameter. By-products (SiCl₄, HCl, etc.) are treated in the off-gas system and recycled, greatly reducing feedstock consumption.

Step 4: Polysilicon rod cutting and crushing After deposition is complete, the polysilicon rods are removed from the furnace and mechanically cut and crushed in a clean environment to produce chunk polysilicon of various specifications — the primary commercial form of polysilicon.

Step 5: Float-zone refining (FZR — for electronic grade only) For semiconductor-grade polysilicon requiring ultra-high purity of 11N, the 9N–10N polysilicon rods obtained from the MSP process undergo float-zone refining (FZR): a narrow molten zone is produced on the polysilicon rod and slowly traversed from one end to the other. Since most impurities (B, P, etc.) have a higher solubility in the liquid phase than the solid phase (segregation coefficient k < 1), they migrate with the molten zone and concentrate at one end of the rod, leaving the solidified section at the other end at extremely high purity. Repeated zone-refining cycles progressively raise the purity from the initial 9–10N to 11N and above.

Granular silicon (FBR, Fluidized Bed Reactor) route

Distinct from the modified Siemens process, the fluidised-bed-reactor (FBR) granular-silicon technology uses silane (SiH₄) as the silicon source; silicon is deposited on small seed particles in a fluidised bed, producing spherical granular polysilicon with a diameter of 1–3 mm: SiH₄ → Si + 2 H₂

Granular silicon advantages over Siemens rod silicon: (1) energy consumption approximately 1/3 that of the Siemens process (no large-current resistively heated rods); (2) continuous production (no need to open the reactor between batches); (3) uniform particle size, ready to use (no crushing required), reducing particulate contamination steps. GCL Technology is the world's largest FBR granular-silicon supplier; its granular-silicon market share rose from approximately 5% in 2022 to approximately 12% in 2025, with production costs falling to RMB 28.17/kg (Q4 2024 data), demonstrating strong cost competitiveness.

However, granular silicon faces unique challenges in electronic-grade applications: in FBR, the contact area between silicon particles and reactor walls/gas distributors is larger, and if equipment material or process control is not managed precisely, the risk of metallic contamination is relatively higher. The high surface area of small particles also means higher surface-adsorbed impurity density. These factors mean that granular silicon's progress in electronic-grade qualification lags, and commercial electronic-grade application of granular silicon remains in the validation and exploration phase.

3.4 Trace Metal Control — The Cleanroom Philosophy Running Through the Entire Chain

In the manufacture of electronic-grade silicon-based chemicals, "trace metal control" is the most central, most precise, and most non-negotiable systemic challenge across all engineering activities.

The semiconductor-damage mechanism of metallic contamination

In wafer fabrication, even ppb-level transition metals (Fe, Cu, Cr, Ni, etc.) entering the SiO₂ dielectric layers or the silicon bulk of integrated circuits produce serious consequences: transition metals introduce deep-level trap states in the silicon bandgap, which capture minority carriers (electrons/holes) and reduce minority-carrier lifetime from the millisecond range to microseconds or even nanoseconds. This leads to a dramatic increase in leakage current, degradation of device off-state characteristics, shorter data-retention times in memory devices (reduced DRAM/Flash data retention), and ultimately a sharp decline in chip yield.

Copper (Cu) in particular is uniquely dangerous: its diffusion coefficient in silicon is extremely large (at room temperature, ~10⁻⁵ cm²/s in Si, one to two orders of magnitude faster than most metals), meaning copper contamination continues to diffuse during device use. This is why, after the industry transitioned from aluminium to copper interconnects (beginning in the late 1990s), the entire fab's copper-contamination management system had to be rebuilt from the ground up.

At advanced nodes below 7 nm, where physical device dimensions have shrunk (equivalent oxide thickness < 1 nm), the impact of ppb-level metallic impurities on device performance is amplified many times over. Some advanced-process chemical-purity requirements have been raised from ppb (parts per billion) to ppt (parts per trillion).

Full-chain trace metal control system

Achieving ppb or even ppt trace metal control cannot be solved by a single post-processing purification step; instead, the entire clean-chain system must be built systematically across six dimensions:

Dimension 1: Upstream feedstock purification. All chemical feedstocks involved in the reaction (HCl, H₂, ethanol, NH₃, etc.) must themselves satisfy — or be below — the metallic specifications required of the final product. If feedstock-borne metallic impurities are introduced, they will be nearly impossible to fully remove in subsequent purification steps. This requires establishing a complete supplier qualification and incoming-material inspection system; some ultra-high-purity feedstocks (such as ultra-pure HCl and electronic-grade ethanol) themselves require specialist qualification.

Dimension 2: Equipment material management. All equipment components in direct contact with process materials (reactors, rectification columns, heat exchangers, pumps, valves, piping) must be made of low-metal-leaching materials: preferably electropolished 316L ultra-pure stainless steel (Ra < 0.4 µm), or perfluoropolymer (PTFE/PFA/PVDF) lined. Ordinary carbon steel, cast iron, brass, or alloys with high Cu/Ni/Pb content are strictly prohibited. Piping connections must use welding (TIG welding, electropolished internal surface) rather than threaded connections (which accumulate deposits), and seals must be PTFE rather than rubber (which leaches plasticisers and metallic ions).

Dimension 3: Production-environment cleanliness control. The entire production, analysis, and filling area must be built and operated to ISO Class 4 (Class 10) through ISO Class 6 (Class 1,000) cleanroom standards, with strict filtration of airborne particulates (especially mineral dust containing metallic components), a comprehensive cleanroom management system including HEPA/ULPA filtration, positive-pressure isolation, and personnel gowning/cleanroom entry protocols.

Dimension 4: High-precision in-line analytical capability. Quality assurance for trace metal control relies on a complete analytical capability system. Electronic-chemicals companies must either operate or commission laboratories equipped with: inductively coupled plasma mass spectrometry (ICP-MS, detection limit down to 0.001 ppb = 10 ppt); gas chromatography–mass spectrometry (GC-MS, for organic impurity detection); Fourier-transform infrared spectroscopy (FTIR, for moisture and functional group detection); laser particle counters (LPC, for particle count and size distribution); and trace moisture analysers (Karl Fischer titration or electrolytic method, detection limit below 1 ppm). These instruments are worth tens of millions of RMB each and represent major capital investments; they are also among the most difficult barriers for smaller domestic companies to clear.

Dimension 5: Packaging and logistics cleanchain management. Finished electronic-grade liquid-chemical products must be packaged in dedicated ultra-clean containers specified by SEMI C8: outer container of HDPE (high-density polyethylene), inner liner of PTFE or PFA, preventing metallic leaching and organic dissolution. Containers must undergo rigorous pre-use cleanroom preparation (acid wash → ultrapure-water rinse → drying → cleanroom seal) to ensure the container itself introduces no new contamination. Storage and transport must maintain low temperature (typically 2–15 °C) and light exclusion, with nitrogen over-pressure to prevent moisture and oxygen ingress.

Dimension 6: Batch traceability and customer audit system. Each batch of electronic-grade chemicals must carry a complete production batch/lot number, recording production date, feedstock source and lot, process parameters, and a final Certificate of Analysis (CoA). Customers (wafer fabs) have the right to conduct on-site audits of suppliers, inspecting cleanroom conditions, equipment status, and quality management systems (QMS) — a mandatory threshold for suppliers to enter the fab's qualified vendor list (QVL).

3.5 Silane Coupling Agent Functional Synthesis Technology

The synthesis of silane coupling agents centres on hydrosilylation addition:

HSiCl₃ (trichlorosilane) or H₂SiCl₂ + CH₂=CH-R-Y → Cl₃Si-R-Y (chlorinated intermediate)

The resulting chlorinated intermediate is then alcoholysed (reacted with methanol/ethanol) to convert Si-Cl bonds to Si-OR bonds, giving the final trialkoxysilane coupling agent product:

Cl₃Si-R-Y + 3 CH₃OH → (CH₃O)₃Si-R-Y + 3 HCl

Key technical challenges of this process include:

Selectivity control in hydrosilylation addition: The reaction must be carried out in the presence of platinum metal catalysts (Speier catalyst or Karstedt catalyst), but the addition of the olefin to TCS can sometimes compete between Markovnikov and anti-Markovnikov orientation, and inadequate selectivity control will generate isomeric by-products that impair product functionality.

Synthesis difficulty for high-end functional groups: Silane coupling agents containing special functional groups such as azido (-N₃-), thiol (-SH), or fluorinated (-CF₃, etc.) groups require difficult-to-source feedstocks and complex synthesis routes; domestic companies in these high-end specialty grades still lag behind international counterparts (Momentive, Dow Corning, etc.).

High-purity refinement: High-purity silane coupling agents for electronics and optical applications must be purified by rectification (for low-boiling grades) or molecular distillation (for high-boiling grades) to remove organic impurities and moisture, ensuring product purity ≥ 99.5% and avoiding organic impurities that would impair interfacial modification effectiveness.

3.6 ALD Precursors: The Chemical Frontier of Next-Generation Silicon Thin-Film Technology

Atomic layer deposition (ALD) has emerged after CVD as a core thin-film deposition technology in advanced semiconductor processes, and has become an indispensable key process for nodes below 7 nm. The fundamental difference between ALD and conventional CVD: CVD is a continuous chemical vapour-phase reaction; ALD is a "stepwise, self-limiting" surface reaction — each reaction cycle consists of two steps: Step 1 "precursor A pulse": precursor A gas enters the chamber and saturatingly adsorbs as a monolayer on the active surface groups, with excess A purged away; Step 2 "precursor B pulse (or plasma activation)": precursor B enters the chamber and reacts with the adsorbed A layer to form a single-atomic-layer-thick film; these two steps are then repeated, with each cycle depositing approximately 0.1–0.2 nm of film (growth per cycle, GPC).

This "stepwise self-limiting" growth mode gives ALD unique capability: uniform conformal deposition on any three-dimensional surface geometry with precise thickness control — the fundamental reason why ALD cannot be replaced by CVD in advanced processes involving extreme-high-aspect-ratio 3D structures (FinFET 3D fin structures, DRAM deep-trench capacitors, 3D NAND gate stacks).

Major silicon-containing ALD precursor types and applications

Among silicon-containing ALD precursors, the following are technologically most advanced and highest in market value:

TEOS-derived ALD precursors (silicon oxide, SiO₂ ALD): While TEOS itself is not suitable for ALD (reaction temperature too high for self-limiting reaction), organoamino silicon-based precursors (such as TDMAS — tetrakis(dimethylamino)silane, and aminosilane derivatives of TEOS) have been applied in SiO₂ ALD for advanced processes. These products can achieve self-limiting ALD reactions at lower temperatures (50–300 °C), suitable for ultra-thin SiO₂ deposition on structures containing copper interconnect layers (temperature-sensitive).

Silicon nitride (Si₃N₄) ALD precursors: Silicon-amino-containing precursors (SiH₂Cl₂ dichlorosilane + NH₃ plasma, or BTBAS — bis(tert-butylamino)silane) for ALD silicon nitride films, widely used in the ONO (oxide-nitride-oxide) stacked charge-trap structures of 3D NAND devices, and as FinFET spacer films — key process materials for 3D NAND evolution toward 200+ layer stacks.

High-k dielectric ALD precursors (metal oxides, non-pure silicon): For advanced-node gate high-k dielectric layers (HfO₂, ZrO₂, Al₂O₃, etc.), ALD metal-oxide precursors (Hf(NMe₂)₄, TEMAH, TMA, etc.) react with O₃ or H₂O to form high-dielectric-constant (high-k) films. Although the primary elements in these precursors are not silicon, they are the highest-value-added, highest-barrier products in the ALD precursor segment — precisely the category Yoke Technology entered after acquiring UP Chemical.

China's ALD precursor market opportunity and domestic progress

ALD precursors are among the categories with the lowest domestic-substitution rates in the electronic chemicals space: the global ALD precursor market is currently dominated by the USA (Air Liquide Advanced Materials, Entegris/Atmi, Sigma-Aldrich, etc.), Japan (Up Chemical Japan, Showa Denko), and South Korea (SK Materials, OCI Materials), with Chinese companies nearly absent — Yoke Technology's acquisition of UP Chemical is the only significant domestic breakthrough.

The market outlook for ALD precursors is extremely attractive: as 3D NAND evolves toward 300+ layers (each 64-layer increase adds at least one ALD-ONO cycle), DRAM approaches the 20 nm physical limit, and logic devices advance to 3 nm/2 nm, global ALD precursor demand is projected to grow at approximately 15–20% per year in 2025–2030, with the market potentially expanding from approximately USD 1.5 billion in 2025 to USD 3.5–4.0 billion by 2030.

Domestic companies able to enter this segment need not only fine-chemical synthesis capability but also ALD process engineers (capable of participating in customer-line process development and qualification), plus specialised gas/liquid purification and analytical equipment investments of tens of millions of RMB. The barriers are so high that this segment will remain heavily import-dependent for the next three to five years — and precisely for this reason, those who succeed in entering will enjoy substantial technology premiums and customer stickiness.

3.7 Anatomy of Wafer-Fab Qualification: The Complete Path from "Sample Delivery" to "Volume Supply"

Understanding the deeper reasons behind slow domestic-substitution progress in electronic chemicals requires a thorough understanding of wafer fabs' supplier qualification process. This process is long and stringent and is the single most important factor determining the pace of domestic substitution — more so than the technology R&D itself.

The five phases of the qualification process

Phase 1: Supplier pre-qualification (~1–3 months). The fab's procurement and materials engineering department first conducts a "paper review" of potential suppliers: checking ISO 9001/14001 certification status, factory cleanroom classification (must generally reach ISO Class 5/Class 3), production-line GMP (good manufacturing practice) compliance, financial soundness (to guard against sudden supplier shutdown causing fab supply disruption), and statements of conformance to relevant SEMI standards such as SEMI S8 (safety) and SEMI C1. Only suppliers passing pre-qualification enter the substantive materials-evaluation phase.

Phase 2: Initial chemical analysis (~2–4 months). The fab's materials laboratory systematically analyses supplier samples: ICP-MS (trace metals in full elements — must meet SEMI C1 Grade 4/5, i.e., critical metals < 0.01–0.1 ppb); GC/GC-MS (full-spectrum organic impurity analysis to rule out organics that might interfere with CVD/photolithography processes); LPCM (liquid particle counter — particle count typically must satisfy ≥0.1 µm particles < 50 per mL); moisture content (Karl Fischer titration, especially for highly hygroscopic materials such as HMDS); and physicochemical parameters (density, refractive index, boiling-point distribution via rectification-GC analysis to determine purity). Initial chemical analysis typically requires multiple batch samples to assess lot-to-lot consistency — one of the quality attributes most closely scrutinised by fabs, since unstable lot-to-lot variation causes process drift and harms product yield.

Phase 3: Small-scale process evaluation (~3–6 months). After chemical analysis, the supplier's material enters process testing on the fab's production line. This phase is conducted on a dedicated "evaluation line" (typically part of the actual production line, isolated for new-material evaluation): in the case of domestic TEOS, process engineers will substitute domestic TEOS for the incumbent imported TEOS in the CVD chamber, depositing SiO₂ films under standard process conditions and then comparing the result against a reference standard (imported TEOS under identical conditions) — film thickness uniformity, dielectric constant (k value), breakdown voltage (> 8 MV/cm), fixed-oxide-charge density (Qf), interface-trap density (Dit), and other parameters must all align with the baseline specification (Spec). The entire process typically involves dozens of wafer lots (at least 25 wafers per lot) and takes several months, with any anomalous data requiring root-cause analysis.

Phase 4: Reliability and long-term stability testing (~3–6 months). Some fabs (especially memory fabs) add long-term reliability testing after Phase 3: thin-film devices made with domestic materials are subjected to accelerated testing (high temperature/voltage stress) including TDDB (time-dependent dielectric breakdown) tests and NBTI/PBTI (negative/positive bias temperature instability) tests to evaluate device service life and reliability. This phase can take up to six months and is the hardest element of the qualification cycle to compress.

Phase 5: Volume supply ramp (~3–6 months). After passing all tests and qualification, the supplier is officially added to the QVL, but the initial production period typically begins with "small-scale introduction" (5–10% share replacement of imports), gradually monitoring quality stability under volume-production conditions; only after confirming no anomalies does the fab gradually expand the substitution share. Full substitution (above 50% share) typically requires a further 6–12 months of stable operation after initial volume introduction.

Structural implications of the qualification cycle

Stacking these five phases, the estimated complete qualification cycle by process and application scenario is:

Mature-process (≥28 nm) silicon-based chemicals (TEOS/HMDS): typically 12–18 months
Mainstream advanced logic (14–7 nm) silicon-based chemicals: typically 18–24 months
Leading-edge processes (≤5 nm) or memory (DRAM/NAND) critical-step materials: typically 24–36 months
Top-tier international fabs (TSMC/Samsung) most advanced process materials: typically 36–48 months

This means that even if a Chinese electronic-chemicals company has technically completed all R&D work, a confirmed waiting period of approximately two to four years (qualification) remains before actual volume supply to a fab. This structural delay is the core reason why domestic substitution appears to "progress slowly" — not a deficiency in technical capability per se.

Chapter 4　The Industry Supply Chain: Monomer → Purification → Packaging → End-Use

4.1 Full Supply-Chain Architecture

The supply chain for electronic-grade silicon-based chemicals follows a linear vertical structure of "mineral → smelting → chemical synthesis → purification → application," with independent technical barriers and commercial profit distribution at every stage.

[Upstream] High-purity quartz sand (raw ore, SiO₂ ≥ 99.9%, North Carolina / Jiangsu Donghai)
         ↓
[Smelting] Metallurgical silicon (electric-arc carbothermal reduction, Si purity ~99%)
         ↓
[Fork node]
 ┌─────────────────────────────────────────────────────────────┐
 │ Organosilicone route    Polysilicon route    TEOS route      │
 │ Rochow process          TCS rectification    SiCl₄ alcoholysis│
 │ D4/D5 intermediates     Siemens reduction    TEOS synthesis   │
 │ Electronic silicone gel  Electronic-grade Si  Electronic TEOS │
 │        ↓                       ↓                  ↓          │
 │ Semiconductor packaging  Single-crystal/wafer  CVD thin film  │
 └─────────────────────────────────────────────────────────────┘
         ↓
[Application] Wafer fabs / PV module fabs / battery fabs / fibre fabs
         ↓
[End products] Integrated circuits / solar modules / power cells / optical-fibre cables

4.2 Upstream: Mineral Resources and Metallurgical Silicon

High-purity quartz sand resources are extremely concentrated, forming a "resource endowment barrier" at the very top of the supply chain:

Spruce Pine, North Carolina (USA): The world's most important source of high-purity quartz sand, jointly controlled by Belgian mining group Sibelco and US-based Covia Corporation (formerly Unimin). Spruce Pine quartz sand has SiO₂ content exceeding 99.99%, uniform mineral grain size, and naturally very low metallic impurity levels — the semiconductor industry's most favoured source. A small fire in 2022 (quickly contained and operations soon resumed) briefly triggered global high-purity quartz supply tightness, clearly demonstrating its strategic vulnerability.

Norwegian deposits: The Fennoscandian geological shield contains high-grade quartzite; companies such as Norway Quartz have developed relatively high-quality resources, making Norway the most important alternative source after Spruce Pine.

Donghai County, Lianyungang, Jiangsu, China: Known as "China's crystal hometown," this is the country's largest natural crystal and high-quality quartz mineral production base, with SiO₂ content reaching 99.9%+, supporting large volumes of domestic demand for optical-fibre quartz sand (core/cladding) and photovoltaic crucible quartz sand. Representative producers include Lianyungang Haoyu Quartz and Jiangsu Zhongsheng Silicon Star Technology. However, the proportion of semiconductor-crucible-grade quartz sand (SiO₂ ≥ 99.999%, total metallic impurities < 10 ppm) from domestic deposits is relatively low; ultra-high-purity crucible-grade quartz sand still relies partially on imports (especially raw ore or refined products from Spruce Pine, USA).

At the metallurgical silicon smelting end, China ranks first globally in capacity and competition is intense:

Hoshine Silicon: industrial silicon capacity 1.22 million tonnes/year (H1 2025), globally largest
Yunnan, Sichuan (hydropower low-cost advantage): Yunnan Hongtai Silicon, Sichuan Lingjiao Industry & Trade, etc.
Xinjiang (coal-power low-cost): multiple industrial silicon companies

Metallurgical silicon purity of approximately 98–99% involves a highly mature smelting process with low technical barriers and thin profit margins. As the most upstream bulk chemical, industrial silicon prices fluctuate sharply with market supply and demand (peak approximately RMB 70,000/tonne in 2022, falling sharply to RMB 15,000–20,000/tonne in 2024–2026).

4.3 Midstream: Synthesis and Purification (the Core Value-Creation Interval)

The midstream is the most technically intensive, highest-value-added, and most domestic-capability-deficient segment of the entire supply chain.

Organosilicone monomer preparation

Industrial synthesis of organosilicone monomers uses the Rochow process (direct process) as its core: metallurgical silicon powder (Si) reacts with methyl chloride (CH₃Cl) under copper catalyst in a fluidised bed at approximately 260–320 °C, producing a mixture dominated by dimethyldichlorosilane (DMDCS, M₂):

Si + 2 CH₃Cl → (CH₃)₂SiCl₂ (main product, M₂) + CH₃SiCl₃ (methyltrichlorosilane, M₁) + other by-products

M₂ selectivity in the product is approximately 80–90%; the components are separated by rectification: M₂ hydrolysis-condensation yields D4/D5 cyclosiloxanes; M₁ hydrolysis-condensation yields T-resins; trimethylchlorosilane (M₃, (CH₃)₃SiCl) is the key feedstock for HMDS.

Hoshine Silicon, Xinghuo Chemical (Zhejiang), Luxi Chemical (Shandong), and Dongyue Silicone (Shandong) are the main domestic organosilicone monomer producers. Hoshine's organosilicone monomer capacity of 1.73 million tonnes/year (globally largest) gives it the strongest cost-scale advantage.

