Why Large Single-Crystal Diamond Is Easier to Exhibit Than to Buy
A large diamond sample can prove technical possibility without creating a purchasable seed product. The missing link is a repeatable system for replication, release, finishing, qualification, and supply.
I recently circulated a specific procurement request: a non-mosaic CVD single-crystal diamond plate larger than 25 × 25 mm, (100)-oriented, 0.5–0.8 mm thick, and suitable for use as a growth seed.
The response was limited. Other researchers reported the same problem when trying to source two-inch single crystal for CMP work. Yet exhibitions and company announcements regularly show “two-inch,” “wafer-scale,” or “large-area” diamond.
The contradiction is only apparent. A sample that can be displayed is not necessarily a product that can be purchased, regrown, and supplied repeatedly.
“Large diamond” describes several different materials
Size alone hides the route and intended function.
- Large polycrystalline diamond can serve thermal, optical, or mechanical applications, but it is not a single-crystal growth seed.
- Smaller homoepitaxial single crystal is widely used in research, but common millimetre-scale formats are far from a non-mosaic two-inch plate.
- Mosaic single crystal joins or overgrows multiple smaller crystals. It expands area while retaining challenges around seams, misorientation, stress concentration, and boundary defects.
- Heteroepitaxial single crystal grows diamond on a non-diamond substrate and may bypass the initial seed-size limit. It must still manage lattice and thermal-expansion mismatch, stress, cracking, release, and subsequent regrowth.
- Non-mosaic, low-defect, regrowth-ready single crystal is the scarce category sought by many growers. Its surface, thickness, orientation, damage state, and stress must support further homoepitaxy.
Calling all five a “large diamond wafer” prevents a useful supply discussion.
A regrowth-ready seed is a capacity asset
A heat spreader or optical component is normally consumed in its intended application. A high-quality seed can be thickened, separated, cut, polished, and used to create additional seeds. If the buyer has the required process capability, one plate can become the starting point for a seed pool.
That gives mother-grade single crystal a strategic value beyond its material price. A supplier may reserve its best plates for internal replication, controlled partnerships, or downstream products rather than offer them as catalogue items.
This is a plausible explanation for limited availability, not a claim about every supplier’s policy. Yield, capacity, qualification, and customer priority can also restrict supply.
Between demonstration and delivery lies an entire process chain
One successful plate is a technical milestone. Repeated plates establish process stability. A commercial seed requires more:
- uniform growth and control of edge polycrystal;
- low and mapped defect density;
- stress and birefringence control;
- release and slicing without unacceptable damage;
- low-damage polishing, flatness, bow, and total thickness variation;
- documented orientation and surface condition;
- successful homoepitaxial regrowth;
- repeatability across lots and delivery dates.
If any one of those steps has poor yield, a visually impressive plate may still be unsuitable as a production seed.
Recent wafer-scale announcements are important—but specific
In June 2026, Element Six and Orbray said they had established a reproducible process for three-inch wafer-scale single-crystal diamond and were developing four-inch substrates. They also said two-inch wafers optimized for epitaxial applications were nearing finalization, while two-inch material for thermal bonding was being prepared for volume production.
Those are meaningful process and commercialization signals. They should not be read as proof that a buyer can already order any non-mosaic, regrowth-ready specification on an open-market basis. “Thermal bonding,” “epitaxial application,” “reproducible process,” and “volume production” describe different product gates.
Better questions than “Do you have two inches?”
When evaluating a large single-crystal claim, I would ask:
- Is it homoepitaxial, heteroepitaxial, or mosaic?
- Is it a display sample, device substrate, thermal plate, or growth seed?
- What area is usable after edge exclusion?
- What are the defect, stress, birefringence, bow, and TTV distributions?
- Has it survived release, slicing, and polishing?
- Has homoepitaxial regrowth been demonstrated across the plate?
- Is the specification orderable, repeatable, and supported by lead-time data?
The future competition is not merely a race to photograph the largest sample. It is a race to build a closed loop for mother-plate growth, replication, separation, finishing, inspection, regrowth validation, seed-pool expansion, and customer delivery.
The closer a large crystal comes to being a reproducible source of future capacity, the less likely it is to behave like an ordinary material sold from inventory.
Continue the research
Evidence limits and uncertainties
- Supplier willingness to sell regrowth-capable seed material is an industry interpretation; individual companies may have different technical, contractual, or capacity reasons.
- Orbray and Element Six describe targets and process milestones; these do not yet constitute open-market availability of every specification discussed here.