Semiconductor-grade quartz and silicon components

Scope and definitions

Semiconductor-grade quartz components in this note mean fused-quartz glass parts used inside wafer-processing equipment: process-tube liners, wafer boats, injectors, showerheads, windows, edge rings, domes, and plasma-confinement rings. The material is normally high-purity SiO₂ glass, either natural quartz purified by fusing or synthetic fused silica.

Semiconductor-grade silicon components mean single-crystal or polycrystalline silicon parts used in etch and deposition chambers: silicon electrodes, focus rings, edge rings, shields, liners, and dummy wafers. These are not the same as silicon wafers that become the substrate for chips, although the high-purity silicon supply chain overlaps.

A granted plasma-processing apparatus patent [S3] shows the chamber context in which these parts are consumed. Company filings from Lam Research, Applied Materials, and Entegris [S4][S5][S6] describe the supply and qualification risk associated with such chamber materials.

Why “grade” matters

Semiconductor-grade quartz is not commodity industrial sand. The critical requirements are:

• Metallic purity: alkali metals, transition metals, and heavy metals must be controlled because they migrate into wafer surfaces or etch chamber surfaces and cause device-yield loss. USGS silicon commodity data [S8] describes the silicon value chain from raw quartz to high-purity silicon, with semiconductor applications demanding the highest product purity.
• Thermal stability: fused quartz has a very low coefficient of thermal expansion, about 0.55×10⁻⁶ K⁻¹ in the commonly cited temperature window [S2]. This makes it tolerant of rapid heating cycles in furnaces and etch tools.
• Plasma resistance: quartz and silicon are consumed by reactive-ion etch and deposition plasmas; the component lifetime is a trade-off between acceptable particle generation and chamber uptime.
• Bubble/inclusion control: gas bubbles, striae, and inclusions in quartz can cause localized etching and particle release. Company product literature [S7] markets “high purity” and “optical-grade” fused silica, but detailed lot-level acceptance data are generally not published in SEC filings.

Single-crystal silicon components need high bulk purity and controlled dopant levels because they contact or shape plasma directly. Ultra-high-purity electronic-grade silicon is often cited as 9N to 11N purity, meaning 99.9999999% to 99.999999999% silicon, though the exact customer specification should be treated as a needs primary source validation unless a wafer supplier’s controlled document is attached.

Material science baseline

The peer-reviewed reference for fused-silica properties is Brückner’s Properties and structure of vitreous silica. I [S2]. It reports the structural and thermal behavior of vitreous silica, including density near 2.20 g/cm³ and a very low linear thermal-expansion coefficient. The low CTE is one reason quartz is preferred over ordinary borosilicate glass for high-temperature semiconductor furnaces.

Key fused-quartz properties used in this note:

• Density ≈ 2.20 g/cm³ [S2]
• Linear CTE ≈ 0.55×10⁻⁶ K⁻¹ [S2]
• Softening point ≈ 1680°C [S2]
• High chemical resistance to many acids, but attacked by HF and by fluorine-containing plasmas [S2][S3]

Single-crystal silicon properties are standard semiconductor references. The theoretical mass fraction of silicon in SiO₂ is computed below from NIST atomic-weight data [S1]. This is the quantitative bridge between quartz raw material and silicon production.

Manufacturing and purification

4.1 Quartz to polycrystalline/electronic-grade silicon

Industrial silicon metal is produced by carbothermic reduction of quartz in an electric arc furnace:

\[

SiO_2 + 2C \rightarrow Si + 2CO

\]

Using NIST standard atomic weights [S1], the molar mass of SiO₂ is 60.084 g/mol and the molar mass of Si is 28.0855 g/mol. The theoretical silicon mass yield from pure quartz is therefore 28.0855 / 60.084 = 46.73%. Real-world yields are lower because of furnace losses, slag, off-gas dust, and downstream purification losses. For semiconductor-grade polysilicon, the process additionally requires chlorosilane distillation or silane decomposition, followed by crystal growth and wafer slicing. The worked calculation in Section 9 makes these yield assumptions explicit and puts them under sensitivity analysis.

4.2 Fused quartz components

Fused-quartz manufacturing uses natural quartz or synthetic silica feedstock. Synthetic fused silica, such as Corning’s HPFS product family [S7], is marketed for semiconductor optics and photomask substrates where high transmission and low inclusion counts matter. In chamber components, fused quartz is machined, then cleaned, annealed, and inspected. The absence of published test data in these marketing documents means any quantitative purity or bubble claim should be marked needs primary source validation unless confirmed against a controlled product datasheet.