Electronic-grade polysilicon purification (highest-value-density step)

Purification of electronic-grade polysilicon is the most technically demanding, capital-intensive, and certification-cycle-heavy step in the supply chain — the most difficult domestic value gap to close:

Xinlightu Silicon (Sichuan, 1,000-tonne/year semiconductor-grade polysilicon project): first batch of samples produced in May 2024, currently in the product-quality ramp-up and customer-qualification phase
Tongwei: core focus is photovoltaic polysilicon; electronic-grade direction is in R&D reserve stage with no scaled capacity
Daqo Energy, GCL Technology, Xinte Energy: all leading PV polysilicon companies; electronic-grade direction is in feasibility study and preliminary exploration

Domestic progress in TEOS purification

Yoke Technology (via UP Chemical) is the most important domestic source of electronic-grade TEOS; its products have entered the procurement lists of certain mature-process fabs. Domestic TEOS substitution rate is estimated at approximately 55–60% overall, but imports (Japan, USA) still dominate absolutely in TEOS qualified for nodes of 14 nm and below. Nantong Chruker and similar companies produce large volumes of industrial/LCD-grade TEOS, but the proportion of semiconductor-grade product is low.

4.4 Downstream: Packaging and Logistics Cleanchain Management

Electronic-grade chemical packaging is the final quality checkpoint in the entire supply chain, and an important hidden barrier that is easily underestimated. SEMI has established the core packaging standards for electronic chemicals:

SEMI C1: Specifies purity requirements by application grade (GRADE 1 through GRADE 5); GRADE 5 (highest purity level) requires metallic ion concentrations of 0.1–1 ppb.

SEMI C8: Specifies container materials for liquid electronic chemicals (PTFE/HDPE-lined drums), cleaning pre-treatment methods, packaging environment requirements (cleanroom operation), and labelling requirements.

Packaging that does not comply with SEMI C8 (e.g., using ordinary PE drums or metal drums) will cause packaging containers to leach metallic ions and organics into the chemicals, creating product secondary contamination and negating all the upstream purification work.

Logistics for electronic-grade chemicals requires dedicated chemical tankers (PTFE-lined interior, temperature-controlled transport) with continuous monitoring and recording. Moisture-sensitive liquids such as TEOS and HMDS in particular require nitrogen-pressurised transport containers with explosion-proof valves, ensuring no moisture ingress, no leakage, and no temperature anomalies throughout transit. Long-distance transport (e.g., from Nantong, Jiangsu to SMIC in Beijing) requires dedicated vehicles, and transport lots must be strictly matched to product test reports to meet the fab's incoming-material traceability requirements.

4.5 Qualification Cycle: The Longest Lead-Time Factor for Domestic Substitution

Even if a domestic supplier demonstrates in laboratory analysis purity and performance comparable to imports, obtaining formal fab purchasing qualification still requires a systematic qualification process of 12–24 months or more. The qualification process typically proceeds as follows:

Phase 1 — Initial screening and evaluation (1–3 months): After the fab's procurement department receives the supplier's sample-submission request, it first conducts a document review (company credentials, ISO certifications, product specification sheets, SDS safety data sheets), followed by a laboratory analysis phase — comprehensive chemical analysis of samples (ICP-MS metals, GC organics, particle size, moisture, etc.) to confirm alignment with the specification.

Phase 2 — Process integration testing (3–6 months): The candidate material is introduced into the fab's actual production process for small-batch trial use, and its effect on key process metrics — thin-film deposition rate, film density, metallic contamination level, device electrical performance — is observed and compared in parallel with an imported reference.

Phase 3 — Reliability and lot consistency validation (3–6 months): Product performance consistency testing across dozens of consecutive lots (typically requiring 30+ lots), assessing whether lot-to-lot variation meets the fab's process window requirements.

Phase 4 — On-site supplier audit (1–3 months): The fab's quality engineers visit the supplier's production base for an on-site audit focusing on cleanroom construction level, analytical instrument provision, production batch traceability, emergency response plans, and production safety capabilities; passing the audit is required before entry to the qualified vendor list (QVL).

Phase 5 — QVL admission and trial supply: After entering the QVL, actual supply typically begins at a small share (e.g., 5–10% replacement of imports); after 6–12 months of stable supply verification, the procurement share may gradually increase.

This lengthy qualification chain is the most critical lead-time factor between "technological breakthrough" and "stable commercial volume supply," and is the fundamental reason why domestic electronic-chemicals companies find it difficult to rapidly increase market share even after achieving technical breakthroughs.

4.6 Packaging Standards: SEMI C8 and Electronic-Chemical Container Technology

Packaging for electronic-grade chemicals is far from simple "drums of liquid" — it is a precisely designed material-contact safety system, the quality-protection element for the "last mile" from producer to end-user.

Overview of the SEMI standards system

The global semiconductor-materials industry universally uses SEMI (Semiconductor Equipment and Materials International) standards, with the core standards directly relevant to electronic-chemicals packaging:

SEMI C1 (Liquid Chemical Classification Standard): Classifies liquid chemicals by metallic impurity content into Grade 1 (< 100 ppb), Grade 2 (< 10 ppb), Grade 3 (< 1 ppb), Grade 4 (< 100 ppt), and Grade 5 (highest purity; some metals < 10 ppt). These five grades correspond directly to different process-node application requirements: Grade 1/2 for 200 mm mature processes; Grade 3 for 300 mm 28 nm/14 nm; Grade 4/5 for advanced nodes of 7 nm and below.

SEMI C8 (Ultra-Clean Liquid Chemical Packaging Standard): Specifies container design requirements, with emphasis on: inner-container material must be HDPE, PTFE, or PFA — ordinary PP or PE must not be used (they leach plasticisers and trace metallic ions into liquids); no lubricants may be used during container manufacture (possible metallic or organic contamination); all inner containers must be rinsed multiple times with ultrapure water (UPW, resistivity 18.2 MΩ·cm) and filled and sealed in a Class 100 (ISO Class 5) cleanroom; each container lot must be accompanied by a Certificate of Analysis (CoA) recording full chemical analysis data (metallic impurities, moisture, particle count, organic content, etc.).

Common electronic-grade chemical packaging specifications

Container type	Material	Applicable products	Common volumes
Ultra-clean PTFE inner-liner drum	HDPE outer + PTFE liner	TEOS / HMDS / high-purity solvents	1 L / 5 L / 18 L / 200 L
Full-PFA container	All-PFA monolayer	Highest-purity HF/TEOS	1 L / 5 L
ISO Container	Electropolished SS outer + PTFE inner bag	Large-volume electronic-grade solvents	1,000 L / IBC
High-pressure cylinder (liquefied grades)	Electropolished SS cylinder	HMDS (some customers use vapour-phase)	1 kg / 5 kg

Gap between China and international packaging standards, and catch-up efforts

Domestic Chinese electronic-chemicals companies are rapidly converging toward SEMI C8 standards, but certain detail gaps persist: some smaller domestic suppliers have not yet fully met container material or cleanroom-filling requirements; the standardisation level of inner-container pre-treatment (acid wash → ultrapure-water rinse → cleanroom seal) varies. Full-PFA containers (needed for highest-purity Grade 4/5 products) have limited domestic production capability and are mostly imported (Japanese and US brands such as those from Entegris dominate the domestic market for clean containers). Standardisation improvement in this "last mile" of the supply chain is also an important part of the domestic electronic-chemicals industry's path to genuine internationalisation.

4.7 Logistics and Supply-Chain Security: Hazardous-Chemical Management System

Electronic-grade silicon-based chemicals in logistics and storage face a dual challenge that ordinary chemicals do not: simultaneously satisfying the safety regulations for hazardous-chemical transport while maintaining the extreme cleanliness required to prevent particulate contamination and moisture ingress during transit.

Hazardous-chemical classification and transport requirements

TEOS (tetraethyl orthosilicate): classified as Class 3 flammable liquid (flash point 48 °C, below 60 °C); must be shipped under flammable-liquid hazmat regulations using certified hazmat vehicles (passing annual inspection, equipped with anti-static facilities and spill-response kits).

HMDS (hexamethyldisilazane): Class 3 flammable liquid (flash point 15 °C, highly flammable) that also reacts with water to generate ammonia (an irritating toxic gas) — dual hazard (flammable + water-reactive toxic), with higher hazard classification than TEOS; transport must strictly prevent moisture contact and ensure container integrity.

Polysilicon (chunk/granular): a solid non-hazmat material, but its dust (ultra-fine silicon powder) is combustible in air and requires explosion-proof handling.

Cold-chain and temperature control management

HMDS and some high-end TEOS products have strict temperature requirements during transport (typically must be maintained at 5–25 °C) to prevent elevated temperatures causing excessive vapour loss or container pressure build-up, while low temperatures may cause some grades to crystallise. This requires temperature-controlled box vehicles, continuous GPS temperature recording, and temperature-compliance confirmation on arrival.

Supply-chain safety inventory strategy

For wafer fabs, a supply interruption from a single electronic-chemicals supplier (especially TEOS or HMDS) directly causes a production line shutdown (Line Down), with enormous losses (a 300 mm fab losing approximately tens of millions of RMB per day from a material-supply-caused stoppage). Domestic fabs therefore typically adopt the following inventory safety strategies for key electronic chemicals (materials that have passed qualification):

Dual-vendor qualification strategy: simultaneously qualifying two or more suppliers (e.g., one domestic + one import), enabling rapid switching if either faces supply disruption; typically requires inventory covering 8–12 weeks of consumption
Strategic safety stock: under geopolitical risk factors such as Japanese earthquake risk or cross-strait tensions, leading domestic fabs have begun extending the inventory cycle for certain critical materials from the previous 4–6 weeks to 12–24 weeks; this further strengthens domestic suppliers' competitiveness on the "supply reliability" dimension — domestic suppliers can achieve shorter replenishment cycles and more flexible emergency-sourcing capability, aligning with fab supply-chain security strategies
Spot market + long-term contract hybrid procurement: large-volume imports (e.g., Wacker TEOS) typically operate under annual or multi-year long-term contracts, while domestic products during initial qualification may begin with small spot batches, gradually transitioning to long-term contracts as qualification progresses and trust accumulates

Chapter 5　Downstream Applications: Wafer CVD / Photolithography Surface / PV Polysilicon / Battery Coating / Optical Fibre

5.1 Wafer Fabs: The Core Battleground for TEOS

In thin-film deposition processes of integrated-circuit fabrication, TEOS is the most important and highest-consumption liquid silicon source. Each 300 mm wafer undergoes dozens of CVD steps during the complete fabrication process, a significant portion of which use TEOS as the silicon source, consuming cumulatively several to tens of grams of TEOS per wafer. Based on global annual production of approximately 90 million 300 mm equivalent wafers and consumption of approximately 10–20 grams of TEOS per wafer, global fab TEOS annual consumption is approximately 900–1,800 tonnes, representing a market of approximately RMB 500 million to 1.5 billion (at electronic-grade TEOS prices of approximately RMB 100–300/kg).

TEOS applications in wafer fabrication, by process type:

PECVD-TEOS (plasma-enhanced CVD): Under plasma activation, TEOS decomposes and deposits SiO₂ films at low temperatures of 350–450 °C. This is currently the most mainstream TEOS application. Typical uses: aluminium-interconnect interlayer dielectric (ILD), chip passivation layer, and partial STI fill. Advantages: low deposition temperature (compatible with temperature windows above aluminium interconnect layers), good film uniformity. Disadvantage: relatively high H content in the film (introduced via TEOS ethoxy groups); without subsequent high-temperature densification annealing, the film's wet-etch rate is somewhat elevated.

SA-CVD TEOS (sub-atmospheric CVD): Under sub-atmospheric pressure (approximately 2,000–15,000 Pa) at higher temperatures (450–750 °C), TEOS thermally decomposes to deposit SiO₂. Suitable for filling high-aspect-ratio trenches (such as copper-interconnect vias and trenches), because sub-atmospheric pressure aids reactive-gas diffusion into high-aspect-ratio structures, achieving seam-free fill. This is the application scenario with the most stringent TEOS purity requirements at advanced nodes (14 nm and below).

HDP-CVD TEOS (high-density plasma CVD): Combines CVD deposition with ion-sputter etching to achieve a flowable fill effect; preferentially used for high-aspect-ratio STI fill. SMIC's 28 nm/14 nm STI processes already use TEOS as the silicon source in this application.

TEOS-SOG (spin-on glass): TEOS formulated with specific solvents and catalysts forms a spin-coatable precursor solution; after spin-coating, it is cured at low temperature (300–450 °C) into SiO₂ glass, used for global surface planarisation — a pre-CMP step.

TEOS demand at China's wafer fabs is growing rapidly: SMIC has expanded to approximately 950,000 monthly 300 mm equivalent wafers; Hua Hong Semiconductor is continuing to expand specialty-process (power devices/CMOS image sensors) capacity; CXMT DRAM capacity is continuously ramping; YMTC NAND Flash capacity is expanding rapidly. These ongoing domestic fab expansions create the most direct domestic demand window for Chinese TEOS.

5.2 Lithography: The Hidden Value of HMDS

Photolithography is the critical patterning step in chip fabrication, and the most costly and precise process of all. Every 1-percentage-point improvement in lithography yield represents substantial cost savings and utilisation-rate gains. HMDS surface treatment is one of the important foundational processes for maintaining lithography yield.

In mature-process fabs (28 nm and above), HMDS treatment is an almost indispensable standard step in the pre-lithography processing flow. A typical pre-lithography processing module (TRACK system) process sequence is:

Incoming wafer → Cleanliness inspection → Dehydration bake (150–200 °C, N₂ atmosphere) → HMDS vapour-phase treatment (90–120 °C, 30–90 sec) → Cool/cold-plate cooling (25 °C) → Photoresist spin-coating → Soft bake (80–110 °C) → EUV/DUV exposure → Post-exposure bake (PEB) → Development (TMAH developer, 2.38%) → Hard bake (optional) → Inspection/metrology

During the HMDS vapour-phase treatment step, wafers receive HMDS vapour treatment in a sealed HMDS treatment chamber (typically a dedicated chamber within the TRACK system). After treatment, wafers transfer to the photoresist coating chamber for spin-coating; the HMDS-modified surface significantly improves photoresist wetting and spreading on the wafer surface, reducing coating bubbles and edge defects.

In practice, HMDS consumption per wafer is relatively small (equivalent to approximately 1–5 mL of liquid HMDS per wafer), but because of its foundational role in lithography yield control, requirements for quality stability and consistency are extremely high. For fabs, switching HMDS suppliers requires a complete qualification process, and even after a successful switch, each new supplier batch must be process-monitored to verify no adverse impact.

Domestic fab qualification and introduction of domestic HMDS is progressing fastest at SMIC's mature-process lines; Hua Hong Semiconductor is also evaluating some domestic HMDS in its power-device lines. Overall, however, HMDS's domestic substitution rate (~25–30%) is the lowest among all major electronic-grade silicon-based chemicals and remains a critical weak link.

5.3 PV Polysilicon: The Largest-Scale Silicon-Material Application

PV polysilicon is the largest-volume and highest-output category among all "silicon materials" and the segment where China's silicon supply-chain advantages are most concentrated.

2025–2026 PV polysilicon supply-demand landscape

In 2025, global photovoltaic installations continued to grow rapidly, with global new PV installations exceeding 700 GW (China accounting for approximately 70%), keeping polysilicon demand persistently high. However, after years of large-scale capacity expansion by Chinese polysilicon companies, supply growth has far outpaced demand growth, creating severe capacity overcapacity:

Global polysilicon nominal capacity (end-2024): more than 2 million tonnes/year
Global polysilicon actual demand (2025): approximately 1.4–1.6 million tonnes
Capacity utilisation rate: approximately 70–80%, with some companies below 60%

This overcapacity has kept Chinese PV polysilicon prices persistently below full production cost (mono-grade polysilicon in China approximately RMB 50–70/kg in 2025, with some companies' full cost approximately RMB 45–55/kg — very close to breakeven). In March 2026, the international price of China's mono-premium polysilicon was approximately USD 6.47/kg (~RMB 47/kg), down more than 13% from the start of the year — a historic trough.

Technology-route evolution for PV polysilicon

With the rapid commercialisation of TOPCon (tunnel-oxide passivated contact), HJT (heterojunction), and perovskite-tandem next-generation high-efficiency solar cell technologies:

N-type mono feedstock demand rising rapidly: TOPCon/HJT cells both require N-type single-crystal silicon substrates (with lower tolerance for B and P dopant impurities), driving the market toward higher quality ("ultra-solar-grade") N-type mono polysilicon approaching electronic-grade thresholds; leading companies (Tongwei, GCL) have raised their N-type share to above 50%
Granular silicon penetration accelerating: GCL Technology FBR granular silicon's cost advantage (RMB 28.17/kg, Q4 2024) is driving a structural shift in the market from rod silicon to granular silicon
Technical requirements for polysilicon from PV are rising, but have not yet reached the electronic-grade (11N) threshold; the technical watershed between the two remains clear

The key role of silane coupling agents in photovoltaics

Photovoltaic-sector consumption of silane coupling agents is closely tied to module shipment volumes:

EVA (ethylene-vinyl acetate copolymer) and POE (polyolefin elastomer) are the core encapsulant materials for PV modules; both require 1–3 wt% silane coupling agent in the formulation to improve interfacial bond strength between the encapsulant and the glass cover, crystalline silicon cells, and back-sheet, ensuring that no delamination occurs during the 25-year outdoor service life.

China's PV module output in 2025 exceeded 800 GW, corresponding to EVA/POE encapsulant demand of more than 2 million tonnes (at 2.5 kg/kW), with supporting silane coupling agent consumption of approximately 20,000–60,000 tonnes/year (estimated from formulation addition ratios) — the single largest domestic application market for silane coupling agents.

5.4 Lithium-Ion Battery Separators: A High-Growth New Battleground for Silane Coupling Agents

Lithium-ion battery separators are the last line of defence for battery safety: as battery temperature rises, the separator must block ion transport through a pore-shutdown mechanism to prevent thermal runaway. Base PE/PP separators have limitations in thermal performance and electrolyte wettability; ceramic coating (Al₂O₃/boehmite) or PVDF coating greatly improves thermal shrinkage resistance and electrolyte imbibition.

In the ceramic-coating (Al₂O₃/boehmite) process, silane coupling agents play a dual role:

Function 1: Surface modification and dispersion of nano-ceramic particles. Nano-Al₂O₃ (particle size 100–500 nm) tends to agglomerate and settle in aqueous slurry, impairing coating uniformity and coating density. Treatment of Al₂O₃ surfaces with KH-570 (methacrylate silane coupling agent) — through chemical bonding between Si-OH and surface aluminium hydroxyl (Al-OH) groups — coats the particles with an organofunctional surface layer, improving the hydrophilic-hydrophobic balance of the particles to maintain uniform dispersion in the slurry and improve coating quality.

Function 2: Interface adhesion between coating and base film. The inherent interfacial energy difference between a ceramic coating (inorganic phase) and a PE/PP base film (organic phase) is large and prone to debonding. The silane coupling agent acts as a molecular bridge: one end chemically bonds to the ceramic particles/base-film SiO₂ via Si-OR, and the other end physically or chemically interacts with the PE/PP base film, improving coating bond strength and preventing the coating from delaminating or cracking during winding, cutting, and charge/discharge cycling.

As EV demand for range and fast-charging capability continuously pushes batteries toward "thinner separators (≤5 µm) + higher energy density," requirements for ceramic-coating quality (uniformity, adhesion, thermal stability) are becoming increasingly stringent, driving sustained growth in demand for high-end silane coupling agents (especially specialty functionalised grades). Domestic lithium-battery-separator silane-coupling-agent demand is estimated to exceed 5,000 tonnes in 2025, with potential to surpass 8,000 tonnes by 2027.

5.5 Optical Fibre Preforms: Silicon-Based Chemicals in Optical Communications

Optical fibre is the physical carrier of modern internet infrastructure; its core material is high-purity fused silica glass (SiO₂ content > 99.99%). Optical fibre works on total internal reflection: light propagates in the higher-refractive-index fibre core (n₁) and undergoes total internal reflection at the core-cladding interface (cladding n₂, n₂ < n₁), locking the light in the core for long-distance propagation. To achieve this, the core must be doped with GeO₂ (higher refractive index than SiO₂), or the cladding doped with F (to lower refractive index), creating the required refractive-index gradient.

The fabrication of optical fibre preforms is the most technically intensive step in fibre production, primarily using two CVD processes:

OVD (Outside Vapour Deposition): Using SiCl₄ as the silicon source, hydrolysed by a hydrogen-oxygen flame to produce SiO₂ soot particles that are deposited layer-by-layer on the outer surface of a rotating target rod to form a porous soot preform; subsequent dehydration (SiF₄ atmosphere) and consolidation yield a transparent solid quartz-glass preform.

MCVD (Modified Chemical Vapour Deposition): Using SiCl₄ and GeCl₄ as feedstocks, vapour-phase deposition on the inner wall of a quartz tube forms a graded-index core/cladding structure, primarily developed by AT&T Bell Labs.

Chinese companies — Yangtze Optical Fibre (the world's largest optical-fibre-cable maker), ZTT, Fiberhome Communications (under China Information Technology), and Hengtong Optic-Electric — are all among the world's leading optical-fibre-cable manufacturers, together accounting for over 50% of global output. Their production processes use large quantities of high-purity SiCl₄ and TEOS (in some processes) as silicon deposition precursors, as well as silane coupling agents (KH-550, etc.) to improve optical-fibre coating adhesion to the quartz-glass surface. Rapid expansion of optical-communications infrastructure (5G backhaul networks, hyperscale data centres, submarine cables) will continue to support demand growth in this direction.

