Scaling Up Silicon 28 For Mainstream Silicon Qubit Chips

by

Transitioning to mass production of ²⁸Si-based spin qubit chips involves leveraging existing semiconductor infrastructure (e.g., CMOS fabs from Intel or TSMC) while addressing isotope-specific challenges. Mass production at consumer electronics levels would require capacities in the hundreds of kg to tons annually of enriched Silicon 28. Production needs to be scaled up by at least 100 times and costs per kilogram to drop by over 10 times (10% or less of current costs).

Here is what is involved in scaling up this production.

Scale Up Enrichment Capacity: Build additional facilities using methods like ASP (gaseous diffusion with flow directors), quantum enrichment (laser-based separation), or centrifugal enrichment.

Current plants (e.g., ASP’s 80 kg/year) need expansion to tons/year for global supply. Diversify suppliers to mitigate risks (e.g., geopolitics affecting Russian sources)

ASP Isotopes started commercial production of Silicon-28 during late March 2025 and has successfully enriched large quantities of intermediate product to 99%. The Company expects to ship finished commercial product enriched to at least 99.995% to its first customers during August 2025. Based on the first three months of commercial production, the Company believes this plant has an annual capacity of greater than 80 kilograms of highly enriched Silicon-28 (enriched to 99.995%) when operating at maximum operating rate versus previous guidance of greater than 50 kilograms and significantly higher than the original expectations of approximately 10 kilograms. This additional capacity expansion cost approximately $4 million in fixed asset investment, which was incurred during 2H 2024 and Q1 2025.

ASP Isotopes is publicly traded and has a current market cap of just under $1 billion.

Develop Supply Chain for Precursors: Convert enriched ²⁸Si into usable forms like silane (SiH₄) gas for CVD. Ensure purity during conversion to avoid reintroducing impurities. Secure long-term contracts, as seen with ASP’s kilogram-scale deals.

Integrate into Fabrication Processes: Use 300 mm CMOS-compatible lines for gate-defined quantum dots or spin qubits. Grow ²⁸Si/²⁸SiO₂ interfaces via CVD, then pattern with lithography. Achieve cryogenic operation (near 0 K) with minimal defects. Companies like SemiQon and Intel are demonstrating this at small scales.

Invest in R&D for higher yields, automated testing, and error correction integration. Lower enrichment costs through volume (target < $1,000/kg). Ensure sustainability, as quantum fabs require ultra-pure, cryogenic environments. Regulatory approvals for nuclear-related isotopes may be needed. Enriched silicon-28 (²⁸Si) is critical for large-scale production of silicon spin qubit chips in quantum computing because it minimizes quantum decoherence caused by the nuclear spin of silicon-29 (²⁹Si), which comprises about 4.7% of natural silicon. High-purity ²⁸Si (typically >99.9% enrichment) enables longer qubit coherence times, higher fidelity, and denser integration of qubits, making it essential for scaling beyond small prototypes toward million-qubit systems. Natural silicon is ~92% ²⁸Si, but quantum applications require depletion of ²⁹Si to levels as low as 50-800 parts per million (ppm) or better to reduce noise and support error-corrected logical qubits.

Global production of highly enriched ²⁸Si is still emerging and focused on R&D and early commercial needs, with capacities in the tens to hundreds of kilograms per year. This is sufficient for prototyping and small-scale quantum chip fabrication but would need expansion for widespread mass production. Key details include:

Other sources include research-scale suppliers like NIST (U.S.), which has produced ultra-pure ²⁸Si at >99.9999% (six nines) for experimental use, but not at commercial volumes. Historical suppliers like Russia’s Isotope JSC have provided material for early qubit demos, but geopolitical issues limit access. Emerging methods, such as focused ion beam implantation or chemical vapor deposition (CVD) from enriched precursors, enable local enrichment on wafers without bulk production.

Quantum chips use thin epitaxial layers of enriched ²⁸Si (typically 100 nm to 1 µm thick) grown on standard silicon substrates via CVD. For a 300 mm wafer (industry standard for scalability), the mass of enriched ²⁸Si needed is minimal: approximately 0.016 g for a 100 nm layer or 0.165 g for 1 µm (based on silicon density of 2.33 g/cm³ and wafer area of ~707 cm²).

A single wafer can yield dozens to hundreds of quantum chips, each potentially hosting thousands to millions of qubits (e.g., up to 10⁶ qubits in a 100 µm² area with advanced designs). Thus, 1 kg of enriched ²⁸Si could support thousands of wafers or millions of chips, making current production scales adequate for R&D and pilot manufacturing but insufficient for hypothetical consumer-level mass production (e.g., billions of chips)

Current prices for 99.99% enriched ²⁸Si range from $10,000 to $30,000 per kg, driven by low-volume production. This is viable for quantum R&D but must drop significantly (e.g., through economies of scale) for broader adoption.