4.3 Silicon components

Silicon electrodes, rings, and shields are machined from high-purity single-crystal or polycrystalline silicon. The production route is tied to the same electronic-grade polysilicon supply chain used for wafers. Equipment makers and component suppliers qualify chamber parts for plasma behavior, particle generation, and service life. The patent [S3] records an apparatus structure but does not qualify any specific silicon or quartz part’s lifetime; it is a configuration claim, not a performance proof.

Patent and peer-reviewed evidence

5.1 Granted patent

• US 6,541,104 B1 — “Plasma processing apparatus and plasma processing method,” granted April 1, 2003, assigned to Tokyo Electron Ltd. [S3].

This is a Tier 1 granted patent, not an application. It documents a plasma-processing chamber and associated components, which is the environment in which semiconductor-grade quartz and silicon consumables are used. The patent should be read as evidence of a recorded apparatus claim, not as evidence of commercial performance, component lifetime, or market success. It does not quantify etch rate, particle count, or quartz/silicon component replacement interval.

5.2 Peer-reviewed paper

• Brückner, R., “Properties and structure of vitreous silica. I,” Journal of Non-Crystalline Solids, Vol. 5, Issue 2, 1970, pp. 123–175, DOI 10.1016/0022-3093(70)90190-0 [S2].

This is a Tier 1, DOI-verified peer-reviewed paper. It supports the structural and thermal property claims for fused quartz. It is not specific to semiconductor chamber components, but it is the material-science anchor for fused-quartz behavior.

Cross-reference to company filings

6.1 Lam Research

Lam Research’s Form 10-K [S4] discusses supply-chain and raw-material risks in its risk-factor section. The company discloses reliance on limited sources for certain components and subassemblies used in its equipment. This is contextual evidence that quartz and silicon chamber parts are not low-risk commodity purchases; they are qualification-constrained components.

6.2 Applied Materials

Applied Materials’ Form 10-K [S5] describes risks from inability to obtain critical materials and components in a timely manner, plus potential price increases. The filing does not publish a bill of materials or component-level purity specification. Therefore, any numerical claim about how much Applied Materials spends on quartz/silicon parts is needs primary source validation.

6.3 Entegris

Entegris’s Form 10-K [S6] identifies advanced materials and components for semiconductor manufacturing as a core product area. The company serves fabs and equipment OEMs with contamination-control, fluid-handling, and specialty-materials products. Its risk disclosures include raw-material availability and supplier concentration. Entegris is not a pure-play quartz-component maker, but it is a public company materially exposed to the same advanced-material qualification economics.

6.4 Discrepancies between marketing and filed disclosures

• Marketing documents for fused silica, such as Corning’s HPFS product page [S7], emphasize “high purity,” “low stress birefringence,” and optical performance. SEC filings [S4][S5][S6] do not quantify component purity or compare quartz grades. This is a disclosure gap, not a numerical contradiction.
• A granted patent [S3] demonstrates a company recorded an apparatus claim; it does not demonstrate that a particular quartz dome or silicon electrode met a customer’s plasma-etch specification. In moat analysis, patents are stronger than unsourced marketing language, but weaker than customer-qualification or lifetime-performance data, which is usually not public.

Supply chain and value chain

The value chain can be separated into:

1. High-purity quartz mining / synthetic silica production — natural quartz feedstock or synthetic silica producers. Public data on high-purity-quartz production volumes is limited; the USGS silicon commodity summary [S8] covers silicon metal and polysilicon categories, but does not separately break out high-purity quartz for semiconductor chamber components.
2. Silicon metal and polysilicon — silicon metal producers and polysilicon refiners feed both wafer production and silicon component machining.
3. Component fabrication — machining, cleaning, annealing, and qualification of quartz and silicon chamber parts.
4. Equipment integration — OEMs such as Lam Research and Applied Materials integrate qualified components into etch, deposition, thermal, and cleaning tools [S4][S5].
5. Fab consumption — semiconductor manufacturers replace quartz boats, tubes, rings, windows, and silicon electrodes on maintenance schedules or condition-based monitoring.

Company risk language [S4][S5][S6] indicates that qualification barriers and supplier concentration are material to the supply chain.

Worked calculations

9.1 Thermal expansion of a quartz wafer boat

For a 300 mm quartz furnace carrier heated from room temperature to a 1000°C process:

\[

\Delta L = \alpha L_0 \Delta T

\]

Inputs:

• \(L_0 = 300\) mm
• \(\alpha = 0.55 \times 10^{-6}\ \text{K}^{-1}\) [S2]
• \(\Delta T = 1000 - 20 = 980\ \text{K}\)

\[

\Delta L = 300 \times 0.55 \times 10^{-6} \times 980 = 0.1617\ \text{mm}

\]

Sensitivity:

• If \(\alpha = 0.50 \times 10^{-6}\ \text{K}^{-1}\), \(\Delta L = 0.147\ \text{mm}\).
• If \(\alpha = 0.65 \times 10^{-6}\ \text{K}^{-1}\), \(\Delta L = 0.191\ \text{mm}\).