Fibre technology evolution and its impact on preform-material requirements

As data-centre scale continues to expand (especially AI large-model training clusters' demand for hyperscale compute), requirements for fibre transmission capacity (400G/800G/1.6T Ethernet) and distance (100 m within the data centre to several km between campuses) are escalating, driving the rapid development of new fibre technologies:

Multi-mode fibre (MMF, OM4/OM5): for short-distance (≤100 m) high-density intra-data-centre interconnection; optical design places extremely high purity and refractive-index-uniformity requirements on SiO₂ and GeO₂
Ultra-low-loss single-mode fibre (ULLSF): transmission loss < 0.16 dB/km, requiring extremely low OH⁻ content in the preform (< 1 ppb) — the most stringent moisture-control requirement on any silicon source (SiCl₄/TEOS)
Hollow-core photonic-bandgap fibre (HC-PBGF): an emerging research direction in which light propagates in air rather than solid glass, with extraordinarily stringent purity requirements on the glass wall; expected to open new applications in ultra-low-latency communications (financial high-frequency trading, quantum communications)

These technology advances place continuously escalating quality demands on fibre-grade high-purity TEOS/SiCl₄, meaning that the fibre market's demand for silicon-based chemicals will not stagnate because "fibre is a mature technology" but will continue to generate new quality-uplift demand as the technology evolves.

5.6 Perovskite Solar Cells: The Next-Generation Growth Driver for Silane Coupling Agents

Perovskite solar cells, having rapidly raised their power conversion efficiency from below 10% to above 26% (single-junction laboratory record) over the past decade, are advancing from laboratory to industrial scale and are widely regarded as the "disruptor" of next-generation photovoltaic technology.

In the structure of perovskite PV devices (N-i-P or P-i-N heterojunctions), silane coupling agents play the following key roles:

Interface passivation enhancement: Interface defects (trap states) between the perovskite absorber layer (ABX₃ type, e.g., methylammonium lead iodide CH₃NH₃PbI₃) and the electron-transport layer (SnO₂, etc.)/hole-transport layer (Spiro-OMeTAD, etc.) are a critical factor affecting device efficiency and stability. Silane coupling agents with specific functional groups (such as amino, carboxyl, or Lewis-base groups) can effectively passivate interface defects by coordinating with defect sites such as lead ions (Pb²⁺) and iodine vacancies (VI) on the perovskite surface, improving device open-circuit voltage (Voc) and fill factor (FF), and thereby improving efficiency.

Moisture/air barrier enhancement: A key weakness of perovskite is its extreme moisture sensitivity — even trace moisture causes perovskite to decompose (CH₃NH₃PbI₃ rapidly decomposes in moisture into PbI₂, CH₃NH₂, and HI). Silane coupling agents with hydrophobic functional groups, applied during device encapsulation, can build a moisture-barrier layer through hydrophobic surface modification, extending device service life.

Integration into silicon-perovskite tandem cells: Tandem cells connecting perovskite (wide bandgap, Eg ~1.6 eV) and crystalline silicon (narrow bandgap, Eg ~1.1 eV) in series have a theoretical efficiency above 40% and represent the most advanced direction in current photovoltaic technology. Tandem cells simultaneously involve silicon-based materials (bottom cell) and perovskite materials (top cell), placing new chemical requirements on the interface treatment between the two layers (and on silicon surface passivation), creating new high-end application scenarios for silane coupling agents.

As perovskite photovoltaics advances from laboratory toward GW-scale production (LONGi Solar, CHN Energy Clean Energy, GCL Optotech, and others have begun layout of perovskite mass-production lines), demand for supporting high-end functionalised silane coupling agents will enter a rapid-growth phase in 2027–2030 — an important strategic window for domestic coupling-agent companies to upgrade to higher-value-added product grades.

Chapter 6　Major Players Roundup

6.1 International Leaders: Multi-Dimensional Technological Moats

Shin-Etsu Chemical (Japan)

Shin-Etsu Chemical is an undisputed super-platform player in the global silicon-based materials sector. Its businesses span: organosilicone (DMC organosilicone fluid, global No. 1, approximately 30% market share); semiconductor wafers (global No. 1, approximately 30% market share); TEOS and CVD precursor materials; and photoresists (KrF/ArF photoresists, global top three) — any one of these standalone businesses would already be among the world's top players in its sub-segment.

Most important recent strategic move: the announcement of an approximately USD 700 million (~JPY 100 billion) domestic Japan investment to expand high-purity silicon-materials capacity, explicitly for supply to top-tier fabs operating at sub-3 nm nodes — including TSMC's Kumamoto Phase II plant (expected opening late 2025) and Samsung's Pyeongtaek facility — signalling Shin-Etsu's long-term optimism about global advanced-process materials demand.

Shin-Etsu's China presence: through local technical service teams and limited in-China production (certain silane coupling agents and organosilicone intermediates supplied locally), deeply serving China's PV module, LED packaging, and semiconductor-packaging markets — but core high-purity polysilicon and advanced-process precursor products are manufactured in Japan, maintaining full control over critical technology.

Wacker Chemie (Germany)

Wacker is Europe's most important polysilicon and organosilicone chemicals company, operating on two parallel tracks: on one side, its Burghausen plant produces semiconductor-grade polysilicon and, together with Hemlock, controls approximately 75% of the global semiconductor-grade polysilicon market; on the other side, its Wacker Silicones business is a top-three global organosilicone monomer and downstream product supplier, with large volumes of D4/D5 base raw materials and various functional silicone oils and gels. Wacker's China strategy is to operate a local organosilicone production base in Yangzhou (Yangzhou Ankang, producing organosilicone intermediates and certain functional products), combining local production with technical support services to cover the Chinese market while maintaining core high-purity polysilicon production in Europe.

Hemlock Semiconductor (USA)

Hemlock Semiconductor — a joint venture of Dow Chemical (63.25% stake) and Shin-Etsu Chemical (36.75% stake), headquartered in Hemlock, Michigan — is North America's largest polysilicon producer, with annual capacity of approximately 25,000–30,000 MT, the overwhelming majority semiconductor-grade. Hemlock's core competitive advantage lies in the quality reputation built over decades: its "Hemlock Brand" semiconductor-grade polysilicon has a track record of decades of supply to Intel, TSMC, Samsung, and other top-tier fabs, deeply embedded in the global semiconductor supply chain. For any new polysilicon supplier seeking to displace Hemlock, the requirement to undergo years of rigorous fab qualification comparison creates extremely deep barriers.

Tokuyama (Japan)

Tokuyama is one of Japan's major chemical companies. Its Shunan plant is Japan's most important high-purity polysilicon base, with annual capacity of 8,000–12,000 MT, all semiconductor-grade, with stable and reliable product quality and supply history to ARM, TSMC, Intel, and other top-tier customers. In Q1 2024, Tokuyama completed a new expansion round at Shunan of approximately JPY 30 billion, dedicated to next-generation polysilicon purification units for sub-3 nm nodes. In July 2025, Tokuyama and OCI (South Korea) announced a joint venture to build a 10,000 MT/year polysilicon plant at the Samalaju Industrial Park in Sarawak, Malaysia, targeting 2027 commissioning — a major step by a Japanese company to establish offshore polysilicon capacity in Southeast Asia to reduce Japan's high labour cost exposure while serving Southeast Asian and mainland Chinese markets from a nearby location.

Entegris (USA)

Entegris is the world's leading comprehensive semiconductor-materials solutions provider. Its 2022 acquisition of CMC Materials for USD 6.7 billion broadened its product lines to cover CMP slurries and polishing pads, high-purity chemical-delivery systems (including piping/filters/pumps), ALD/CVD precursors (TEOS, HMDS, and metal precursors), and photoresist ancillary materials. Entegris's TEOS and HMDS products have passed qualification for sub-7 nm advanced processes and are core procurement items at the world's top-tier fabs; in the Chinese market (SMIC, Hua Hong advanced-process lines) they are also a major foreign-funded supplier.

Merck KGaA / Versum Materials (Germany/USA)

Germany's Merck KGaA (a science and technology specialty-chemicals giant unrelated to the US pharmaceutical company of the same name) acquired Versum Materials (spun off from Air Products' semiconductor business) in 2019 for USD 5.7 billion, becoming one of the world's most important semiconductor specialty-materials suppliers. Versum's HMDS, low-k dielectric precursors (DEMS, OMTS, etc.), fluorinated etchants, developers, and other products are core slots in the qualified-supplier lists of TSMC and Intel for advanced processes.

OCI (South Korea)

OCI is South Korea's largest polysilicon producer, with significant positions in both PV and semiconductor-grade polysilicon. In January 2024, OCI implemented an improved Siemens process, improving production energy efficiency by approximately 15% and polysilicon yield by approximately 12%. As noted above, OCI is partnering with Japan's Tokuyama to jointly build new capacity in Malaysia's Sarawak, positioning a lower-cost Southeast Asian supply-chain node. OCI's semiconductor-grade polysilicon primarily supplies South Korean fabs (Samsung, SK Hynix).

6.2 China Corporate Tier Analysis

Yoke Technology (002409.SZ — breakthrough precursor player)

Yoke Technology started as a flame-retardant manufacturer (phosphorus-based, nitrogen-phosphorus synergistic), then around 2017 transformed into a semiconductor-precursor supplier through the acquisition of Korea's UP Chemical, becoming China's most important ALD/CVD precursor listed company.

Yoke's precursor capacity currently ranks first domestically and third globally, with main products supplying SK Hynix's HBM (high-bandwidth memory). In the TEOS direction, domestic TEOS pricing is approximately RMB 70–80/kg (vs. imported approximately RMB 100/kg, a cost reduction of approximately 20–30%), and Yoke has entered the procurement lists of certain customers, demonstrating substantive domestic-substitution progress. The company is further developing High-k metal-oxide precursors (ALD hafnium oxide, zirconium oxide, etc. for advanced-node gate dielectric layers) and ultra-high/low-temperature CVD silicon-type precursors, with some products already in customer-side testing, preparing for the next phase of upgrading to advanced-node precursors.

The explosion of AI large models driving continued high growth in HBM market demand is a strong tailwind for Yoke: according to guidance from Lumentum, SK Hynix, and others, the global HBM market is projected to grow from approximately USD 1 billion in 2021 to approximately USD 13 billion by 2030 (CAGR ~30%), and Yoke, as one of the core HBM precursor suppliers, will directly benefit.

Hoshine Silicon (603260.SH — vertically integrated organosilicone leader)

Hoshine Silicon is China's largest integrated industrial-silicon-to-organosilicone-monomer company, and the globally largest single-plant organosilicone monomer producer. In H1 2025, its industrial silicon capacity was 1.22 million tonnes/year and organosilicone monomer (DMC) capacity 1.73 million tonnes/year — both globally first.

The company's 2025 strategic focus is extending from cyclical-bulk organosilicone toward higher-value-added electronic chemicals: in the organosilicone deep-processing direction, it has successfully industrialised amino silicone oil and silicone emulsion to internationally leading quality levels; it is developing electronic-grade silicone gel product lines for the medical, electronics, and new-energy markets, with semiconductor-packaging and LED-application qualification samples expected to be launched in 2026. Hoshine's 2025 full-year revenue was approximately RMB 26.6 billion, with net profit approximately RMB 1.7 billion, demonstrating good cost control and business resilience against a cyclical low in the organosilicone industry. Its R&D pipeline into third-generation semiconductor (SiC, silicon carbide) related chemicals (such as SiC epitaxy precursors) is also an important future watch-point.

Tongwei (600438.SH — global polysilicon leader)

Tongwei is the world's largest producer of high-purity crystalline silicon (polysilicon), with more than 900,000 tonnes/year of high-purity crystalline silicon capacity in 2025, integrated upstream-downstream with solar cell sheet capacity (more than 80 GW/year), forming a "polysilicon-cell" vertically integrated business — one of the most vertically integrated companies in China's photovoltaic supply chain.

Tongwei's core competitive strengths are extremely low polysilicon production cost (full cost in the industry's lowest range, built on Sichuan hydropower low cost and accumulated process optimisation at the Leshan base) and highly stable product quality, making it one of the preferred core polysilicon suppliers for domestic PV module companies (LONGi, JA Solar, Jinko, etc.). Its granular-silicon (beads polysilicon) technology has formed an independent technology pathway competing for market share against Siemens rod silicon. In the electronic-grade polysilicon direction, Tongwei's focus remains on PV-grade products, with electronic-grade research in the reserve stage and no scaled electronic-grade polysilicon capacity or market yet.

GCL Technology (3800.HK — granular-silicon pioneer)

GCL Technology is the world's largest granular-silicon supplier, with four granular-silicon production bases (Xuzhou, Leshan, Baotou, Hohhot) totalling approximately 300,000–400,000 tonnes/year capacity, primarily FBR granular silicon, with low cost and scale as core competencies. In 2025, granular-silicon market share is approximately 12%, with production costs already down to RMB 28.17/kg (Q4 2024) — the lowest-cost polysilicon form at present.

Whether GCL Technology can achieve a breakthrough in the electronic-grade direction for its granular silicon will be an important observation point for tracking market dynamics in 2026–2027: if the FBR process can achieve a breakthrough in electronic-grade purity control, granular silicon's cost advantage could become a disruptive force in the electronic-grade polysilicon market.

Daqo New Energy (688303.SH)

Daqo New Energy, headquartered in Shanghai with core capacity in Shihezi, Xinjiang, is one of the world's top-five polysilicon producers. In 2025, its polysilicon capacity is approximately 150,000 tonnes/year (Siemens rod silicon), primarily high-quality N-type mono feedstock — a premium raw-material supplier for TOPCon/HJT cell technology routes. In Q3 2024, Daqo's polysilicon unit cash cost was RMB 38.93/kg, demonstrating good cost control, though profitability pressure is significant with industry prices persistently below cost.

Shanghai Sinyang (300236.SZ — high-growth precursor company)

Shanghai Sinyang has deep expertise in semiconductor electronic-chemicals subsegments, with products spanning copper-interconnect electroplating solutions, ALD precursors, specialty cleaning chemicals, and photolithography ancillary materials. In 2025, the company achieved revenue of RMB 1.937 billion (up 31.28% year-on-year) and net profit of RMB 301 million (up 71.12% year-on-year), one of the fastest-growing and most closely followed domestic electronic-chemicals listed companies in the capital market. Shanghai Sinyang's precursor-category breakthroughs are concentrated primarily in electroplating solutions and specialty additives for copper interconnect processes and certain ALD precursors, with less dedicated focus on TEOS — but the rapid revenue and profit growth reflects its overall technical-R&D and customer-introduction capability.

Jianghua Microelectronics (603078.SH — steady wet-electronic-chemicals player)

Jiangyin Jianghua Microelectronics Materials Co., Ltd. is one of the most representative listed companies in China's wet electronic-chemicals (WEC) industry. Products span ultra-clean high-purity reagents (hydrofluoric acid, hydrochloric acid, nitric acid, ammonia, hydrogen peroxide, etc., SEMI Grade 1 to Grade 4) and photoresist ancillary reagents (developer diluents, photoresist-strip solutions), primarily for wet processing steps in display panels (LCD/OLED), semiconductor chips (mature processes), and solar cells. H1 2025 revenue was RMB 459.06 million (up 5.80% year-on-year), gross margin approximately 39.29% — above-average profitability for the electronic chemicals industry. Customers include SMIC, Hua Hong, TCL CSOT, and BOE among China's leading panel and semiconductor companies.

Xingfa Group (600141.SH — phospho-silicone synergy player)

Xingfa Group is one of Hubei Province's most important comprehensive chemical companies, headquartered in Yichang and built on Yichang's abundant phosphate and silicon mineral resources. It has constructed a synergistic industrial system spanning phosphochemicals (phosphoric acid, phosphates), organosilicone (Xingfa Organosilicone, with Xingshan/Yidu as core production bases), and electronic chemicals (electronic-grade hydrogen fluoride HF). Xingfa's electronic-chemicals segment uses organosilicone products (silane coupling agents, specialty organosilicone) and electronic-grade HF as two main tracks, making it an important player in the domestic organosilicone-fluoride supply chain. With Hubei Province's semiconductor ecosystem (Wuhan National Memory Technology Base, YMTC, etc.) continuing to be built out, Xingfa has geographic and policy-support advantages in the regional market.

6.3 Korean Electronic-Chemicals Companies: Active Participants in the Chinese Market

South Korea is, after Japan, the most deeply engaged external source of competition in the electronic-grade silicon-based chemicals space relative to the Chinese domestic market. Korean companies supply TEOS and certain ALD precursors to the Chinese market, and their commercial openness is relatively high (benefiting from their own semiconductor industry's strong domestic demand, Korean companies have accumulated advanced process capability while having commercial motivation to expand into the Chinese market).

SK Materials / SK Specialty (South Korea)

SK Materials (now restructured as SK Specialty) is SK Group's specialty electronic-chemicals and specialty-gases subsidiary, with products spanning high-purity TEOS (mature and certain advanced processes), high-purity ammonia (NH₃), nitrogen trifluoride (NF₃, for CVD-chamber cleaning), hydrogen fluoride (HF), and multiple ALD precursors (including metal-organic precursors). SK Materials has a deep supply relationship with Samsung Electronics and SK Hynix, with products qualified under both companies' rigorous certification systems — important credential backing for its entry into Chinese fab supply chains. Currently, SK Materials primarily exports to China (no domestic production facilities), supplying certain electronic-grade materials to SMIC, Hua Hong, and other Chinese fabs through agents or direct sales.

Soulbrain (South Korea, formerly Dongchin SemiChem)

Soulbrain is another significant South Korean listed electronic-chemicals company, with revenue of approximately KRW 500 billion (~RMB 2.7 billion). Main products include etchants (HF/BHF/NH₄F systems for SiO₂ and Si₃N₄ etching), developers (TMAH, tetramethylammonium hydroxide), TEOS (mature and certain advanced processes), and certain ALD precursors. Soulbrain's market-opening approach in China is similar to SK Materials — primarily export supply — and in its business relationships with SMIC, Hua Hong, and other domestic fabs, it occupies the "party being substituted" role in the domestic-substitution wave. As China's chemical domestic-substitution rate rises, Soulbrain's China market share faces sustained pressure from domestic companies such as Yoke Technology and Shanghai Sinyang.

OCI Materials (subsidiary of OCI Group, South Korea)

OCI Materials (OCI Group's electronic-chemicals subsidiary) focuses on liquid-phase precursors for semiconductor CVD processes (including certain TEOS grades) and certain specialty gases, synergistic with OCI's polysilicon main business — using SiCl₄ by-product from polysilicon production to extend downstream into TEOS and other liquid silicon sources. This "by-product value-enhancement" business model is logically highly similar to China's Xingfa Group's "phosphochemicals + organosilicone synergy" model.

6.4 Emerging Players and New Segments: Third-Generation Semiconductor Supporting Chemicals

SiC (Silicon Carbide) Epitaxy Growth Precursors

Third-generation semiconductors (SiC, GaN) are a strategic-emerging-materials direction receiving priority policy support in China. SiC, because of its high-temperature endurance, high breakdown electric field, and high thermal conductivity, is extremely valuable in new-energy-vehicle power devices (IGBT/SiC MOSFET) and fast-charging applications.

SiC epitaxial layer growth uses chemical vapour deposition (SiC CVD/SiC-EP), with main precursors:

Silicon source: silane (SiH₄, gas), methyltrichlorosilane (CH₃SiCl₃, MTS, liquid)
Carbon source: propane (C₃H₈, gas), ethylene (C₂H₄, gas), some processes use methane (CH₄)
Dopant source: nitrogen (N₂, n-type doping), trimethylaluminium (TMA, p-type doping)

Among these, MTS (methyltrichlorosilane, CH₃SiCl₃) is an emerging precursor in the liquid organosilicone-chemical category, containing both silicon and carbon and able to simultaneously supply both major elements for SiC epitaxial growth while reducing CVD-system complexity — an increasingly important feedstock for high-speed SiC epitaxy processes. Domestic MTS sources: Xinghuo Organosilicone (Jiangxi Anyuan) and Bluestar Chemical (Beijing) have industrial-grade MTS supply capability, but electronic-grade (ultra-high-purity) MTS currently relies primarily on imports (Japan Shin-Etsu, US Air Liquide).

GaN (Gallium Nitride) Precursors

MOCVD (metalorganic CVD) growth of GaN-based semiconductors (used in 5G radio-frequency, GaN fast-chargers, power devices), with key precursors:

Group III metal-organic sources (MOCVD-grade): trimethylgallium (TMGa), triethylgallium (TEGa) — these liquid organometallic compounds are the core Group III precursors for GaN MOCVD, with extremely stringent purity requirements
Nitrogen source: high-purity ammonia (NH₃, gas)
Dopant source: bis(cyclopentadienyl)magnesium (Cp₂Mg, for p-type GaN Mg-doping)

TMGa/TEGa are highly hazardous organometallic compounds (pyrophoric in air, explosive on contact with water), with extremely stringent storage and transport requirements. They are currently 100% import-dependent (primarily from Germany's Umicore and Dow Chemical / Rohm and Haas Electronic Materials / Entegris) — one of the most urgently needed "chokepoint" materials in China's third-generation semiconductor supply chain.

Overall assessment: The specialty chemical precursors required for third-generation semiconductors (SiC/GaN) are an important extension of the electronic-grade silicon-based chemicals supply chain. Their technical barriers are no lower than those of TEOS/HMDS, and some categories (such as TMGa) have a domestic substitution rate approaching zero, placing them at extremely high strategic priority. With Big Fund Phase III support, dedicated breakthrough plans for this segment are expected to emerge in 2026–2028, potentially attracting substantial industrial capital from the new-energy-vehicle supply chain.

