A 12-Inch Diamond Claim Is Not the Same as Manufacturing Maturity

An industry item recently claimed that a company had achieved mass-production capability for 12-inch CVD diamond. My first reaction was enthusiasm. Anyone who has spent years around MPCVD wants to see credible signs that the field is moving forward.

My second reaction was more cautious: does a larger growth area, a higher-power reactor, or a more impressive demonstration necessarily mean that manufacturing maturity has improved?

The original item is not identified in the Chinese article, and I could not independently confirm its source during this review. The 12-inch statement should therefore be read as the prompt for this analysis—not as a verified production milestone.

A sample proves possibility; production proves control

A laboratory result can establish that a material can be made. Production has to establish that it can be made repeatedly, across long campaigns, tools, operators, and material lots, while remaining inside an agreed specification.

That difference is easy to underestimate in MPCVD. A large plate or a strong short-duration run is visible and easy to communicate. The less visible questions determine whether customers can actually use the output:

• Does the process remain stable over consecutive batches?
• Can two reactors produce equivalent material?
• Are edge-to-center differences characterized and controlled?
• Are recipes, maintenance changes, and deviations traceable?
• Can the supplier grade the output consistently and reproduce qualification results?
• When a lot fails, is there enough evidence to identify and correct the cause?

A reactor that can run is not yet a production system that can deliver.

The industry’s three roles see different problems

One reason progress is uneven is that universities, equipment suppliers, and material producers are optimized for different outcomes.

Universities are often strongest at mechanisms, characterization, and demonstrating a new process window. That work is essential, but an excellent paper or record sample does not automatically answer whether 100 or 500 consecutive parts can remain within specification.

Equipment companies are strongest at plasma platforms, fields, hardware interfaces, and application engineering. Their natural milestone is a machine that operates and reproduces a demonstration. The material producer must answer a different question: can the resulting lot be accepted, delivered, and supported when something goes wrong?

Material producers sit between those groups and the customer. They carry the consequences of yield, grading, delivery, qualification, cash flow, and warranty responsibility. In my experience, many problems described as isolated technical limitations are actually failures to connect these three roles into one manufacturing system.

What the SiC transition can teach MPCVD

Silicon carbide is not a direct technical template for CVD diamond. Its useful lesson is organizational. Industrialization required the sector to move beyond record samples and nominal dimensions toward defect maps, specifications, qualification systems, statistical process control, fault detection and classification, and tool-to-tool matching.

The breakthrough was not simply learning how to make a better crystal. It was learning how to place crystal growth inside a controlled manufacturing order.

For MPCVD, a similar transition would require:

• platform and chamber-to-chamber consistency;
• recipes managed as controlled production assets;
• standardized reference samples and characterization protocols;
• statistical process control rather than end-point inspection alone;
• fault detection and classification tied to corrective action;
• change control for hardware, consumables, maintenance, and software;
• stable grading and a reproducible qualification evidence chain.

These capabilities rarely make headlines, but they distinguish an industrial product from a sequence of successful samples.

Large-area growth still matters

None of this makes large-area progress unimportant. Uniform plasma over a larger area, thermal-field control, edge management, and substrate handling are real engineering achievements. Reactor scale may also improve the economics of some products.

The mistake is to let area stand in for all the evidence that is missing. “We can grow it large” and “we can deliver it reliably” are two different statements. The first is a platform milestone. The second requires yield data, repeatability, acceptance criteria, qualification, and customer history.

My working test for a manufacturing claim is therefore simple: after the headline demonstration, what remains true across hundreds of pieces, multiple tools, controlled changes, and repeated customer acceptance?

The next milestone is manufacturing discipline

MPCVD does not need less ambition. It needs a more demanding definition of progress. Larger chambers and higher power deserve attention, but the industry’s durable value will come from making growth stable, measurable, transferable, and certifiable.

That is the shift from knowing how to grow diamond to knowing how to manufacture and deliver it.

Continue the research

• How MPCVD growth works
• Why substrate stress must be evaluated as a production problem
• Explore Wonsen’s research areas