This calculation explains why fused quartz is used for high-temperature wafer carriers: it distorts less than ordinary glass for the same thermal load.

9.2 Quartz-to-silicon mass requirement, with sensitivity

Using the theoretical Si yield from SiO₂:

\[

\mu_{Si} = \frac{M_{Si}}{M_{SiO_2}} = \frac{28.0855}{60.084} = 0.4673

\]

For 1 kg of finished semiconductor-grade silicon, if the following yield factors are assumed:

• Reduction furnace silicon recovery: \(\eta_{red} = 0.85\)
• Polysilicon purification/deposition recovery: \(\eta_{poly} = 0.90\)
• Crystal growth/cropping/grinding recovery: \(\eta_{cz} = 0.60\)
• Wafer/component machining and slicing recovery: \(\eta_{slice} = 0.90\)

Then:

\[

M_{quartz} = \frac{1.0}{\mu_{Si} \times \eta_{red} \times \eta_{poly} \times \eta_{cz} \times \eta_{slice}}

\]

\[

M_{quartz} = \frac{1.0}{0.4673 \times 0.85 \times 0.90 \times 0.60 \times 0.90}

\]

Denominator = \(0.4673 \times 0.4131 = 0.19298\).

\[

M_{quartz} \approx 5.18\ \text{kg quartz per kg finished silicon}

\]

Sensitivity to reduction-furnace recovery:

• \(\eta_{red} = 0.80\): denominator = \(0.4673 \times 0.3888 = 0.1817\); quartz required ≈ 5.50 kg/kg.
• \(\eta_{red} = 0.90\): denominator = \(0.4673 \times 0.4374 = 0.2044\); quartz required ≈ 4.89 kg/kg.

Important caveat: the reduction, polysilicon, crystal-growth, and machining recovery factors are assumed for sensitivity and are needs primary source validation unless a specific plant-level recovery study is cited. The molar masses and theoretical 46.73% yield are sourced to NIST atomic weights [S1].

Parameter table

Unit economics / business model

Who monetizes it: The technology is monetized through product sales, not royalties or subscriptions. Component suppliers sell consumable quartz and silicon parts to equipment OEMs and directly to fabs. Equipment OEMs monetize the integrated tool, aftermarket service, spare parts, and upgrades [S4][S5]. Materials suppliers such as Entegris sell contamination-control and advanced-material components [S6].

How value is captured: Value is captured at the qualified-component tier because a fab cannot swap in an unqualified quartz ring or silicon electrode without risking chamber damage or yield loss. Qualification creates switching costs. This is visible in the supplier-concentration risk language in SEC filings [S4][S5][S6], which is a stronger signal than marketing language.

Margins: Public component suppliers report gross margins at the company level, and SEC filings contain the financial statements. However, this note did not retrieve a segment-level gross-margin line item specifically for semiconductor-grade quartz/silicon components. Therefore, any component-level margin figure is needs primary source validation.

Typical component prices: Quartz boats, domes, and silicon electrodes are widely reported in the industry as high-value consumables, but no primary source retrieved here provides a verifiable price. Any figure such as “$X per dome” is needs primary source validation.

Cost structure: The high cost arises from raw-material purity, cleaning, annealing, machining, metrology, and qualification overhead. The low thermal expansion and chemical resistance [S2] are the technical basis for the material choice; the business-model basis is consumable replacement during preventive maintenance.

Disconfirming Evidence

The thesis that semiconductor-grade quartz and silicon components are a durable, qualification-protected product category would be weakened or falsified by:

1. Material substitution in chamber parts: If silicon carbide, yttria-coated alumina, or other materials demonstrate longer plasma lifetimes and lower particle generation across leading-edge etch tools, demand for quartz and silicon electrodes could fall.
2. EUV/logic process shifts away from plasma-intensive steps: If future manufacturing flows reduce etch/deposition chamber hours per wafer, consumable replacement demand could slow.
3. Synthetic fused silica becomes a commoditized alternative: If synthetic quartz fabrication removes the scarcity or qualification premium associated with semiconductor-grade natural quartz, pricing power for quartz-component suppliers could decline.
4. Patent claims are invalidated or circumvented: If the apparatus claims in US 6,541,104 B1 [S3] are designed around, or if key patents in quartz/silicon component architecture expire without replacement IP, the moat weakens.
5. Customer qualification proves less sticky than assumed: If fabs routinely dual-source or open-spec quartz/silicon replacement parts without yield penalty, the switching cost implied by the supply-chain risk language [S4][S5][S6] would not support pricing power.