Chapter 7　Domestic Substitution Profile and Manufacturing Database Insights

7.1 Domestic Substitution Rates: Ice and Fire Across Category Tiers

Based on the Research Institute's systematic analysis, the domestic substitution rates across electronic-grade silicon-based chemical categories currently vary dramatically, presenting a stratified pattern of "bulk/commodities fully substituted, specialty/fine chemicals weak, advanced-process weakest":

Category	Estimated domestic substitution rate	Primary constraining factors
Industrial-grade TEOS	≥95%	No significant barriers, fully competitive
Electronic-grade TEOS (≥28 nm mature)	~55–65%	Purity qualification, lot consistency
Electronic-grade TEOS (<14 nm advanced)	~15–25%	Ultra-high purity + extremely long customer qualification
Industrial/electronic-grade HMDS	~25–30%	Purification process + customer admission barriers
Organosilicone monomer D4/D5 (industrial)	≥90%	Essentially self-sufficient
Electronic-grade silicone gel (semiconductor packaging)	~40–50%	Low impurity + customer qualification
Silane coupling agents (standard bulk grades)	≥85%	Mature technology, sufficient domestic supply
Silane coupling agents (high-end specialty functionalised)	~50–60%	Special functional-group synthesis
PV-grade polysilicon (6N–8N)	≥90% (global capacity)	Absolute global dominance, overcapacity
Electronic-grade polysilicon (9N–10N)	~30–40%	Process qualification, scale still weak
Semiconductor-grade polysilicon (11N)	~15–25%	Ultra-high purification + qualification
High-purity quartz sand (fibre/PV crucible ordinary grade)	≥70%	Resources relatively abundant, technology achievable
High-purity quartz sand (semiconductor crucible ultra-pure grade)	~30–40%	Ultra-pure ore scarce, high technical barriers

The core pattern revealed by this table is: the closer to the highest specification (advanced process), the lower the domestic substitution rate. This is because higher-specification products have higher technical barriers, longer customer qualification cycles, higher qualification-process technical thresholds, and less accumulated experience on the part of domestic companies relative to international giants. In short, "what was achievable domestically has already been substituted; what remains unsubstituted is the genuinely hard challenge."

7.2 The 4.8-Million Factory Database: A Map of China's Silicon-Based Chemical Suppliers

Tianxia Gongchang's Research Institute has systematically mapped and analysed domestic silicon-based chemical manufacturing companies and their regional distribution using its real-time database covering 4.8 million verified in-production Chinese factories. These 4.8 million factories have all been validated by a proprietary "factory identification algorithm" that distinguishes them from trading companies, distributors, and other non-manufacturing entities — making it one of China's most complete in-production factory atlases.

Electronic-grade chemicals and wet electronic chemicals: The database records more than 80 in-production factories, highly concentrated in:

Jiangsu Province (southern and central Jiangsu): approximately 40–45% of in-production factories are in Jiangsu, centred on the Jiangyin (Jianghua Micro headquarters), Nantong (Chruker TEOS, etc.), and Suzhou (concentrated foreign-invested companies) triangle
Shanghai and Yangtze-Delta Zhejiang: Shanghai Sinyang and certain fine-chemicals companies are distributed here
Guangdong (Pearl River Delta): electronic-chemicals suppliers serving South China's semiconductor and display-panel customers

PV polysilicon: In-production factory distribution in the database is concentrated in Leshan/Xichang, Sichuan (Tongwei polysilicon main production base), Shihezi/Urumqi, Xinjiang (Daqo Energy, Xinte Energy), Inner Mongolia (GCL polysilicon), and Yunnan (certain organosilicone and polysilicon).

Silane coupling agents: This is the sub-category with the largest number of market participants and the most complete domestic substitution among all silicon-based chemicals. The Research Institute's database records more than 130 in-production factories, widely distributed (Zhejiang, Jiangsu, Shandong, Guangdong all have presence) — a typical category exhibiting "fully competitive" market characteristics.

High-purity quartz sand: Over 100 in-production factories recorded, heavily concentrated in Donghai County, Lianyungang, Jiangsu (the country's largest high-quality quartz-sand mineral production area), as well as parts of Zhejiang (Shaoxing/Quzhou) and Anhui (Bengbu/Suzhou), forming the country's largest quartz industry cluster centred on Donghai.

The in-production factory distribution in the Research Institute's database strongly confirms the stratified domestic-substitution logic of "bulk/commodities adequate, specialty weak": in fully competitive categories such as silane coupling agents, factory counts are high (130+); in elite categories such as electronic-grade polysilicon, in-production factory counts are extremely small (single digits), with both technical and capital barriers combining to create extremely high entry thresholds.

7.3 Three Core Pathways for Domestic Substitution

Based on systematic analysis of the supply-chain structure and competitive landscape above, the Research Institute has identified three main pathways currently driving domestic-substitution progress in electronic-grade silicon-based chemicals:

Pathway 1: M&A-driven technology and qualification acquisition (the "Yoke model")

Acquiring or taking a controlling stake in overseas companies with advanced electronic-chemicals process capability (exemplified by Yoke Technology's 2017 acquisition of Korea's UP Chemical) to bring in the acquired team's process system, R&D capability, customer relationships, and qualification credentials in one step — effectively bypassing the super-long cycle of "starting from zero R&D + qualification."

Core advantage: extremely low time cost (compared with 10+ years of self-developed technology accumulation, M&A integration typically achieves substantive supply capability in 3–5 years); inheritable customer relationships (existing fab qualifications held by the acquired company can be transferred as part of business assets, subject to re-review and confirmation). Limitation: geopolitical factors mean that target countries (USA/Japan/Netherlands) may impose security-review barriers on cross-border M&A of critical-technology companies; cultural and management integration challenges are significant — if core team members leave during integration, the accumulated process know-how may leave with them; and this M&A window is narrowing under the current geopolitical tensions.

Pathway 2: Joint R&D + domestic-fab co-qualification (the "Shanghai Sinyang / Greentek model")

Companies establish joint-R&D and co-qualification mechanisms with leading domestic fabs (SMIC, CXMT, Hua Hong, etc.), with "vendor doing full-process process verification on the fab's production line" as the core, jointly advancing process adaptation and lot-stability validation for domestic materials.

Core advantage: deep domestic substitution — the qualification ultimately obtained is a genuine qualification for the Chinese fab's specific process conditions, creating the strongest customer stickiness; once stable supply is established, it is very difficult to replace. Also helps build a Chinese domestic electronic-chemicals quality evaluation system, progressively aligning domestic standards with international standards. Limitation: qualification cycles are still long (12–24 months unavoidable); depends on the pace of domestic fab process node advancement (if domestic fabs' advanced-process ramp-up is slow, demand for supporting material qualification is correspondingly delayed); domestic fabs' own advanced-process material qualification experience and evaluation capability are themselves still being built — there is a "joint learning" dimension.

Pathway 3: Incremental technology upgrading from PV/LED mature applications (the "Hoshine/Xingfa model")

In electronics application markets with relatively lower purity requirements (though still well above industrial-grade), such as PV modules, LED packaging, and lithium batteries, domestic companies accumulate cleanroom production management capability, lot stability control experience, and customer quality-qualification reputation as a foundation from which to technically upgrade toward semiconductor-grade purity.

Core advantage: lowest entry cost; initial market scale is large (PV/LED demand is robust), allowing technology experience to be accumulated at volume-production scale, building the foundation for upgrading to higher-end applications. Suitable for organosilicone supply-chain integrated companies (e.g., Hoshine Silicon) to fully leverage their supply-chain-integration advantages. Limitation: the technical upgrade from "PV/LED electronic-grade" to "semiconductor process grade" is still a clear step-change threshold — it cannot be linearly extrapolated; a cognitive pitfall of "meeting PV standards and mistakenly assuming semiconductor standards are achievable" can arise, requiring clear understanding of the essential differences between the two.

7.4 Deep Structural Bottlenecks for Domestic Substitution

At the technical level, the deep factors affecting domestic-substitution progress can be summarised as four structural bottlenecks:

Bottleneck 1: "Pre-investment" nature of analytical capability building

ICP-MS, GC-MS, laser particle counters, and other high-precision analytical instruments cost hundreds of millions of RMB in total; operating and maintenance costs are substantial; and specialist analytical chemistry talent teams are required. These instruments are prerequisites for product R&D and batch quality control — not something that can be acquired after volume supply is secured. In other words, companies must invest tens of millions to over 100 million RMB in analytical capability before securing their first order — a typical "high-risk up-front investment" that constitutes a heavy entry barrier for smaller companies.

Bottleneck 2: "Comprehensive requirement" nature of cleanroom infrastructure

Building cleanrooms (ISO Class 4 and above), ultrapure water systems (resistivity 18.2 MΩ·cm), inert-gas (N₂, Ar) supply networks, and full-PTFE piping systems requires holistic planning and one-time completion — it cannot be built and qualified in stages. This "comprehensive requirement" means that even for well-capitalised companies, cleanroom infrastructure construction takes 2–3 years, further lengthening the timeline from "project approval" to "formal supply."

Bottleneck 3: Incomplete electronic-grade raw-material supply chain

The feedstocks for electronic-grade chemicals (e.g., ultra-pure ethanol, ultra-pure hydrochloric acid, ultra-pure ammonia) must themselves meet electronic-grade specifications, but domestic suppliers of such high-purity raw materials are limited, and some key electronic-grade raw materials still need to be imported from Japan or Germany. This creates a "chokepoint matryoshka doll" within the supply chain: producing electronic-grade TEOS requires ultra-pure ethanol, and some ultra-pure ethanol is also partially import-dependent.

Bottleneck 4: "Qualification barriers" are structural, not temporary

The fab's 12–24-month qualification cycle will not be materially shortened by national policy support or Big Fund investment — this cycle is determined by the objective requirements of process-quality verification; it is a technical assurance mechanism for chip reliability, not an artificially erected barrier. Therefore, any domestic supplier, regardless of technical strength, must complete the qualification pathway step by step with no shortcuts available. This means that even if domestic companies achieve technical breakthroughs in 2026, the earliest they could achieve stable volume commercial supply is around 2028.

7.5 Industry Cluster Perspective: Yangtze Delta, Beijing-Tianjin-Hebei, and Southwest Polysilicon Belt Division of Labour

The industrial geography of China's electronic-grade silicon-based chemicals exhibits a clear pattern of regional specialisation, with different product categories concentrated in different industry clusters — a pattern that profoundly influences supply-chain logistics, regional policy support, and localised service capability.

Yangtze Delta: densest electronic-chemicals region

The Yangtze Delta (Shanghai + Jiangsu + Zhejiang) is China's most important electronic-chemicals industry cluster, with the following core sub-clusters:

Jiangsu Southern (Jiangyin-Changzhou-Suzhou-Nanjing): Jianghua Micro (Jiangyin, wet electronic chemicals), Nantong Chruker (TEOS, Sino-German JV), Suzhou Keji Chemical (various fine chemicals), Thermo Fisher (analytical instruments), and Chinese production bases of foreign companies including Air Liquide and Honeywell — constituting the core electronic-chemicals supplier agglomeration in the northern wing of the Yangtze Delta. Leveraging the excellent business environment of Taicang and Suzhou Industrial Park and their complete chemical-industry ecosystem, this is the most mature domestic electronic-chemicals manufacturing region.

Shanghai: Shanghai Sinyang (Shanghai Free Trade Zone, electronic-chemicals R&D and sales headquarters), Greentek (electronic-grade TMAH developer, primarily Hangzhou but with Shanghai sales team), and multiple foreign-company (BASF, Dow Chemical, Merck, etc.) China headquarters and R&D centres — Shanghai is an R&D and innovation high ground and sales centre for electronic chemicals rather than a purely manufacturing base.

Zhejiang (Tongxiang-Shaoxing-Quzhou): Xinghuo Chemical (organosilicone monomer, Tongxiang), Quzhou Fluorosilicone (organosilicone + fluorochemicals), Juhua Group (fluorochemicals), and Dongyue Silicone (Zibo, Shandong, with close supply-chain links to Zhejiang) — an important production area for organosilicone monomers and specialty silicon chemicals in East China.

South China (Pearl River Delta): display-panel and packaging supporting chemicals cluster

The Shenzhen-Dongguan-Guangzhou industrial belt in Guangdong, with display panels (TCL CSOT, BOE Shenzhen, Visionox) and packaging/testing (ASE, Amkor, etc.) as core demand, is served by large numbers of wet-electronic-chemicals suppliers (etchants, developers) for liquid-crystal display applications, as well as some LED-packaging electronic-grade silicone gel and silicone rubber suppliers (Shin-Etsu's Guangzhou plant is the most important source).

Southwest polysilicon belt (Sichuan-Yunnan-Guizhou)

Leshan/Meishan in Sichuan (Tongwei polysilicon), Mount Emei in Sichuan (Xinlightu Silicon electronic-grade polysilicon pilot base), Qujing in Yunnan (GCL polysilicon), and Weng'an in Guizhou (phosphochemicals/some organosilicone) form China's most important polysilicon production belt. The Southwest region's core advantages are cheap clean hydropower (average industrial electricity price approximately RMB 0.30–0.45/kWh, far below the Eastern region) and abundant water resources (polysilicon production is water-intensive), making Southwest China the energy-cost low-ground of China's polysilicon supply chain.

Northwest energy belt (Xinjiang-Inner Mongolia)

Shihezi in Xinjiang (Daqo Energy, Xinte Energy polysilicon), Urumqi (some industrial silicon), and Baotou/Hohhot in Inner Mongolia (GCL polysilicon, Zhonghuan Semiconductor) form the Northwest polysilicon and industrial-silicon belt, primarily powered by coal (though Inner Mongolia's wind power is also rapidly developing), with low electricity costs but higher carbon footprint — facing energy-structure transition pressure as global green-supply-chain requirements tighten.

Northeast organosilicone (Jihua-Daqing)

Jilin Chemical (Jihua, under PetroChina) and Daqing organosilicone production, primarily trichlorosilane (TCS) and certain organosilicon chlorides — the historical origin of China's organosilicone monomer industry, though in recent years cost competitiveness has weakened relative to the Northwest and Southwest, and market position has declined; primarily serving local and regional markets and exporting raw organosilicone materials.

The strategic implication of this regional division of labour is that different regional industry clusters correspond to different domestic-substitution breakthrough pathways and policy support priorities. Understanding this geographic landscape helps in assessing the policy-orientation of different provinces (Sichuan focused on electronic-grade polysilicon, Jiangsu focused on fine-chemical electronic chemicals, Guangdong focused on display and packaging chemicals) and the priority investment regions for future industrial capital.

Chapter 8　Pricing Tiers and Business Models: Tonne-Scale vs. High-Purity 11N Kilogram-Scale

8.1 Price Ladders and Quality-Premium System

The pricing structure of electronic-grade silicon-based chemicals exhibits an extremely clear "quality-ladder premium" configuration. For the same chemical compound, from industrial-grade to the highest-specification semiconductor-grade, the price gap often reaches 5–10 times, and the origin of this premium is precisely the technical barriers and qualification moats.

Polysilicon price ladder (2025–2026 actual market data)

Specification	Purity	Price range (RMB/kg)	Notes
PV-grade (N-type mono)	6N–8N	45–70	2025–2026 historical low
Semiconductor-grade (9N–10N)	9N–10N	150–200	Scarce domestically, very limited volume
Semiconductor-grade (11N EG)	11N	200–280	Mainly import-dependent domestically
Top-grade EG (Wacker/Hemlock)	11N+	280–350	Premium for most authoritative qualification brand

Note: In March 2026, the international price of Chinese PV-grade polysilicon (mono-premium) was approximately USD 6.47/kg (~RMB 47/kg) — a near-historical low. The absolute price differential between electronic-grade and PV-grade has therefore widened further; the electronic/PV price ratio exceeded 4–5x in 2026.

TEOS price ladder

Specification	Metallic-impurity requirement	Price range (RMB/kg)	Major suppliers
Industrial-grade	ppm level	30–50	Multiple domestic companies
LCD/panel-grade	sub-ppm level	60–90	Domestic + South Korean
Electronic-grade (mature process)	ppb level	80–120	Yoke (domestic) / imports
Advanced-process (7 nm and below)	ppt level	150–300	Import-dominated (Entegris/Wacker)

HMDS price ladder

Specification	Purity/impurity	Price range (RMB/kg)	Major suppliers
Industrial/agrochemical-grade	≥98%	30–60	Domestic
Reagent-grade	≥99.9%	80–120	Domestic + imports
Electronic-grade (LCD)	≥99.99%, metals at ppb	120–200	Some domestic
Semiconductor-grade (advanced process)	≥99.999%, metals sub-ppb	200–400	Merck/Shin-Etsu dominated

Silane coupling agent price range (KH-560 example)

Industrial bulk grade (PV/construction-use): approximately RMB 20–35/kg
Electronic/optical specialty grade: approximately RMB 60–150/kg

The 4–5x price gap between PV-grade (bulk) and semiconductor-grade (specialty) silane coupling agents reflects the difference in functional synthesis difficulty.

8.2 Bulk-Commodity Model vs. Technical-Service Supply Model

The electronic-grade silicon-based chemicals industry contains two fundamentally distinct business models; understanding this distinction is critical for predicting profit allocation and competitive intensity at each supply-chain segment.

Bulk-commodity model (PV polysilicon)

PV polysilicon exhibits fully bulk-commodity characteristics: prices are highly transparent, published weekly by industry associations and data providers (InfoLink, PV InfoLink, CPIA); customers (module manufacturers) procure through annual/multi-year long-term contracts and spot markets flexibly; the competitive dimension among suppliers is primarily cost (who has the lowest cost) and scale (who can guarantee stable large-volume supply); prices are highly sensitive to supply-demand balance — during industry overcapacity, prices rapidly fall below cost, triggering widespread losses.

Tongwei's long-term supply contracts with major downstream customers such as LONGi Solar and JA Solar are a textbook example of this bulk-commodity supply relationship — Tongwei locks in the largest downstream buyers with extremely low cost and stable quality, forming a "volume and price dual-assurance" mechanism. Technical-Service Supply Model (Electronic-Grade Polysilicon / TEOS / HMDS)

Semiconductor-grade chemicals exhibit all the hallmarks of a technical-service product: prices are non-transparent (negotiated bilaterally between supplier and customer, no public quotes), reviewed annually, and relatively stable once set. The customer's (wafer fab's) primary concern is not price but rather: "Can this supplier consistently deliver, batch after batch, a product that meets my exact process requirements?" Once a supplier passes qualification, the supply relationship becomes extremely stable — switching costs are high and fabs are extremely reluctant to replace a qualified supplier. Even when a domestic product is 20–30% cheaper than an imported alternative, a fab will not readily switch, because the risk of switching far outweighs the cost savings.

The defining characteristic of this "technical-service" business model is: high gross margins (differentiation premium created by the qualification barrier), but a very high entry threshold (long qualification cycles, exacting technical requirements), and once entry is achieved, extremely high customer stickiness.

Implications for Capital Allocation

Bulk-commodity model (PV polysilicon): currently at a cyclical trough — focus on tracking the industry consolidation process and margin recovery at leading companies rather than chasing short-term profit. Electronic-grade chemicals (TEOS/HMDS/semiconductor-grade polysilicon): supply-demand dynamics are far healthier than the PV-grade segment, with no overcapacity pressure. Companies that have successfully passed wafer-fab qualification are high-value investment targets; the critical variable is "who completes qualification first."

8.3 Cost Structure and Profitability Analysis

Cost Structure of Electronic-Grade Polysilicon (11N)

Electronic-grade polysilicon carries several additional cost items relative to PV-grade polysilicon:

The TCS feedstock must undergo extra distillation to remove electrically active impurities (B, P, etc.), adding approximately RMB 5–15/kg
Float-zone (FZ) purification steps (where required) add power and equipment depreciation costs of approximately RMB 20–50/kg
Analytical testing costs (ICP-MS etc.): thousands to tens of thousands of RMB per batch
Qualification-maintenance costs (ongoing technical communication with wafer fabs, sample shipments, periodic audits): fixed annual cost of several million RMB
Off-spec loss rate: roughly 10–20% of production volume fails to meet the highest specification and is downgraded to PV-grade, representing a downgrade loss of approximately RMB 200/kg

Combining these additional costs, a domestic company aiming to produce 11N electronic-grade polysilicon faces an estimated all-in cost of approximately RMB 120–200/kg (on top of a PV-grade all-in cost of approximately RMB 45–55/kg). Current international market procurement prices for electronic-grade polysilicon stand at approximately RMB 200–280/kg, which means profitable margins of approximately 20–50% gross margin are achievable — provided the company can secure stable customer qualification and achieve minimum economic-scale output of approximately 1,000 MT/year.

8.4 Long-Term Supply Agreement Mechanics and the Procurement Ecosystem

Typical LTSA Structure in Electronic Chemicals Supply

In semiconductor-grade chemical supply chains, the Long-Term Supply Agreement (LTSA) is the legal vehicle binding wafer fabs and core suppliers, and the formal mechanism that locks in their technical-collaboration relationship. Unlike the annual pricing contracts typical of bulk commodities (PV polysilicon), electronic-chemical LTSAs characteristically contain several unique clauses:

Technical Specification and Change Notification Process (CNP): Contracts specify detailed technical specifications (metal impurity ceilings, organic impurity limits, particle counts, moisture levels, etc.) and require that any changes — including changes to raw-material sources, production processes, or production facilities — be notified to the fab 90–180 days in advance and subjected to re-evaluation and approval. This clause means the supplier is subject to a "process lock-in constraint" during the contract period: production methods cannot be freely altered, and operational flexibility is curtailed.

Take-or-Pay Clause: The wafer fab commits to purchase a minimum volume during the contract period (e.g. not less than XXX MT per year); even if actual demand falls below that level, it must pay the equivalent or a liquidated-damages penalty. This clause gives the supplier a guaranteed revenue base that makes it willing to invest in dedicated capacity (such as a purpose-built production line) for a specific customer.

Supplier Scorecard Mechanism: Quarterly or semi-annually, the fab's procurement department evaluates the supplier on multiple dimensions — quality, on-time delivery rate, technical-response speed, price competitiveness, etc. Scorecard results directly determine the supplier's volume share in the next period. Suppliers must assign dedicated customer-service engineers to maintain regular contact with the fab's process engineers and rapidly resolve incoming-material anomalies (Lot Hold events).

IP and Confidentiality Clauses: Technical outcomes potentially generated jointly during the qualification-cooperation process (process-window parameters, improved formulations, etc.) must have ownership assigned in advance. The fab strictly protects formulation and process information provided by the supplier; the supplier equally protects process-requirements information provided by the fab — preventing competitive information leakage in both directions.

Single-Source Risk Management Clause: For product categories with low domestic-substitution rates (e.g. high-purity HMDS), where the fab's primary source is a single foreign supplier, procurement strategy typically requires simultaneously advancing qualification of a second (domestic alternative) supplier to reduce "supply-disruption risk." This mechanism is a key policy-driven entry point for domestic suppliers seeking to enter the wafer-fab supply system.

Market-Entry Strategy Insights for Domestic Suppliers

Based on the LTSA mechanics above, domestic electronic-chemical companies need to understand the following key strategic dynamics when developing market-entry strategies:

First, "qualification approval" and "procurement volume share" are two separate milestones. Entering the Qualified Vendor List (QVL) does not mean large orders follow immediately. Typically only a 5–10% initial share is granted (as a "backup supplier"), and the company must accumulate a track record of high-quality supply over time before progressively increasing its share.

Second, "supply-chain security" has become the most significant new scoring weight since 2022. Fab procurement decisions have added a "supplier geographic-risk diversification" dimension — deliberately avoiding excessive concentration in any single country (e.g. Japan or the USA) — which objectively creates incremental share opportunities for China-based domestic suppliers.

Third, domestic suppliers that can offer "co-development" value (willingness to work jointly with the fab on next-generation process formulations, provide process engineering support) will score significantly higher on the fab's scorecard than a pure product-supply model. This is a potential differentiation strategy for domestic precursor companies (Yoke Technology, Shanghai Xinyang, etc.) relative to foreign bulk suppliers.

8.5 Global Electronic-Chemical Trade Flows and China's Import-Export Structure

Current Import Dependence

In 2025, China's electronic-chemicals imports in the high-end semiconductor-grade segment are estimated at approximately RMB 6–9 billion, with the following primary sources:

Japan: the largest import source, supplying high-purity TEOS (Shin-Etsu, Wacker's Japan plants), semiconductor-grade polysilicon (Tokuyama), and photolithography materials (JSR, Shin-Etsu photoresists, etc.)
Germany: predominantly Wacker's semiconductor-grade polysilicon, Merck/Versum's HMDS and low-k dielectric precursors, and Evonik's TEOS
USA: primarily Hemlock's semiconductor-grade polysilicon and Entegris's TEOS and HMDS — the country most exposed to potential US export-control impact
South Korea: primarily Soulbrain's TEOS and OCI Materials' precursors, with relatively strong local-service capabilities in the Chinese market

Export Opportunities

China holds significant export-competitive advantages in PV-grade polysilicon, silane coupling agents (standard grades), and organosilicone monomers (D4/D5):

PV polysilicon: China exported approximately 85,000 MT of polysilicon in 2023, primarily to Europe, India, and Southeast Asia (Vietnam, Malaysia, and other PV-module manufacturing countries)
Standard-grade silane coupling agents: exports of bulk organosilicone coupling agents (KH-550/560/570 etc.) to Japan, South Korea, and Europe represent a textbook example of "Made in China" competing on cost-performance in fine chemicals
Organosilicone intermediates (D4/D5): large-volume exports to Europe and North America, for use in local daily-chemical products (shampoos, skincare) and industrial silicone-rubber production. However, as noted earlier, EU SVHC restrictions on D4 are affecting European daily-chemical demand, and future export structure may increasingly concentrate on industrial applications.

Chapter 9　Customer Case Studies

9.1 Case One: Greentek — A Domestic-Substitution Blueprint Starting from Developer Fluid

Greentek (Hangzhou), one of China's representative companies in electronic-chemical domestic substitution, provides a high-value reference pathway through its track record in lithography developer-fluid substitution.

Greentek's G5-grade TMAH (tetramethylammonium hydroxide) developer fluid initially entered the supply system at Hua Hong Semiconductor's power-device lines, passing approximately 18 months of process-adaptation verification before breaking into Hua Hong's qualified supplier list. The company subsequently extended into SMIC's 14 nm/28 nm logic nodes, and is currently in the ramp-up stage of small-volume supply to CXMT's DRAM lines as well as completing final verification for Yangtze Memory Technology's 3D NAND lines.

The core lesson from the Greentek case is its clear "three-step" strategy: Step One — enter through process lines with relatively less demanding specifications (Hua Hong power devices), earning the first wafer-fab qualification certificate and building a baseline quality reputation; Step Two — upgrade to mainstream logic processes (SMIC 14 nm/28 nm), establishing a stable foothold in the mainstream mature-process market; Step Three — penetrate memory processes (CXMT DRAM / Yangtze Memory NAND), entering fabs with higher demand volumes and a stronger domestic-substitution preference.

This "staircase penetration" path is the most instructive Chinese reference model for domestic-substitution strategies for TEOS, HMDS, and other precursor suppliers. The domestic-substitution push for TEOS and HMDS will most likely follow a similar sequence: specialty processes → mature logic → memory → advanced logic — progressing step by step rather than leapfrogging directly to the most advanced nodes.

9.2 Case Two: Yoke Technology and SK Hynix — A Cross-Border Qualification Headstart

Yoke Technology's deep cooperative relationship with South Korea's SK Hynix is a textbook example of a Chinese precursor supplier entering the supply chain of a top-tier global customer through cross-border M&A — and is currently one of the furthest-advanced cases of domestic electronic-chemical companies achieving qualification at the world's most demanding customer tier.

SK Hynix is a core HBM (High Bandwidth Memory) supplier globally. Against the backdrop of the AI large-model boom (NVIDIA H100/H200/B200 GPUs paired with HBM2e/HBM3/HBM3E), SK Hynix's HBM shipments have grown rapidly, driving parallel high growth in ALD/CVD precursor procurement volumes. Yoke Technology (through its UP Chemical subsidiary) supplies silicon-containing ALD precursors and other thin-film deposition materials to SK Hynix, occupying a position within SK Hynix's core supply system.

Yoke is further developing High-k metal-oxide precursors (for advanced-node gate High-k dielectric layers using ALD hafnium oxide, zirconium oxide, etc.) and ultra-high/low-temperature silicon-type CVD precursors for different temperature-window processes, with some products already in customer-side process testing — demonstrating Yoke's sustained deepening of its precursor technology roadmap.

From a secondary-market perspective, Yoke Technology's valuation is driven by the dual narratives of AI-industry demand (HBM precursor growth driven by GPU volume expansion) and domestic substitution (TEOS and other import-substitution progress) — making it one of the A-share semiconductor-materials stocks with the clearest industrial logic at present.

9.3 Case Three: Tongwei × LONGi — The Win-Win and Structural Limits of the Bulk Polysilicon LTSA

The long-term polysilicon supply agreement between Tongwei and LONGi Green Energy (the world's largest mono-crystalline silicon wafer and PV module company) is a landmark example of vertical integration in China's PV supply chain — and simultaneously reveals the deep structural limitations in migrating the PV bulk model to electronic-grade applications.

In the PV bulk segment, Tongwei builds a "volume and price dual-assurance" cooperation framework with LONGi on the basis of the lowest all-in production cost in the industry and the largest production scale (annual capacity exceeding 900,000 MT), maintaining high capacity utilisation even during industry downturns while giving LONGi stable low-cost raw-material supply. This model perfectly fits the "cost + scale + volume lock-in" logic of bulk-commodity markets.

However, the logic of the electronic-grade polysilicon segment is completely different. Wafer-fab customers (SMIC, TSMC, etc.) are far more conservative in their procurement decisions than PV module manufacturers; supplier qualification cycles (approximately 12–24 months) and technical requirements (11N purity plus a complete quality system) vastly exceed those of PV customers. Even with a clear price advantage, a fab will not accept a new supplier that has not passed full qualification. More importantly, global electronic-grade polysilicon annual demand is approximately 30,000–40,000 MT — compared to the PV-grade market's multi-hundred-thousand-MT scale — which means that even if Tongwei achieves an electronic-grade breakthrough, it will not fundamentally reshape its overall business structure.

Tongwei's pathway to electronic-grade polysilicon therefore requires not a direct extension of its PV bulk model, but the construction of an entirely new "technical-qualification-type" business system: dedicated electronic-grade production lines physically isolated from PV-grade lines, independent analytical testing capabilities, dedicated electronic-grade sales and technical-service teams, and the patience to absorb long qualification-cycle investment. This is a new commercial capability that must be built from scratch, not a simple extension of existing PV bulk-production competencies.

9.4 Case Four: Xinlightu Silicon — Polysilicon EG Domestic Substitution "From Zero to One"

Sichuan Xinlightu Silicon (a joint venture of Tongwei, Xinte Energy, Emeishan State-Owned Asset Operating Company, and others) is currently the fastest-moving domestic electronic-grade polysilicon project, carrying the strategic mission of achieving China's "from zero to one" self-supply capability in the 11N ultra-high-purity polysilicon segment.

Xinlightu Silicon's 1,000 MT/year semiconductor-grade polysilicon project, after a long period of process research and equipment commissioning, produced its first batch of samples in May 2024 — marking a preliminary breakthrough in the domestic process flow for electronic-grade polysilicon. However, the distance from "producing samples" to "stable large-volume commercial supply" involves a long qualification journey still to be completed. The project currently remains in the stages of product-quality ramp-up (improving batch consistency, gradually increasing the 11N-grade coverage rate from its current sub-50% level) and conducting sample evaluations with potential customers (domestic and international wafer fabs). Achieving formal entry into a wafer fab's Qualified Vendor List and stable volume supply is estimated to require a further 18–24 months.

Based on analysis by the Research Institute, the customers where Xinlightu Silicon (and future domestic entrants) is most likely to first achieve volume-supply breakthrough are domestic DRAM and NAND Flash fabs (CXMT and Yangtze Memory Technology), rather than advanced-process international fabs such as TSMC and Samsung — the latter have longer qualification-cycle requirements (typically 36+ months), stricter quality standards, and represent a market target achievable only at a later stage for domestic companies.

9.5 Case Five: Hoshine Silicon's Push Up the Value Chain into Electronic-Grade Organosilicone

Hoshine Silicon's strategic exploration — using its unmatched organosilicone monomer production scale (173,000 MT/year D4/D5 cyclics and related products) as the base from which to extend into higher-value electronic-grade downstream products — is pursuing the development of low-impurity, high-purity silicone gel products targeting semiconductor packaging and LED applications.

Compared to directly targeting 11N electronic-grade polysilicon or advanced-process TEOS (both categories with extremely high technical barriers), the technical threshold for electronic-grade organosilicone gels (used in power semiconductor module encapsulation) is relatively lower. The core specifications centre on ionic impurities (Na⁺, K⁺, Cl⁻ etc., typically required at < 10 ppm), mechanical properties (post-cure hardness/elasticity), and high/low-temperature stability. These requirements are orders of magnitude more accessible in engineering terms than the ppt-level trace-metal control needed for advanced-process TEOS — making this the shortest-path route by which Hoshine can leverage its existing organosilicone platform to reach "high-end extension."

In 2025, Hoshine Silicon successfully industrialised amino silicone oils and silicone emulsions, achieving product quality at internationally leading levels. In 2026, the company plans to introduce certified samples of electronic-grade silicone gels targeting domestic automotive power-module packaging companies (BYD Semiconductor, Starpower, CRRC Times Electric, etc.) for first-customer qualification breakthroughs. If successful, Hoshine will become a platform-type silicon chemicals enterprise with full-chain coverage from industrial silicon → D4 → electronic-grade organosilicone deep-processing, and its valuation narrative will shift from "cyclical organosilicone play" toward "electronic-chemical growth company."

9.6 Case Six: China's High-Purity Quartz Breakthrough — The Lianyungang–Donghai Cluster

High-purity quartz sand may appear to be the most "upstream" mineral raw material, but it is in fact one of the hidden critical chokepoints constraining China's semiconductor supply-chain autonomy. Understanding this category's domestic-substitution pathway is essential for grasping the overall strategic picture of China's electronic-grade silicon-based chemical industry.

Extreme Concentration in Global Supply

The Spruce Pine mining district in North Carolina, USA, is universally recognised as the world's highest-quality high-purity quartz source, controlled by Belgium's Sibelco group and US-based Covia (formerly Unimin). Spruce Pine quartz is celebrated for its "exceptional chemical purity" — natural SiO₂ content exceeding 99.99%, complete crystal structure (rock crystal / vein quartz morphology), and naturally low metallic impurity levels (Fe, Al, Ti, etc.) — which after beneficiation can directly meet the most stringent semiconductor-crucible grade requirements.

In September 2022, a hurricane forced a brief shutdown (a few weeks) at one Spruce Pine facility. The announcement immediately put the entire semiconductor-materials industry on high alert, and prices for high-purity quartz crucibles surged — demonstrating the strategic influence of this single mining district on the global semiconductor supply chain. Notably, even though Spruce Pine production was not fully halted, the mere threat of a shutdown was sufficient to trigger severe market turbulence, reflecting the true fragility of the global high-purity quartz supply chain.

China's Lianyungang: The Closest Domestic Contender

Jiangsu Province's Donghai County in Lianyungang City is renowned as "China's Crystal Capital," the largest natural crystal and premium quartz mineral region in the country, supplying large volumes of high-purity quartz sand for domestic optical fibre and solar-PV crucibles for many years. Donghai quartz sand, characterised by uniform particle size, high natural SiO₂ content (≥ 99.9%), and low metallic impurities, possesses the mineralogical foundation for processing into semiconductor-grade high-purity quartz sand.

However, there remains a gap between "having the foundation" and "meeting the highest standards." Semiconductor-grade quartz crucibles require quartz sand with SiO₂ ≥ 99.99% (99.9999% at the highest grade) and total metallic impurities < 10 ppm, including Fe < 1 ppm, Al < 3 ppm, Ti < 0.5 ppm. The proportion of Donghai quartz ore that naturally meets 99.99% purity is limited, and the share meeting the highest semiconductor standard is smaller still.

Currently, representative Chinese high-purity quartz companies include: Silica Stock (603688.SH, Suzhou), which holds multiple mining rights in Donghai and is the leading domestic company in high-purity quartz beneficiation and quartz components; Lianyungang Haoyu Quartz, Jiangsu Zhongsheng Silicon Star Technology, and others continuously advancing high-purity quartz beneficiation and purification R&D toward semiconductor-crucible-grade standards; and Yantai Skraa, which has made breakthroughs in semiconductor-grade quartz components.

Core Technical Challenges in Beneficiation

The core technical pathway for beneficiating natural quartz sand to semiconductor grade includes:

Physical separation: using gravity separation (density-liquid separation), magnetic separation (high-gradient magnetic separators to remove iron-bearing minerals), and flotation (to remove feldspar, mica, and other aluminium-bearing minerals) to remove non-quartz minerals from run-of-mine ore, raising SiO₂ purity above 99.9%;

Acid leaching: immersing physically separated quartz sand in high-purity acid solutions (HCl + HF mixed acid, or high-purity HNO₃) to dissolve and remove metallic impurities (Fe, Al, Ca, Mg, etc.) remaining at crystal-grain surfaces and micro-fractures; the purity of the acid used is critical (the acid itself must not introduce new metallic contamination) — electronic-grade high-purity acids are required;

High-temperature chlorination refining (key advanced step): processed quartz sand is exposed to high-purity HCl gas at high temperature (~1,100°C) to form volatile metal chlorides that are driven off, removing residual metallic impurities. This step is the core technology for refining quartz sand to ppb-level ultra-high purity meeting the highest semiconductor-crucible standards. Domestic companies remain significantly behind international leaders such as Germany's Heraeus and US-based Momentive (formerly GE Quartz) in this process step;

Ultra-pure-water washing and drying: the finished product must be washed multiple times with 18.2 MΩ·cm resistivity ultra-pure water to remove residual acid and ions, then low-temperature dried in a clean room before packaging.

As domestic semiconductor wafer fabs rapidly expand (pulling single-crystal silicon ingots requires large quantities of high-purity quartz crucibles — each 300 mm single-crystal ingot consumes 1–2 crucibles), annual domestic demand for high-purity quartz sand is growing rapidly, driving companies such as Silica Stock to significantly expand beneficiation capacity and progressively move toward the highest semiconductor-crucible specifications.

Outlook: By 2027–2028, domestic Lianyungang high-purity quartz sand crucible-grade refining technology is expected to meet 70–80% of domestic wafer-fab (CZ crystal-pulling) demand, with the remainder still imported from Spruce Pine (USA) or Norway. Complete domestic self-supply of ultra-high-purity semiconductor-crucible-grade quartz (ppm/ppb-level metallic impurities) is unlikely to be achieved before 2030.

9.7 Case Seven: Xingfa Group — Platform Synergy from Phospho-Chemistry to Electronic-Grade Silicon Chemicals

Xingfa Group's strategy exemplifies the model of leveraging local resource endowments to build a multi-category synergistic chemical platform. Xingfa has combined Yichang's rich phosphate reserves (approximately 6 billion MT, one of China's most important phosphate production regions) and quartz resources with the low-cost hydropower advantages of the Three Gorges grid to form an integrated "phospho-chemistry + organosilicone + electronic chemicals" platform.

In its organosilicone segment, Xingfa Organic Silicon (with Yichang Xingshan and Yidu as core production bases) produces approximately 500,000 MT/year of organosilicone monomers (predominantly M₂), making it one of China's important organosilicone monomer suppliers. The product portfolio covers: industrial-grade silane coupling agents (KH-550/560/570 standard grades), various silicone oils/silicone rubbers/silicone gels, and electronic-chemical-supporting organosilicone materials.

In its electronic-grade fluoride segment, Xingfa's electronic-chemicals subsidiary produces electronic-grade hydrofluoric acid (UPW-grade, SEMI Grade 1–4) and ammonium fluoride (NH₄F), and is a key domestic supplier of raw materials for semiconductor wet-etching chemicals. Electronic-grade HF is essential for every wafer fab for SiO₂ film etching (HF dilute in RCA cleaning, BHF buffered oxide etch, etc.).

Xingfa Group's domestic-substitution strategy for silicon-based chemicals benefits from several distinctive advantages: raw-material synergies between phospho-chemistry and organosilicone; electricity-cost advantages (Three Gorges grid and local hydropower, yielding electricity costs below those of East-China peers); and regional policy support (Hubei Province has clear policy direction favouring Yichang chemical-company transformation, benefiting Xingfa in capital and land resources).

Xingfa's next strategic priorities are expected to centre on: upgrading to high-purity silane coupling agents (semiconductor-packaging specialty grades), advancing electronic-grade HF purification breakthroughs toward Grade 5 (highest purity), and building localised electronic-chemical supply relationships with nearby customers — particularly Yangtze Memory Technology's Wuhan base.

9.8 Case Eight: OLED and Flexible Electronics Packaging — Organosilicone Thin Films as a New Frontier

Beyond traditional semiconductor wafer manufacturing, organosilicone-based electronic chemicals are opening entirely new market territory in flexible display and OLED panel encapsulation. This direction differs significantly from wafer-level electronic chemicals in its technical pathway, but imposes equally stringent requirements on material purity, thin-film uniformity, and long-term reliability — a dimension essential for a complete picture of China's silicon-based electronic-chemical landscape.

Material Challenges in OLED Encapsulation

Organic light-emitting diode (OLED) core materials (emitting-layer organic small molecules/polymers) are extremely sensitive to moisture and oxygen: even water-vapour transmission of just 10⁻⁶ g/(cm²·day) is sufficient to cause "dark spots" (pixel failure) and brightness degradation in OLED devices within hundreds of hours. OLED panel encapsulation (barrier-layer design) is therefore the critical engineering challenge determining device lifetime, with extremely demanding requirements on thin-film encapsulation material water-vapour transmission rate (WVTR) — flexible OLED requiring WVTR < 10⁻⁶ g/m²/day, far beyond the barrier capability of ordinary thin-film materials.

In rigid OLED panels (smartphone screens, high-end TV displays), traditional encapsulation uses a "glass cover + sealant" approach; sealant materials are typically epoxy-resin-based or UV-cure organosilicone adhesives. Organosilicone encapsulant (e.g. Dow Corning/Momentive's optically transparent SYLGARD® series) is the preferred material for high-end OLED panel-edge encapsulation, due to its high light transmittance, wide-temperature-range applicability (-40°C to +200°C), and chemical stability.

For flexible OLED (fOLED) encapsulation, Thin Film Encapsulation (TFE) has become the industry-standard solution, using alternating deposition of inorganic barrier layers (SiNₓ, SiO₂, Al₂O₃) and organic buffer layers (acrylic polymer) in an "organic-inorganic multilayer" structure (typically 3–7 bilayers), achieving overall WVTR below 10⁻⁶ g/m²/day while maintaining the flexibility to withstand bending without cracking.

The Central Role of PECVD Silicon-Based Thin Films in TFE

The inorganic barrier layers in TFE (SiNₓ, SiO₂) are deposited by plasma-enhanced chemical vapour deposition (PECVD), and the core silicon source in PECVD precursors is exactly this report's subject of study — TEOS and silane (SiH₄). In TFE processes, TEOS as the silicon source for SiO₂ barrier-layer deposition offers: lower deposition temperatures (typically 100–150°C, compatible with flexible-substrate thermal-budget constraints), good film density, and excellent step coverage; SiNₓ barrier layers are typically deposited from SiH₄ and NH₃ precursors, covered by the electronic-specialty-gas segment's nitride PECVD process.

The rapid expansion of OLED panel TFE production therefore directly drives parallel growth in TEOS and silane-type precursor consumption. Global OLED panel capacity is expanding rapidly (Samsung/LG Display's premium panels plus BOE/CSOT/Visionox's fast-growing domestic OLED push), creating structural demand growth for related PECVD precursors.

Domestic-Substitution Status for China's OLED Encapsulation Materials

Compared with wafer-manufacturing-level electronic chemicals, the domestic-substitution progress in OLED encapsulation materials is relatively more advanced:

Panel-grade TEOS (for LCD/OLED use, not wafer-level semiconductor-grade TEOS; purity requirements are lower but specific metallic-impurity limits still apply) has seen substantial domestic uptake, with Yoke Technology and Nantong Sogeler among companies maintaining stable supply — domestic substitution estimated at over 70%;

For organosilicone encapsulant (UV-cure/thermally cured, for panel edge sealing), Chinese organosilicone companies (Hoshine, Xinaan Chemical, locally produced Dow products) can technically satisfy low-to-mid-end panel (LCD/ordinary phone OLED) requirements; however, top-end smartphone OLED encapsulants (requiring high refractive index, ultra-low impurity content, ultra-long service life) are still predominantly sourced from Shin-Etsu and Dow Corning international brands;

ALD aluminium oxide (Al₂O₃) barrier layers rely on trimethylaluminium (TMA) precursor, and a handful of domestic companies (including certain business lines at Huate Gas) can supply TMA, though the quality gap relative to international leaders (BASF, Entegris/Sigma-Aldrich) remains notable, with purity pass rates and batch stability still in need of improvement.

The Next Wave: Mini-LED / Micro-LED Opportunities

After OLED, Mini-LED backlighting and Micro-LED self-luminous display technologies represent the next wave of display industry advancement, with Apple, Samsung, and Sony all introducing Mini-LED backlit products in flagship devices (Pro iPads, high-end TVs, professional displays). Micro-LED volume production is expected to reach scale between 2027 and 2030.

For silicon-based electronic chemicals, Micro-LED packaging similarly demands high-performance organosilicone encapsulation materials, with even stricter requirements for light transmittance (> 95%) and low birefringence. Additionally, the temporary-bonding adhesives and release materials used in Micro-LED chip mass-transfer processes have organosilicone-based materials (e.g. Nitto Denko's thermally releasable tape) as a mainstream technical option, creating emerging demand for specialty functional organosilicone materials. Domestic companies (Yantai Debang Technology, Shengyi Technology, etc.) have begun developing domestic alternatives for temporary-bonding adhesives, though significant progress before entering the Micro-LED supply chain at volume will require further development time.

Chapter 10　Investment, Financing and M&A

10.1 Primary Market: Fundraising Intensity in the Semiconductor-Materials Track

Since 2022, as the semiconductor domestic-substitution investment theme has taken root, primary-market financing activity in electronic-grade silicon-based chemicals has remained consistently active, concentrated in the following funding directions:

High-purity precursor / TEOS: Multiple early-stage companies developing high-purity TEOS, HMDS, and ALD precursors have successively closed angel-through-Pre-A/A-round financing, with individual-round sizes typically RMB 50 million–200 million. Representative investors include: SMIC-Junyuan (SMIC's vehicle specialising in semiconductor materials and equipment), WI Harper (semiconductor-focused VC), Legend Capital, and SAIF Partners. Because the precursor track carries extremely high technical barriers and extremely long qualification cycles, investors apply a "scientific technology, long-cycle" investment framework at the project-selection stage — emphasising research-team technical credentials (whether lab-bench results or process development experience exist) rather than near-term revenue, which is typically minimal.

Electronic-grade polysilicon: Projects such as Xinlightu Silicon have received support from local government industrial-guidance funds, with funding sizes significantly larger (estimated at RMB 1 billion-plus), reflecting the direct capital-commitment of local government to strategic materials domestic substitution. Structurally, local state-owned capital and industrial capital are the primary investors in electronic-grade polysilicon projects; pure market-oriented VC participation is limited, as investment payback periods exceed 10 years and commercial uncertainty is high.

High-purity quartz sand: Quartz beneficiation companies in Donghai County, Lianyungang, Jiangsu, have benefited directly from the semiconductor-materials investment wave. Listed companies such as Silica Stock (603688.SH) saw extremely high A-share market enthusiasm during 2023–2024. Other high-quality unlisted companies such as Pacific Quartz have also closed multiple rounds of strategic financing, building valuation foundations for future IPOs or acquisition transactions.

10.2 Big Fund Phase III: A Historic Intervention

The establishment of the National Integrated Circuit Industry Investment Fund Phase III (Big Fund III, officially registered in June 2024, registered capital RMB 344 billion) is the largest national industrial fund in China's semiconductor industry history. Its significance lies not only in the capital scale but in a structural shift in investment priorities:

Investment focus shifting from "chip design + manufacturing" toward "materials + equipment + components": Big Fund I (RMB 98.7 billion, established 2014) and Big Fund II (RMB 204.2 billion, established 2019) primarily invested in chip design (Huawei HiSilicon, Cambricon, SMIC, etc.) and wafer manufacturing. Big Fund III continues manufacturing investment while significantly increasing allocations to semiconductor materials, semiconductor equipment, and key components (high-precision parts, vacuum systems, etc.) — precisely because these are the "chokepoint" segments with the lowest current domestic-substitution rates and highest import dependence.

Specifically for electronic chemicals, Big Fund III's potential investment pathways include:

Direct equity stakes: through secondary-market purchases or targeted share placement in listed electronic-chemical companies such as Yoke Technology (002409.SZ), Shanghai Xinyang (300236.SZ), and Jianghua Micro (603078.SH)
Sub-fund investment: through regional Big Fund vehicles (Shanghai, Beijing, Shenzhen, Hefei, etc.) providing early growth capital to R&D-stage electronic-chemical companies
Major project direct investment: providing co-investment or credit endorsement for strategically significant domestic-substitution projects (such as Xinlightu Silicon's electronic-grade polysilicon project)

Market estimates suggest Big Fund III allocates approximately 15–25% of total capital to materials and components, implying approximately RMB 50–86 billion directed to the materials/components segment over the fund's investment horizon — making it the single most important policy-capital force driving accelerated electronic-chemical domestic substitution over the next 3–5 years.

10.3 M&A: Two-Way "Technology Acquisition" Strategies

Chinese Outbound M&A

As noted earlier, outbound M&A is the fastest route for Chinese companies to acquire advanced electronic-chemical technology — Yoke Technology's acquisition of UP Chemical (2017) is the most successful textbook example. Future potential outbound targets may concentrate on:

Mid-sized Japanese/Korean TEOS, HMDS, and ALD precursor companies: these companies often possess decades of process know-how and core patents, but face intense competition from larger peers and trade at relatively low independent valuations — creating M&A willingness. In periods of lower geopolitical pressure, M&A windows for Japanese and Korean companies may open somewhat in 2026–2028.
European fine-silicon-chemical companies: high energy costs and rising manufacturing costs across Europe are pressuring profitability for mid-sized European silicon-chemical companies, increasing their willingness to sell — potentially creating technology-transfer opportunities for Chinese acquirers.

Domestic Industry Consolidation

Domestic electronic-chemical sector consolidation is still at an early stage, but as the sector heats up and competitive dynamics clarify, meaningful consolidation is expected in 2026–2028:

Leading companies (Yoke Technology, Shanghai Xinyang) acquiring early-stage companies with specific technical capabilities to rapidly fill product-portfolio gaps
Integrated chemical groups (Hoshine, Xingfa Group) taking strategic stakes in R&D-stage electronic-chemical companies to accelerate high-end product-line extension

10.4 Secondary Market and Valuation Logic

In the A-share market, the electronic-chemical segment is an important sub-sector within the broader semiconductor-materials theme:

Yoke Technology (002409.SZ): The market assigns high semiconductor-materials valuation premiums (P/E typically 30–50×), with the core thesis being AI-driven HBM precursor demand growth and TEOS/ALD precursor domestic-substitution progress.

Shanghai Xinyang (300236.SZ): Net profit grew 71.12% year-on-year in 2025; this exceptional earnings elasticity drives valuation premium — the market views it as the high-growth name in electronic chemicals.

Jianghua Micro (603078.SH): Characterised by stability and high gross margins; P/E valuation relatively conservative, a "value-oriented" holding. Gross margin of approximately 39% ranks among the better performers in the listed electronic-chemical universe.

Hoshine Silicon (603260.SH): Mixed cyclical-plus-transformation valuation; P/E is heavily influenced by the organosilicone industry cycle (currently at a trough), with the electronic-chemical transition narrative being the key potential valuation re-rating catalyst — 2026 electronic-grade silicone-gel certification progress is the key variable to monitor.

Silica Stock (603688.SH): High-purity quartz sand leader; valuation surged during the 2023–2024 semiconductor-materials boom and has since partially corrected. Watch certification progress for semiconductor-crucible-grade ultra-pure quartz sand and capacity expansion timelines.

10.5 ESG Dimension: Green-Transition Pressures and Opportunities for Electronic-Chemical Companies

As the electronic-chemical industry progresses from "catch-up domestic substitution" toward "global competition," ESG (Environmental, Social, and Governance) factors will rise from a secondary concern to a mainstream competitive dimension.

EU REACH and RoHS Chemical Compliance Pressure

The EU's Substances of Very High Concern (SVHC) list continues to expand. Beyond the already-listed D4 (octamethylcyclotetrasiloxane), certain silane coupling agent grades (particularly methoxy-containing grades such as KH-530/570, due to methanol hydrolysis product toxicity concerns) are also on EU regulatory radars. Companies exporting to Europe must complete REACH registration (required for chemical export volumes exceeding 1 MT/year) and continuously track SVHC list updates to maintain compliance in the European market.

RoHS (Restriction of Hazardous Substances Directive) primarily targets lead, mercury, cadmium, hexavalent chromium, and similar substances in electrical and electronic equipment; however, with the potential expansion of RoHS 3.0 to additional chemical substances, silicon-based chemical suppliers should proactively conduct product compliance assessments to avoid involuntary market withdrawal in Europe.

Green Supply Chain Certification: Carbon Footprint Pressure Cascading Upstream

TSMC, Samsung, Intel, and NVIDIA have committed to Scope 3 greenhouse-gas emission compliance (covering the entire supply chain) by 2030–2050, and are requiring core suppliers to provide Product Carbon Footprint (PCF) data. Electronic chemicals, as direct input materials for wafer manufacturing, are a significant contributor to Scope 3 emissions. Suppliers will inevitably be required to: establish and disclose PCF data (per ISO 14067 or Ecoinvent databases); provide PCF verification reports to wafer-fab customers demonstrating compliance with green-supply-chain requirements; and actively adopt green electricity (REC/I-REC certified) in production to reduce product carbon intensity and improve scores on wafer-fab green-supplier scorecards.

For Chinese polysilicon companies, production bases powered predominantly by Xinjiang coal-fired electricity carry product carbon footprints far higher than those of bases powered by Sichuan/Yunnan hydropower. This carbon-footprint differential will become an increasingly significant competitive factor over the next 3–5 years — as green-supply-chain requirements penetrate Asia-Pacific fab procurement decisions (South Korea, Taiwan), high-carbon-intensity products will face additional premium requirements or procurement obstacles.

Circular Economy: Recycling of Silicon-Material Waste

Electronic-chemical production inevitably generates off-spec waste (downgraded product) and silicon-containing waste streams. Recycling high-value silicon-containing waste liquids (such as TEOS-containing organic-solvent waste) is becoming a dual driver of raw-material cost reduction and environmental compliance:

TCS/SiCl₄ recycling: Silicon tetrachloride (SiCl₄), a by-product of polysilicon production, is a severe pollutant if directly discharged; conversion back to trichlorosilane via "hydrogenation" (SiCl₄ + H₂ → SiHCl₃) substantially lowers raw-material costs and eliminates emissions. Large polysilicon companies (Tongwei, Xinte, etc.) have built SiCl₄ recycling systems achieving ≥ 99% SiCl₄ recycling rates — the regulatory threshold below which discharge penalties apply.
TEOS production solvent recovery: Dilute ethanol waste streams (containing water, ethanol, trace silicates) generated in TEOS production can be distilled to recover ethanol for re-use, reducing per-unit ethanol input costs and reducing waste-stream discharge.
CMP spent slurry recovery: CMP polishing slurry (containing ultra-pure SiO₂ abrasive) becomes a complex waste stream containing abrasive particles, copper/tungsten, and other metals after use; solid-liquid separation, metal removal, and SiO₂ abrasive concentration/recovery are important dimensions of semiconductor-factory wastewater treatment and materials circularity.

These ESG pressures will compel Chinese electronic-chemical companies to upgrade from "meeting product-quality standards" to "meeting full-lifecycle sustainability requirements" — further raising industry entry barriers and creating new competitive dimensions (green certification credentials), which will be a new source of long-term competitive moat for companies with foresighted positioning.

Chapter 11　Policy and Standards

11.1 Domestic Policy Support Framework

National Integrated Circuit Industry Investment Fund Phase III (Big Fund III)

In June 2024, Big Fund III was officially registered with paid-in capital of RMB 344 billion — a record for any Chinese industrial fund. From a sector-allocation standpoint, Big Fund III continues to support wafer manufacturing (SMIC, Hua Hong Semiconductor, etc.) while explicitly listing "key materials and components domestic substitution" as a strategic priority direction.

For electronic chemicals specifically, Big Fund III is expected to concentrate investment in 2025–2028 on: photoresists (ArF/EUV photoresist domestic substitution); precursors (high-purity ALD/CVD precursors); ultra-pure solvents (ultra-pure IPA, ultra-pure NMP, etc.); high-purity reticle materials; and high-purity target materials. Electronic-grade TEOS and HMDS, as indispensable standard process chemicals for wafer fabs, also fall within Big Fund's materials-domestic-substitution scope.

In 2025, Big Fund III had deployed approximately RMB 164 billion in AI chip design and HBM memory (per media disclosures through end-2025); substantial investment in the materials/components direction is expected to concentrate during 2026–2027.

MIIT's Petrochemical and Chemical Industry Stable-Growth Plan (2025–2026)

In September 2025, the Ministry of Industry and Information Technology and six other ministries jointly issued the Petrochemical and Chemical Industry Stable-Growth Work Plan (2025–2026), establishing the following policy directions highly relevant to electronic chemicals:

Expanding high-end supply: focusing on downstream needs across integrated circuits, new-energy vehicles, and new-display sectors, with priority support for technology breakthroughs and engineering implementation of key electronic-chemical products
Promoting "specialised, refined, unique, and innovative" (SRNI) development: encouraging companies in organosilicone, electronic chemicals, and fine chemicals to advance toward the SRNI and "single-champion" development pathways, improving competitiveness in sub-niche categories
Supporting integrated industrial-park development: around western-China low-cost-energy regions (Sichuan, Yunnan, Inner Mongolia, Xinjiang), supporting integrated polysilicon–cell–module industrial park development to enhance industrial clustering and synergistic cost reduction

This plan covers 2025–2026, providing policy coverage and funding guidance for domestic-substitution technology breakthroughs in key electronic chemicals.

Domestic Standards System Development

MIIT's Electronics and Information Department, together with the China Electronic Materials Industry Association, is continuously advancing development of the national electronic-chemical standards system toward international SEMI standard alignment:

Already issued: national standards for electronic-grade hydrofluoric acid (UPW-grade), electronic-grade ammonium hydroxide, electronic-grade hydrogen peroxide, and electronic-grade hydrochloric acid (ultra-pure solvents)
In development: national standard for electronic-grade TEOS (expected publication 2026–2027); national standard for electronic-grade HMDS (expected 2027); high-purity polysilicon quality-grading standards

The establishment of domestic standards will provide a common language and reference framework for quality alignment between domestic suppliers and wafer fabs, while supporting broader market recognition for domestic products.

Electronic-Chemical Certification Body and Testing Capability Development

To support the certification journey of domestic electronic-chemical companies, authoritative Chinese institutions including the China Building Materials Academy (organosilicone direction), the National Semiconductor Materials Quality Supervision and Inspection Centre (MIIT-authorised), the Chinese Academy of Sciences' Shanghai Institute of Microsystem and Information Technology, and the Guangdong Semiconductor Lighting Standardisation Technical Committee are actively building SEMI-standard testing capabilities:

ICP-MS trace-metal analysis: multiple domestic laboratories can now achieve ppb-to-ppt-level metal detection; however, institutions capable of providing full-item SEMI C1 Grade 4/5 chemical certification-testing services to external clients remain limited and urgently need to be expanded
Ultra-trace organic analysis: sub-ppb-level organic impurity analysis via GC-MS/MS (triple-quadrupole gas chromatography-mass spectrometry) currently relies mainly on third-party analytical services, as in-house investment costs are high for suppliers
Particle reference standards: PSL calibration standards for particle counters are currently predominantly imported (Duke Scientific in the USA, JSR in Japan, etc.), and domestic development of ultra-pure particle reference standards needs to be strengthened to support the completeness of domestic electronic-chemical analytical-testing infrastructure

In 2025, the National Key R&D Programme's "New-Generation Display Devices and Semiconductor Lighting" direction explicitly supports R&D on "electronic-chemical precision analytical-testing key technologies" — with more domestic authoritative testing bodies expected to enter SEMI certification-testing services over the next 2–3 years, reducing supplier dependence on external certification-testing costs.

11.2 PV Overcapacity and Polysilicon Industry Policy

During 2025–2026, China's PV polysilicon industry has experienced the most severe overcapacity cycle in its history. Policy-level responses have taken several forms:

Industry-Association Coordinated Production Restraint: The China Photovoltaic Industry Association (CPIA) has repeatedly called on major polysilicon companies to exercise "self-disciplined production restraint" through meetings and public statements during 2025, to prevent continuing to sell below cost from damaging the entire industry. Leading companies including Tongwei and Xinte (GCL) have also clearly stated intentions to proactively constrain production in annual reports and investor communications, though implementation has been uneven due to varying company-level interests.

"Three Gorges Power Price" Model Exploration: Some local governments and companies are exploring binding lower-cost green electricity (hydropower, PV, wind) at even more favourable rates, driving polysilicon production power costs below RMB 0.10–0.20/kWh in certain locations. Tongwei's Leshan base and cooperation projects with Dadu River hydropower development exemplify how the southwest region maintains a relatively low-cost production advantage using abundant hydropower resources.

Export-Tax Rebates and Carbon-Footprint Certification: Targeting European markets, Chinese PV polysilicon companies are actively advancing green-electricity certification (I-REC International Renewable Energy Certificate) and carbon-footprint calculation to comply with EU Carbon Border Adjustment Mechanism (CBAM) requirements and avoid additional tariffs on "high-carbon polysilicon" entering Europe.

Special Policy Support for Electronic-Grade Polysilicon: In contrast to the "capacity management and de-stocking" policy orientation for PV polysilicon, policy for semiconductor-grade (electronic-grade) polysilicon is actively encouraging new investment. Under Big Fund III, technology-innovation special loans, and similar policy instruments, electronic-grade polysilicon projects (such as Xinlightu Silicon) receive priority support status — reflecting the national intent to achieve self-supply capability in this category as quickly as possible.

11.3 Export Controls: Two-Way Impact

The Impact of US Export Controls on China's Electronic Chemicals

The US Department of Commerce Bureau of Industry and Security (BIS) export controls on China's semiconductor supply chain have limited direct impact on most liquid electronic chemicals (the majority of liquid electronic chemicals are not themselves on the EAR control list), but indirect effects are significant:

Analytical-instrument restrictions: Certain high-precision analytical instruments used for electronic-chemical quality control (specific ICP-MS models, certain advanced gas chromatographs, etc.) are dual-use items subject to export-licence requirements. This increases procurement difficulties for domestic electronic-chemical companies building complete analytical-testing capabilities, objectively slowing establishment of the quality-assurance systems required for domestic-substitution progress.

Wafer-equipment restrictions affecting downstream demand scale: Equipment restrictions on Chinese advanced-process wafer fabs (such as SMIC) indirectly limit the pace of capacity expansion at sub-14 nm process lines, correspondingly constraining the scale of domestic demand for advanced-process electronic chemicals (particularly advanced-process TEOS/HMDS) — even if domestic production of higher-purity electronic chemicals becomes feasible, demand increments will remain relatively limited if advanced-process fab expansion itself is constrained.

Potential chemical-substance control risk: The USA and its allies are actively studying bringing certain semiconductor-process-specific chemicals (including specialty photoresists and certain ALD precursors) under export-control scope. If such controls materialise, Chinese wafer fabs' procurement options would be directly impacted, further intensifying the urgency of domestic substitution.

China's Counter-Measures

From 2023–2024, China began imposing export controls on certain critical minerals (gallium, germanium, antimony, nickel, etc.) — raw materials important to semiconductor and chemical industries in Europe, the USA, Japan, and other countries. While China has not yet imposed export controls on polysilicon or organosilicone monomers, China's absolute dominance in these categories (85–95% of global production capacity) itself represents a form of implicit "market leverage." In the PV sector, China's competitively priced polysilicon supply has been a key driver of rapid global PV cost reduction; any artificial disruption to this supply chain would significantly raise global PV deployment costs — providing China with a degree of strategic leverage in diplomatic and trade negotiations.

11.4 Dual-Carbon Policy: Green Constraints on High-Energy Silicon-Material Industries

Polysilicon smelting (particularly the electric-arc-furnace stage of industrial-silicon smelting) is a high-energy, high-emission process, creating inherent tension with China's Dual-Carbon targets (peak emissions by 2030, carbon neutrality by 2060):

Regional Differences in Energy-Intensity Control Implementation: In Sichuan and Yunnan (hydropower-rich, low-carbon), and Inner Mongolia (surplus wind/PV power), polysilicon/organosilicone production carries relatively low carbon footprints and "dual-control" constraints are applied with greater flexibility. In Xinjiang (predominantly coal-fired power), polysilicon production carries a relatively higher carbon footprint, and with the EU's Carbon Border Adjustment Mechanism (CBAM) gradually taking effect, Xinjiang-produced polysilicon entering Europe will face incrementally higher carbon costs.

Green Polysilicon Certification: Leading companies including Tongwei and Xinte have proactively pursued "green-electricity polysilicon" certification — procuring large proportions of hydropower, wind, and solar electricity (certified via I-REC) to reduce their polysilicon carbon footprint toward near-zero levels, satisfying European PV market carbon-footprint requirements.

Organosilicone Energy-Efficiency Improvement: Organosilicone monomer production (Rochow process) carries relatively high energy intensity. Hoshine Silicon, Xinaan Chemical, and others are pursuing improvements in by-product utilisation (chloromethane recovery from off-gas, process-heat utilisation) and energy-efficiency retrofits, addressing Dual-Carbon constraints through per-unit energy-consumption reduction while simultaneously lowering production costs.

Chapter 12　Trends and the Research Institute Verdict

12.1 Five Structural Trends

Trend One: Electronic-Grade Polysilicon (11N) Will Achieve Substantive Domestic Breakthrough Between 2026 and 2028

For a long time, domestic substitution of electronic-grade polysilicon has been the most conspicuous gap in China's semiconductor-materials map. With the advancement of the Xinlightu Silicon project (1,000 MT/year, first batch of samples produced in 2024) and Big Fund III's attention to this track, the Research Institute expects a "from zero to one" commercial-supply breakthrough for domestic electronic-grade polysilicon by end-2026 to 2027: the first domestic wafer fab (most likely CXMT or Yangtze Memory Technology) will complete preliminary qualification and accept small-volume domestic 11N polysilicon supply. By 2028, domestic electronic-grade polysilicon's share of domestic wafer-fab procurement could rise from the current near-zero to approximately 5–15%, marking the true opening of semiconductor-grade polysilicon domestic-substitution history.

Trend Two: TEOS Domestic Substitution Reaches 70–80% in Mature Processes, with Advanced Process Remaining a Gap

Yoke Technology has made substantive domestic-substitution progress in the TEOS direction. As the company continues to improve batch-to-batch consistency and expand capacity, the domestic penetration rate for TEOS in 28 nm and above mature processes is projected to rise from approximately 55–65% currently to 70–80% by 2026–2027. However, at 7 nm and below advanced-process TEOS qualification, the barriers from import leaders (Entegris, Wacker, etc.) will not easily erode — advanced-process domestic substitution rates may remain below 25–35% through 2027, forming a bifurcated landscape of "mature-process breakthrough first, advanced-process gradual progress."

Trend Three: HMDS Is the Biggest "Blue Ocean" for Domestic Substitution in 2026–2028

Current HMDS domestic-substitution rates of approximately 25–30% are the lowest among the major silicon-based chemical categories. Yet HMDS synthesis technology (trimethylchlorosilane ammonolysis) is relatively mature; the primary barriers concentrate in purification processes and wafer-fab qualification. Compared with TEOS, HMDS customer qualification cycles may be relatively shorter, because HMDS treatment is only an auxiliary step prior to photolithography — not a core step forming final thin-film structures — and is therefore more loosely coupled to downstream process integration. On this basis, the Research Institute judges HMDS to be the category with the greatest acceleration potential for domestic substitution in 2026–2028, with the domestic substitution rate expected to rise from the current 25–30% to 50–60% — making it the "highest-return breakthrough point" in semiconductor chemical domestic substitution.

Trend Four: Granular Silicon (FBR) Will Become the Mainstream PV Polysilicon Form, but EG Application Is a Separate Technical Question

Granular silicon, with its extremely low production cost (RMB 28.17/kg, Xinte Q4 2024 data), continues to displace traditional Siemens rod silicon's market share. By 2027, granular silicon's global market share is projected to exceed 20%. The rise of granular silicon in the PV segment will accelerate the exit of uncompetitive Siemens capacity and drive industry consolidation toward lower-cost, higher-efficiency production.

However, a breakthrough for granular silicon in electronic-grade applications is a technical question independent from the PV track — it will not be automatically achieved as PV share increases. The engineering challenge of metallic contamination control in the FBR (Fluidised Bed Reactor) process (risk of metal leaching from reactor walls and distributor plates) requires Xinte Energy to make targeted technical breakthroughs in material selection, process control, and product testing — making it the most noteworthy independent technical milestone for Xinte in the 2026–2028 timeframe.

Trend Five: Organosilicone Premiumisation — Chinese Organosilicone Monomer Giants' Strategic Reinvention

Chinese organosilicone monomer companies including Hoshine, Xinaan Chemical, and Dongyue Silicones face the "volume without value" bulk-commodity trap (the organosilicone cycle is at a historical trough; D4/D5 and other intermediate prices have fallen sharply; industry-wide profits are at historical lows). The path out lies in extending downstream into high-end products (electronic-grade silicone gel, specialty functional silicone oils, medical-grade silicone rubber, optical-grade silicone gel, etc.) to increase per-tonne value added. Hoshine's integrated industrial-silicon → organosilicone monomer → electronic-grade silicone gel deep-vertical pathway is the most complete embodiment of this transformation strategy, and will be one of the central storylines of organosilicone industry competitive-landscape evolution in 2026–2028.

12.2 Core Judgments from the Research Institute

Judgment One: "PV Polysilicon Dominance" Does Not Equal "Semiconductor Polysilicon Leadership" — Conflating the Two Is a Strategic Risk

Outsiders frequently cite "China's polysilicon production capacity exceeds 90%+ of global supply" to assert that "China's silicon materials are now fully import-independent." This is a serious misreading that risks distorting both policy decisions and industrial strategy. China's global polysilicon dominance is built entirely on PV-grade (6N–8N) products. For the 11N electronic-grade polysilicon required for semiconductor wafer manufacturing, China currently imports approximately 50% from Germany's Wacker and the US's Hemlock. If this supply chain were severed by external political factors, China's semiconductor industry's silicon-wafer supply would face severe disruption. This reality is a structural vulnerability that must be clearly recognised in any serious domestic semiconductor-materials strategic planning.

Judgment Two: Qualification Cycles Are a "Speed Ceiling," Not a Humanly Circumventable Obstacle

Some voices suggest that with sufficient policy support and capital investment, domestic-substitution progress can be "accelerated" to completion in a matter of months. This reflects a fundamental misunderstanding of how electronic-chemical qualification works. A wafer fab's qualification cycle (12–24 months) stems from objective technical requirements: process-integration testing, batch-consistency verification, reliability assessment — all engineering safeguards for chip quality and reliability. These requirements do not shrink in response to external policy pressure. What policy support and capital can achieve is enabling more companies to enter qualification concurrently and speeding up pre-qualification technical preparation — but the time cost of qualification itself is non-compressible.

Judgment Three: PV Polysilicon Overcapacity Does Not Equal Electronic-Grade Polysilicon Overcapacity — The Two Markets Are Completely Separate

Some perspectives suggest that PV polysilicon production overcapacity will "spill over" into electronic-grade polysilicon, thereby pushing down electronic-grade prices. This logic is fundamentally flawed. PV-grade polysilicon (6N–8N) and electronic-grade polysilicon (11N) differ completely in process routes, equipment configurations, purity specifications, and customer-qualification standards — they are not interchangeable, and capacity cannot be "migrated" between the two. PV polysilicon production lines cannot directly convert to electronic-grade production; entirely new dedicated purification equipment must be installed and an independent quality-control system established. PV-segment supply glut therefore has absolutely no bearing on supply-demand balance in the electronic-grade segment — the latter will continue to maintain a relatively tight supply, stable-high-price environment for a long time.

Judgment Four: HMDS Is Currently the "Highest-Return Breakthrough Point" in Domestic Substitution

Among the three core categories — TEOS, HMDS, and electronic-grade polysilicon — HMDS presents the relatively lowest technical difficulty for domestic substitution (synthesis routes are relatively mature; purity specifications, while demanding, are less daunting than 11N polysilicon), while simultaneously having the lowest current domestic-substitution rate (~25–30%) among the three, implying the greatest market-capture upside. The Research Institute judges that if domestic electronic-chemical companies focused resources in 2026–2027 on HMDS semiconductor-grade purification processes and customer qualification, HMDS could be the first of the three categories to achieve a significant leap in domestic-substitution rate — making it the single sub-category in the electronic-grade silicon-based chemicals sector currently most worthy of concentrated strategic focus.

Judgment Five: Supply-Chain Security Premiums Are Reshaping Wafer-Fab Procurement Decision Weightings

Before 2022, the dominant weighting in domestic wafer-fab chemical procurement decisions was "performance-to-price ratio": performance meeting specification plus a price advantage were the conditions for considering a supplier switch; switching costs (qualification cycles, engineering risk) were the primary resistance.

After 2022, "supply security" weighting has risen sharply. Even where a domestic product's purity specification or batch stability is slightly below that of an import (within an acceptable quality tolerance margin), the structural shift in procurement strategy toward "avoiding excessive dependence on any single import source" has significantly increased domestic wafer fabs' willingness to introduce domestic materials and commit resources to advancing qualification. This structural change is the most powerful market signal that the window for domestic electronic chemicals has arrived, and that Tianxia Gongchang's factory database confirms the accelerating activity of qualified domestic precursor and silicone-chemical producers across the supply chain.

12.3 The AI and HBM Era: Structural Impact on Demand for Electronic-Grade Silicon Chemicals

Since 2023, the explosive development of AI large language models (LLMs) has profoundly reshaped the demand structure of the semiconductor industry, creating quantifiable demand increments that are transmitting through the supply chain to the electronic-chemicals market.

The AI Compute Arms Race and Wafer-Fab Capacity Expansion

AI large-model training and inference demands for compute power are growing exponentially: GPT-4's training consumed approximately 10²³ floating-point operations (FLOPs), while next-generation models expected to be trained within two years will consume 10²⁵ or more. This explosive demand for compute is directly driving surging shipments of GPUs (NVIDIA H100/H200/B200 series) and dedicated AI accelerators (Google TPU, Amazon Trainium, Huawei Ascend, etc.), in turn propelling TSMC, Samsung, and other fabs into unprecedented capacity expansion. TSMC in 2025 announced a five-year capital expenditure plan exceeding USD 200 billion, a large proportion directed toward advanced-process (3 nm/2 nm) capacity build-out — the largest single-company wafer-capacity investment in history.

HBM: The Demand Multiplier for Electronic Chemicals

High Bandwidth Memory (HBM) serves as the high-speed bandwidth bridge between GPUs and large-capacity DRAM in AI training clusters, and is currently the highest-demand memory type. HBM manufacturing (3D stacked packaging, TSV through-silicon via) creates a significant "demand multiplier" for chemicals compared with conventional planar DRAM:

More CVD film deposition layers: The TSV process for HBM (etching vias approximately 5 μm diameter and ~50 μm depth into silicon, then filling with copper) involves multiple SiO₂ dielectric-layer depositions (predominantly using TEOS as the silicon source); each HBM wafer consumes approximately 30–50% more TEOS than a conventional DRAM wafer
Significantly higher ALD precursor consumption: HBM's high-aspect-ratio TSV sidewall liner and diffusion-barrier layers (TiN, TaN, WN, etc.) rely on ALD processes for conformal deposition, driving rapid growth in ALD precursor demand
Ultra-thin wafer thinning and polishing: Before HBM stacking, each DRAM wafer layer must be thinned to approximately 50–80 μm (versus approximately 780 μm for conventional DRAM); the thinning process consumes large quantities of high-precision CMP polishing slurry (containing ultra-pure SiO₂ abrasive) — a significant incremental demand driver for electronic-grade SiO₂ polishing materials
Silicone-gel applications in wafer-level packaging: HBM multi-layer stacks (4-Hi/8-Hi/12-Hi) use high-purity electronic-grade silicone gels for inter-layer bonding and edge sealing, with ultra-low ionic impurity requirements (given the extremely high signal-transmission speeds in HBM, any trace ionic contamination can impair signal integrity)

The Significance of Rapid Catch-Up by Domestic Memory Fabs

Domestic memory fab capacity ramp-up (CXMT and Yangtze Memory Technology) is the fastest-growing sub-segment of domestic electronic-chemical demand:

CXMT (DRAM): launched LPDDR5X products in 2024 (optimised for AI on-device inference), continuing to ramp capacity through 2025–2026 toward a monthly capacity target of approximately 150,000 12-inch wafers and advancing DRAM process below 17 nm — driving sustained growth in domestic procurement demand for TEOS, HMDS, and other chemicals.

Yangtze Memory Technology (3D NAND): 232-layer 3D NAND was brought to market in 2024; by 2025 advancing toward 266-layer and higher stack counts. Each additional 64 layers requires approximately 8–12 additional ALD process steps, cumulatively driving growing demand for ALD precursors (silicon-containing ALD precursors, High-k metal precursors) — making Yangtze Memory one of the fastest-growing ALD precursor customers domestically.

From an electronic-chemical demand perspective, the transmission chain — AI compute boom → wafer-fab large-scale expansion → surge in advanced packaging demand → higher electronic-chemical volumes — is clear, quantifiable, and represents the most powerful underlying demand driver for the electronic-chemicals market through 2024–2028. For domestic suppliers, the significance is that the absolute demand increment brought by the AI era (not merely "domestic-substitution rate improvement" driven share gains) means domestic suppliers completing qualification face a rapidly expanding market, not a finite share of a static pie — which is extremely positive for the sector's overall outlook.

Chapter 13　Risks

13.1 Risk One: Deep PV Polysilicon Overcapacity Weighing on the Entire Supply Chain

During 2024–2026, the PV polysilicon segment is experiencing the most typical "overcapacity–price collapse–industry losses–consolidation" cycle seen in Chinese manufacturing expansion:

Severe Glut: Global nominal polysilicon capacity exceeded 2 million MT/year at end-2024; actual demand was approximately 1.4–1.6 million MT — nominal capacity utilisation only approximately 70%, and effective capacity utilisation even lower. China's five largest polysilicon producers (Tongwei/Xinte/Daqo/Xinte Energy/Hoshine) have combined nominal capacity of approximately 1.4 million MT/year, already approaching total global demand — and together with other mid-sized producers, supply is severely oversupplied.

Prices Below Cost: During most of 2025–2026, Chinese polysilicon all-in costs are approximately RMB 45–55/kg, but average selling prices have been at or below this range for extended periods, with the industry broadly in loss or near-breakeven. In June 2026, PV N-type polysilicon prices were approximately USD 4.45/kg (~RMB 32/kg) — more than 30% below all-in cost, with industry-wide losses persisting for approximately 18 months. Multiple companies have announced production curtailments or postponed new-project commissioning.

Transmission to the Electronic-Chemical Supply Chain:

Low PV polysilicon prices → organosilicone monomers (D4/D5, with certain overlapping applications) demand impacted → companies such as Hoshine see revenue and margin pressure
PV supply-chain capital contraction → EVA/POE encapsulant expansion slowing → silane coupling agent PV-application demand growth decelerates
PV-segment talent migration (some engineers transitioning to electronic-grade chemical roles) — objectively building a degree of human-resource reserve for domestic electronic-grade polysilicon substitution

Consolidation Forecast: The Research Institute expects the PV polysilicon shakeout to reach a critical juncture in 2026–2027: mid-sized and smaller capacity (all-in cost above RMB 60/kg) will accelerate exit (shutdown, closure, or absorption), industry capacity contracting from the current >2 million MT/year to approximately 1.2–1.5 million MT/year, with top-3 concentration (CR3) rising from approximately 50% currently to 65–70%. Survivors among leading companies (Tongwei, Xinte, Daqo, etc.) will enjoy volume-and-price upside in the next demand-recovery cycle (expected 2027–2028).

13.2 Risk Two: Electronic-Grade Purification Technical Barriers — Not Overestimating the "Breakthrough Space"

Although the Research Institute maintains cautious optimism on the domestic-substitution outlook for electronic-grade silicon-based chemicals, the following deep technical challenges deserve clear-eyed recognition:

FZ Purification Limits for 11N Purity: Raising polysilicon purity from 9N to 11N requires multiple float-zone (FZ) purification passes beyond the standard Siemens process. Each FZ operation achieves limited incremental purification, and the FZ equipment itself (high-frequency induction coils, quartz atmosphere tubes) may introduce new contamination at high temperatures — creating an engineering paradox of "purification then re-contamination." Resolving this requires meticulous engineering control of FZ equipment materials, operating parameters, and chamber atmosphere — embodying decades of process know-how from Wacker and Hemlock that cannot be quickly replicated by reverse engineering.

Non-Documentable Process Know-How: Large amounts of critical knowledge in electronic-chemical manufacturing exist as "tacit knowledge" — operator experience, equipment calibration parameters, judgment calls on process anomalies — that does not appear in patent literature or academic papers. Accumulating this tacit knowledge requires years to decades of actual production operation. It is the deepest part of the technology moat, and is the core capability that domestic new entrants cannot quickly acquire simply by importing equipment or reverse-engineering patents.

Non-Compressible Qualification Timelines: As noted, the 12–24-month wafer-fab qualification timeline is an objective technical requirement that cannot be compressed by domestic-substitution urgency. This is a "speed ceiling" that requires realistic timetables when reporting expected commercialisation timelines to decision-makers.

13.3 Risk Three: Overseas Giants on the Offensive — Widening the Technology Lead and Capturing Capacity

Overseas electronic-chemical majors are not standing still; they are actively increasing investment to strengthen and maintain their technological leadership:

Shin-Etsu Chemical's ~USD 700 Million Expansion: Explicitly targeting top-tier fabs at 3 nm and below, deepening qualification moats at the advanced-process level and pushing the domestic-substitution window for advanced-process TEOS and polysilicon further out in time.

Tokuyama + OCI Malaysia Plant: Building lower-cost polysilicon capacity in Southeast Asia with the aim of competing with Chinese products on cost.

Entegris's Continuous Qualification Investment: Through ongoing advanced-process qualification work (TEOS, HMDS for sub-2 nm nodes), Entegris is continuously deepening its qualification relationships with top-tier fabs — making "switching suppliers" progressively more costly.

Technology-Roadmap Lock-in: Through patent filing (global ALD precursor-related patent applications surging 2020–2025), international majors are expanding intellectual-property lock-in of advanced-process precursor core know-how across more dimensions — increasing the technical-workaround difficulty for domestic entrants.

These moves mean that even if domestic companies achieve mature-process electronic-chemical domestic-substitution breakthroughs in 2026–2028, full domestic substitution in advanced processes (sub-7 nm) may not arrive until after 2030, and will require facing continued competitive pressure from international majors.

13.4 Risk Four: Rising Environmental Compliance and Occupational-Safety Costs

EU SVHC Restriction on Organosilicone D4: D4 (octamethylcyclotetrasiloxane) has been listed as a Substance of Very High Concern (SVHC) by the European Chemicals Agency (ECHA), with restrictions on its use in personal-care products under REACH, and potential extension to industrial applications in Europe. While direct restriction of D4 in semiconductor applications is unlikely, this regulatory trend increases compliance costs for Chinese D4 producers exporting to Europe, and may also affect future European market access for D4-derived electronic-grade silicone gels.

Environmental Infrastructure Investment for Polysilicon Production: The polysilicon manufacturing process (particularly the modified Siemens process) generates silicon tetrachloride (SiCl₄) as a by-product — a severely corrosive and toxic substance if not properly managed. As Chinese environmental regulation tightens (fixed-pollution-source discharge-permit management continuing to improve), SiCl₄ recycling rates must exceed 99%, and related environmental-infrastructure investment is a significant fixed cost for polysilicon companies.

Occupational Safety Management for HMDS and Trimethylchlorosilane: HMDS, TMS-Cl, and related organochlorosilanes have strong pungent odours and corrosive properties. Production requires comprehensive enclosed production systems, personal protective equipment (PPE), and emergency-response capabilities; occupational health and safety management costs continue to rise as regulation tightens.

13.5 Risk Five: Dual Constraints of Talent Scarcity and IP Barriers

Structural Talent Scarcity

The core talent required for electronic-grade silicon-based chemicals is a highly interdisciplinary group with simultaneously deep backgrounds in fine organosilicone synthetic chemistry, ultra-pure chemical purification engineering, semiconductor process chemistry, and high-precision analytical testing — an extremely scarce resource domestically.

From the existing talent pool, domestic professionals with integrated R&D and process-engineering capabilities in electronic-grade organosilicone chemicals are concentrated primarily in graduates from: the Shanghai Institute of Organic Chemistry (SIOC, CAS, organosilicone direction), the Institute of Chemistry (CAS, silicon-containing polymers), East China University of Science and Technology (fine chemicals), Dalian University of Technology (organosilicone monomer processes), Nanjing Tech University (fine chemicals), and similar institutions. However, the number of graduating students each year with research experience specifically in electronic-grade directions is extremely limited — estimated nationally at no more than 200–300 per year.

By comparison, international leading companies (Shin-Etsu, Wacker, Entegris, etc.) have decades-deep R&D teams and continuously recruit fresh talent from top chemical engineering programmes in Japan, Germany, and the USA — a decisive talent-pipeline advantage.

Three pathways currently address this talent scarcity:

Returnee talent recruitment: attracting Chinese researchers working in semiconductor-chemical R&D in Japan, Germany, and the USA back to China — currently one of the most important talent-sourcing channels for domestic electronic-chemical companies
Internal cultivation: a "new recruits + mentoring + overseas research visits" system to progressively develop chemical-engineering graduates into competent electronic-chemical process engineers over 3–5 years — the most reliable approach for internalising process know-how, but costly and slow
Team acquisition through M&A: as with Yoke Technology's UP Chemical acquisition, which effectively imported the Korean technical team wholesale — the fastest route, but increasingly exposed to geopolitical headwinds that could narrow this M&A window

IP Barriers: Patent Thickets and Technical Workarounds

In ALD precursors, high-purity TEOS purification processes, and ultra-high-purity polysilicon refinement, international leading companies have laid extensive global patent protection. For ALD precursors alone, global patent applications grew at an average annual rate of approximately 25% between 2015 and 2025; companies from the USA (Entegris, Air Liquide, etc.), Japan (Shin-Etsu, JSR, etc.), and South Korea (SK Materials, etc.) have cumulatively filed more than 3,000 ALD-precursor patents (across US, European, Japanese, Korean, and Chinese patent offices), creating a "patent thicket" covering precursor molecular structures, synthesis processes, purification methods, and ALD process parameters.

Practical impacts on domestic companies: high technical-workaround costs — avoiding existing patents may require developing alternative synthesis routes that are at a cost or performance disadvantage; and infringement risk — if domestic companies' products are found to infringe by international competitors upon entering international markets (e.g. supplying to TSMC or Samsung), patent litigation exposure may follow. This risk warrants particular attention for companies such as Yoke Technology and Shanghai Xinyang that have already entered international supply chains.

Effective IP counter-strategies include: systematic Freedom-to-Operate (FTO) analysis of existing patents at the product-development stage; building a portfolio of original patents through proprietary R&D, progressively accumulating "patent-for-patent" cross-licensing negotiating leverage; and actively participating in international standard-setting (SEMI standards working groups, IEC/ISO technical committees) to embed proprietary technology in standards and create Standard-Essential Patents (SEPs) — achieving a proactive position in global competition.

R&D Intensity Gap

Domestic electronic-chemical companies' R&D intensity (R&D expenditure/revenue ratio) lags significantly behind international peers: A-share listed electronic-chemical companies typically have R&D intensity of approximately 4–8%; international leaders such as Shin-Etsu and Entegris spend approximately 10–15%, and because of their far larger revenue bases (tens of billions to hundreds of billions of USD), their absolute R&D spending is 10–50 times that of Chinese peers.

Bridging this gap requires: leveraging Big Fund III policy capital to mobilise greater R&D investment; building university/research-institute–industry consortia (analogous to Taiwan's ITRI Chemical Materials Division model) to pool resources; and achieving focused "leapfrog" innovation in niche sub-categories (concentrating resources for local breakthroughs in specific categories), a strategy of outsize impact through concentrated effort.

13.6 Supplementary Appendix: Key Process Parameters, Material Specifications, and Supply-Chain Company Reference Index

Appendix A: Electronic-Grade TEOS Core Process Parameters

In real wafer-fab production environments, TEOS process parameters vary considerably across process nodes and thin-film types. The following summarises typical parameter windows for different CVD process types:

PECVD-TEOS Process Parameters (Typical Values)

Process Parameter	Typical Range	Control Objective
Reaction temperature	350–450°C	Ensure complete TEOS decomposition; avoid residual organic groups
Chamber pressure	200–1,000 Pa	Control reaction rate and film uniformity
TEOS flow rate	500–3,000 mg/min	Adjusted to deposition-rate target
O₂/TEOS flow ratio	4:1 to 8:1	Ensure complete oxidation; avoid carbon residues
Plasma power	200–1,500 W	Control deposition rate and film stress
Deposition rate	50–300 nm/min	Throughput efficiency
Film refractive index	1.44–1.46 (@633 nm)	Confirm SiO₂ stoichiometry

PECVD-TEOS-deposited SiO₂ films typically contain 4–8 wt% residual OH groups in the as-deposited state (from TEOS ethoxy-group hydrolysis), resulting in wet-etch rates (WER) 2–5 times higher than thermally grown SiO₂. Subsequent high-temperature annealing (700–900°C, N₂ atmosphere, 30–60 minutes) can reduce WER to near thermal-oxide levels while improving film density and insulation — an important post-treatment step in PECVD-TEOS integration processes.

SA-CVD TEOS Process Parameters (Typical Values)

Process Parameter	Typical Value
Reaction temperature	450–750°C
Chamber pressure	2,000–15,000 Pa
O₃/TEOS molar ratio	5:1 to 15:1 (SA-CVD uses O₃ to improve gap-fill capability)
Deposition rate	50–200 nm/min

The advantage of SA-CVD is that with O₃ participation, TEOS-derived SiO₂ has a degree of surface migration capability (surface flow), enabling void-free filling of high-aspect-ratio structures (aspect ratio > 8:1) — a capability beyond PECVD-TEOS, making SA-CVD-TEOS indispensable for high-aspect-ratio Pre-metal Dielectric (PMD) gap-fill in advanced-process front-end integration.

Appendix B: Electronic-Grade Polysilicon Purity Specification Reference

Polysilicon Purity Grade Specification Table

Grade	Purity (N-value)	B content (ppba)	P content (ppba)	Fe (ppbw)	Primary Application
PV Grade A	≥ 6N (99.9999%)	≤ 0.3	≤ 0.3	≤ 1.0	P-type polycrystalline/monocrystalline PV
PV N-type	≥ 7N (99.99999%)	≤ 0.1	≤ 0.1	≤ 0.5	N-type TOPCon/HJT PV
Semiconductor 9N	≥ 9N (99.9999999%)	≤ 0.01	≤ 0.01	≤ 0.1	MCZ single crystal, power devices
Semiconductor 10N	≥ 10N (99.99999999%)	≤ 0.005	≤ 0.005	≤ 0.05	CZ single crystal, DRAM/logic
Semiconductor 11N	≥ 11N (99.999999999%)	≤ 0.001	≤ 0.001	≤ 0.01	FZ single crystal, most advanced processes

The gap between PV N-type (7N) and semiconductor 9N is approximately 10-fold in B and P allowable content, and approximately 5-fold in Fe allowable content — with correspondingly significant differences in purification investment. The final two orders-of-magnitude improvement from 9N to 11N represent the steepest part of the technical barrier — the core moat of Wacker, Hemlock, and Tokuyama.

It is also worth noting that some market commentators (including Bernreuter Research) have pointed out that claims of "12N polysilicon" in the market involve some exaggeration. Under laboratory conditions, float-zoned local sample measurements may reach 12N levels; however, as a stable commercial product produced at manufacturing batch scale, 12N-equivalent production through any commercially viable, scalable route does not currently exist — a point requiring critical reading when reviewing supplier marketing materials.

Appendix C: Silane Coupling Agent Variety Classification and Selection Logic

Silane Coupling Agent Classification by Functional Group

Functional Group Type	Representative Grades (China/International codes)	Primary Organic-Phase Substrate	Typical Applications
Amino (-NH₂)	KH-550 / A-1100	Epoxy resin, phenolic resin	Optical fibre coating, fibreglass reinforcement, mineral wool
Epoxy (-CH₂CHCH₂O-)	KH-560 / A-187	Epoxy resin, acrylate	PV EVA encapsulant film, chip underfill adhesive
Methacrylate (-OCOC(CH₃)=CH₂)	KH-570 / A-174	Unsaturated polyester, polyamide	Lithium-battery separator coating, optical adhesive
Vinyl (-CH=CH₂)	A-151 / VTMO	Polyethylene, polypropylene, silicone rubber	XLPE insulation, silicone-rubber reinforcement
Mercapto (-SH)	KH-580 / A-189	Natural rubber, synthetic rubber	Tyre white-carbon-black modification
Isocyanate (-NCO)	KH-A / A-1310	Polyurethane	PU product-silica filler interface
Chloropropyl (-CH₂CH₂CH₂Cl)	KH-550 precursor	Epoxy/amine derivatives	Organic-synthesis intermediates

In PV and lithium-battery applications — two major new-energy downstream markets — silane coupling agent selection trends are moving toward increasing specialisation and functional enhancement. In PV encapsulant films, traditional EVA formulations (with KH-560 addition) are shifting toward POE (polyolefin elastomer) plus specialty coupling agent formulations, because POE has lower reactivity with conventional KH-560 and requires purpose-developed vinyl or acrylate coupling agents with higher reactivity — pushing domestic silane coupling agent companies into the specialty-grade development competitive track.

Appendix D: Global and China Supply-Chain Company Reference Index

Global Major Electronic-Grade Silicon-Based Chemical Company Index

Company	Country/Region	Core Products	Market Position
Shin-Etsu Chemical	Japan	Organosilicone, semiconductor wafers, TEOS, photoresists	Organosilicone/wafers global No. 1; TEOS top 3
Wacker Chemie	Germany	EG polysilicon, organosilicone (Wacker Silicones)	EG polysilicon global No. 1 (tied with Hemlock)
Hemlock Semiconductor	USA	EG polysilicon	North America's largest EG polysilicon supplier
Tokuyama	Japan	EG polysilicon	Japan's most important EG polysilicon supplier
Entegris	USA	TEOS, HMDS, CMP slurry, filtration systems	Global leader in semiconductor-materials solutions
Merck KGaA / Versum	Germany	HMDS, low-k dielectric precursors, etch gases	HMDS global top 3; Versum brand well-known
Air Liquide	France	ALD/CVD precursors, specialty gases	Precursors global top 3; Suzhou China JV
Dow Inc.	USA	Organosilicone (Dow Silicones), polysilicon (Hemlock stake)	Organosilicone global top 3
OCI Company	South Korea	Polysilicon (PV + EG), chemicals	South Korea's largest polysilicon producer
Evonik Industries	Germany	TEOS, fumed SiO₂ (HDK)	TEOS global top 3
Soulbrain	South Korea	TEOS, etchants, precursors	TEOS global top 6; South Korea leader

China Major Electronic-Grade Silicon-Based Chemical Company Index

Company	Listed Code	Core Products	Market Position
Yoke Technology	002409.SZ	ALD/CVD precursors (incl. TEOS), flame retardants	China's No. 1 precursor company
Hoshine Silicon	603260.SH	Industrial silicon, organosilicone monomers D4/D5, electronic-grade silicone gel	Global largest organosilicone industrial-chain company
Tongwei	600438.SH	High-purity polysilicon (PV-grade), solar cells	Global No. 1 polysilicon capacity
Xinte Energy	3800.HK	Granular silicon (FBR), TCS	Global No. 1 granular silicon
Daqo New Energy	688303.SH	Polysilicon (Siemens rod)	Global top-5 polysilicon supplier
Shanghai Xinyang	300236.SZ	Precursors, copper interconnect plating solutions, cleaning chemicals	High-growth semiconductor-chemical stock
Jianghua Micro	603078.SH	Ultra-clean high-purity reagents, wet electronic chemicals	Domestic leader in wet electronic chemicals
Xingfa Group	600141.SH	Organosilicone, phospho-chemicals, electronic-grade HF	Phospho-silicon chemical synergy platform
Silica Stock	603688.SH	High-purity quartz sand, quartz components	Domestic high-purity quartz sand leader
United Xin Chemistry	Unlisted	Silane coupling agents (standard + specialty grades)	Domestic coupling agent leading company
Nantong Sogeler	Unlisted	TEOS (industrial/LCD grade)	Global top-6 TEOS supplier
Xinlightu Silicon	Unlisted (preparing)	Electronic-grade polysilicon (11N)	China's EG polysilicon breakthrough project

Appendix E: Supply-Chain Risk Matrix and Response Strategies

Supply-Chain Risk Matrix (2026 Perspective)

Risk Dimension	Specific Risk Factor	Risk Level	Response Strategy
Supply side	PV polysilicon overcapacity; prolonged price depression	Medium–High	Cost control + await consolidation
Supply side	EG polysilicon (11N) domestic capacity negligible; high dependence on Wacker/Hemlock	High	Domestic R&D investment + Xinlightu qualification
Supply side	High-purity quartz sand premium ore (Spruce Pine, USA) concentration risk	Medium	Develop domestic Lianyungang refining + Norway alternative
Demand side	Advanced-process wafer-capacity expansion limited by export controls	Medium	Focus domestic substitution on mature-process + memory fabs
Geopolitics	Potential US expansion of chemical export-control scope	Medium	Accelerate domestic-substitution qualification timelines
Geopolitics	Narrowing outbound M&A window (tightening acquisition review)	Medium	Shift toward proprietary R&D + domestic collaboration
Technology	Qualification cycles (12–24 months) are non-compressible	Low (inherent)	Plan ahead, advance with patience
Technology	Process know-how accumulation gap; not quickly closable	High (long-term)	Long-term R&D investment + talent acquisition

Recommended Response Priority Sequence:

Near-term (1–2 years): Focus on categories already in the qualification ramp-up phase (mature-process TEOS, HMDS), driving ongoing wafer-fab qualifications to completion as quickly as possible; simultaneously maintain close tracking of Xinlightu Silicon's EG polysilicon qualification progress, ensuring adequate resource allocation.

Medium-term (3–5 years): Increase R&D investment in HMDS and sub-14 nm advanced-process TEOS, targeting domestic HMDS substitution rate exceeding 50% in advanced processes by 2029; drive the first domestic EG polysilicon (11N) into formal commercial supply at CXMT/Yangtze Memory.

Long-term (5–10 years): Pursue qualification breakthroughs at global top-tier fabs (TSMC, Samsung) for electronic-grade silicon-based chemicals, extending the domestic-substitution strategic high ground from "Chinese wafer fabs" to "global wafer fabs"; simultaneously advance domestic high-purity quartz sand (ultra-pure crucible-grade) refining technology, progressively reducing dependence on Spruce Pine (USA) ore.

Appendix F: Key Milestones through 2030

2026:

Xinlightu Silicon's first EG polysilicon samples complete initial chemical evaluation acceptance at a domestic wafer fab (expected to be CXMT or Yangtze Memory)
Yoke Technology's TEOS domestic-substitution rate at 28 nm/14 nm nodes exceeds 65%
Domestic mature-process (28 nm and above) HMDS qualified suppliers increase to 3 or more
Hoshine Silicon launches first certified-sample batch of electronic-grade silicone gel
Big Fund III completes its first concrete investment closes in the electronic-chemicals direction

2027:

Xinlightu Silicon's EG polysilicon (11N) completes formal qualification at CXMT or Yangtze Memory Technology; enters small-volume official commercial supply
Domestic mature-process TEOS domestic-substitution rate reaches 70–75%
HMDS domestic-substitution rate rises to 40–50%
Specialty high-grade silane coupling agents (solid-state battery, HJT cell specialty grades) achieve substantive domestic-substitution progress
High-purity quartz sand crucible-grade refining technology achieves domestic self-supply for 90%+ of domestic wafer-fab demand

2028:

Domestic EG polysilicon (11N) annual capacity exceeds 5,000 MT; domestic market share at domestic wafer-fab procurement rises above 15%
Advanced-process (sub-14 nm) TEOS domestic-substitution rate exceeds 30%
HMDS domestic-substitution rate exceeds 55%; Chinese companies become mainstream suppliers at mature-process customers
Initiate first sample-submission evaluations at TSMC/Samsung-level international top-tier fabs for EG polysilicon and high-purity TEOS

2029–2030:

Domestic EG polysilicon (11N) annual capacity exceeds 10,000 MT; semiconductor-grade polysilicon import volumes substantially compressed (domestic-substitution rate exceeding 40%)
Mature-process TEOS and HMDS near-complete domestic supply achieved
Advanced-process (7 nm) TEOS and HMDS domestic-substitution rates reach 50%+
First preliminary qualification by an overseas top-tier fab (TSMC overseas plant or Samsung facility) obtained for domestic electronic chemicals

The above milestone projections are benchmarked to "historically credible rates of progress" rather than "politically optimistic estimates," intended to provide industry decision-makers with a realistic and actionable timeline reference. If domestic technical breakthroughs and policy-support intensity exceed expectations, some milestones may be achieved ahead of schedule; if international controls escalate or technical difficulties exceed expectations, timelines may also be pushed out.

Appendix G: Notes for Investors in China's Electronic-Grade Silicon-Based Chemical Sector

In China's A-share and Hong Kong equity markets, electronic-grade silicon-based chemicals is an actively covered thematic sector, but the following cognitive blind spots warrant attention:

Blind Spot One: Applying "earnings fundamentals" frameworks to "qualification-type growth" companies. For companies in the qualification ramp-up phase (no large-scale revenue until qualification is complete), current revenue scale does not reflect future revenue potential. A more appropriate valuation framework is "technical qualification milestones" — qualification completion is the triggering event for a revenue inflection — rather than conventional P/E valuation.

Blind Spot Two: Conflating "PV-grade polysilicon" and "electronic-grade polysilicon" earnings drivers. The financial performance of polysilicon majors such as Tongwei and Xinte is >99% driven by PV-grade polysilicon. Valuing these stocks should be based on PV bulk-cycle logic (supply-demand balance, price trends) — not semiconductor domestic-substitution narratives, which currently contribute negligibly to these companies' actual results but are frequently amplified by market commentary.

Blind Spot Three: Overestimating TEOS's share of the PV market. Some research attributes the entire TEOS market size (~USD 270 million) to the semiconductor category; in reality, optical-fibre-grade TEOS accounts for a significant proportion of total consumption. True high-purity semiconductor-grade TEOS market size is considerably smaller. Individual stock analysis must carefully distinguish between LCD-grade/optical-fibre-grade/semiconductor-grade revenue composition in each company's TEOS product mix.

Blind Spot Four: Ignoring the "bidirectional nature of qualification barriers." The qualification barrier not only protects already-qualified international suppliers; it will also durably protect the first domestic companies to achieve qualification breakthroughs. Once a domestic company completes wafer-fab qualification and enters the QVL, later domestic entrants must undergo equally long qualification processes and cannot quickly replicate the first mover's success. This means "first-mover advantage" is highly pronounced in the electronic-chemical sector — the first domestic company to pass qualification will sustain competitive advantages rather than face rapid commoditisation.

Data Sources

Data in this report is drawn from the following sources, current as of 19 June 2026:

I. International Industry Research Data

Valuates Reports: Global TEOS for CVD Market Report (2025) — Asia-Pacific 84% share; top-six suppliers 75% market share
Bernreuter Research: Global Polysilicon Producer Rankings and Polysilicon Price Trends (2025–2026) — price data and Wacker/Hemlock share data
MarkWide Research: Electronic-Grade Polysilicon Market Forecast 2026–2036 — USD 1.074 billion (2026); USD 1.576 billion (2035)
GlobalGrowthInsights: Global HMDS Market Size and Share Report 2025–2035 — USD 178 million / 8.4% CAGR
QY Research: Electronic-Grade HMDS Market Report (2025) — electronic-grade HMDS CAGR data
FutureMarketInsights: PV Silane Coupling Agent Market Report 2025–2035 — USD 1.4 billion (2025)
6W Research: China Silane Coupling Agent Market Report 2025–2031 — China leading at 7.8% CAGR
GlobeNewswire / Verified Market Research: High-Purity Quartz Sand Market Report (2025) — USD 6.26 billion by 2031
24ChemicalResearch: Global D4 Market Report 2025–2032 — USD 1.34 billion (2024)
MarketsAndMarkets: Global Silane Coupling Agent Market Report (2024–2025) — USD 1.65 billion (2025)
IntelMarketResearch: Electronic-Grade TEOS Market Outlook 2026–2032
ArchiveMarketResearch: Semiconductor-Grade TEOS Market Report (2025)
DataIntelo: Semiconductor-Grade Polysilicon Market Report (2025)
IMARC Group: Polysilicon Price Trends (May–June 2026 real-time data)

II. Chinese Company Public Financial Information

Hoshine Silicon (603260.SH) 2024 Annual Report (April 2025) — industrial silicon 1.22 million MT/year; organosilicone monomer 1.73 million MT/year
Hoshine Silicon 2025 financial analysis (Sina Finance, April 2026) — revenue RMB 26.6 billion; net profit RMB 1.7 billion
Jianghua Micro (603078.SH) 2025 Semi-Annual Report — H1 revenue RMB 459 million, +5.80% YoY
Jianghua Micro (603078.SH) 2025 Annual Report (April 2026)
Shanghai Xinyang (300236.SZ) 2025 net-profit growth data (Electronic Engineering Zhuanji, 2026) — net profit RMB 301 million, +71.12% YoY
Tongwei (600438.SH) 2025 capacity announcement — over 900,000 MT/year
Eastmoney: "Silicon Precursor Materials" analysis (January 2026)
Future Think Tank: Yoke Technology Research Report 2023 — domestic precursor No. 1; TEOS price data

III. Policy and Macro Data

21 Economic Network: "Big Fund III RMB 344 Billion" (June 2024)
Sina Finance: "Big Fund III Strategic Investment Overview" (December 2025) — approximately RMB 164 billion deployed to date
MIIT and six other ministries: Petrochemical and Chemical Industry Stable-Growth Work Plan (2025–2026) (September 2025)
Semiconductor Industry Association (SIA): Polysilicon Section 232 Investigation Public Comments (August 2025) — Wacker + Hemlock 75% share

IV. Market Prices and Industry Developments

PV Magazine: "China Polysilicon Prices Fall More Than 13%" (March 2026) — USD 6.47/kg price data
PV Magazine: "China Polysilicon Prices Rise on Policy Signals" (July 2025) — USD 4.77/kg
CSIS: "China Semiconductor Localisation Progress" analytical report (2025)
Futunn Global News: "Electronic Chemicals: The Key to Breaking the Semiconductor Bottleneck" (September 2025) — RMB 174.08 billion China market
TrendForce: "China Semiconductor Equipment Domestic Self-Sufficiency Rate" (February 2025)
BusinessWire / ResearchAndMarkets: PV Polysilicon Industry Forecast 2025–2034
PV Magazine: "Tokuyama–OCI Malaysia Plant Construction" (July 2025)

V. Research Institute Proprietary Data

Tianxia Gongchang Manufacturing Research Institute: drawing on a real-time database of 4.8 million verified in-production Chinese factories to analyse the distribution and regional clustering of in-production factories across electronic-grade chemicals, wet electronic chemicals, PV polysilicon, silane coupling agents, high-purity quartz sand, and related categories (data current as of 19 June 2026)

This report is produced by the Manufacturing Research Institute and published on 19 June 2026. Third-party market forecast data cited in this report is for reference only and does not constitute investment advice. Reproduction must cite the source: Manufacturing Research Institute (faxiangongchang.com). For further information on in-production factory data and supplier identification in China's electronic-grade silicon-based chemicals sector, the platform database — covering 4.8 million in-production factories — can help you precisely locate supply-chain partners across the relevant industry segments.