Yantai Likai Semiconductor Technology Co., Ltd. completed line qualification for 6-inch sapphire substrates and 310 × 310 mm sheets in the fourth quarter of 2025, with volume scheduling from this month. The company, established in October 2025, specialises in precision slicing of semiconductor materials and had previously shipped mainly 2- and 4-inch formats.
What drove the change was not an equipment list but a shift in the customer mix. Customers patterning substrates (PSS) are moving to larger panels, while optics customers want more usable area from a single piece. Both point at the same problem: usable area per piece is too small, while handling and mounting events stay the same.
A question that comes up often: with larger ingots available, why not go straight to 8 inch? The company's answer is that slicing itself is not the barrier — the downstream chain is. Epitaxy, lithography and PSS patterning lines have not yet caught up with 8-inch formats, and pushing a larger diameter first would only shift the tolerance burden onto customers. Published specifications currently stop at 6 inch, while ingot handling keeps 12-inch capability.
What changed on the line
The first change is ingot mounting. On smaller diameters the fixture holds the ingot with its own weight plus adhesive; at 6 inch the ingot weight and cutting reaction both step up, and once face run-out drifts the whole batch inherits a thickness bias after slicing. The new mounting procedure writes adhesive thickness and cure time into the process card, and face run-out is measured after mounting, before any cut is committed.
The second change is the X-ray orientation fixture. Orientation must be confirmed before slicing, and at a larger diameter the same angular error translates into a bigger edge displacement, so centring and angular repeatability both had to be re-qualified. The reworked fixture cuts the reference flat in the same setup that measures orientation, so customers can reproduce the orientation the same way instead of inferring it from a mark.
The third change is the polishing carrier. With more area, pressure uniformity directly sets TTV and flatness, so the carrier grooves, mounting method and pressure distribution were all revised. There is no shortcut here: mount the slices, polish, demount, measure, change one parameter, repeat — until the thickness distribution across the carrier settles.
Qualification followed the same route as any new part: cut, lap, polish, measure, in a loop. First the kerf and damage layer were checked against the expected range, then the thickness distribution after lapping, and finally TTV, flatness and surface roughness after polishing. Volume production was not released until the numbers settled — a rule written plainly into the internal process card.
What the customer receives
Published values are unchanged: C-plane (0001) offcut within ±2°, thickness 700 / 1000 μm with ±3 μm tolerance, polished Ra ≤ 0.3 μm, and transmission above 70 % at 355 nm and 1064 nm. A larger format that holds a looser tolerance is not progress, so the expansion left these numbers alone.
The 310 × 310 mm sheet holds an edge tolerance of ±0.2 mm, with corners and edges chamfered to drawing, so it can enter a customer's existing mounting and exposure flow for PSS patterning. For optics customers, one sheet yields more — and larger — usable windows, which reduces edge waste.
Inspection is piece by piece: orientation by X-ray, thickness at five points with a micrometer, TTV and flatness from the gauge, Ra for roughness, transmission for optical grades, and a full visual check under controlled lighting. The data sheet travels with each batch, and the batch number traces back to a specific ingot.
For downstream users, the practical value of the expansion is fewer handling events. For the same usable area, 6-inch wafers and 310 × 310 mm sheets mean fewer mounts, fewer cleans and fewer transfers — steps that cost both time and yield inside a cleanroom. Feedback from one customer after trials was that larger piece area visibly reduced manual intervention in the front-end process.
On capacity: 6-inch wafers and sheets share the same multi-wire slicing line and the same lapping and polishing tooling. Scheduling is grouped by thickness and surface requirement, and lead times for standard formats are quoted in weeks. Customers who need a trial slice can start with a small batch and fix volume parameters once the measured data is in hand.
For new customers the recommendation is a small trial batch first: material taken from the same ingot, cut, lapped and polished to the target thickness and surface, with thickness distribution, TTV, Ra and transmission reported before volume parameters are discussed. The same route applies to 6-inch wafers and 310 × 310 mm sheets; a trial slice usually ships within two weeks, and customers can take it into their own epitaxy or patterning runs.
The two customer groups look at the same sheet rather differently. PSS customers care about surface defect density and uniformity after patterning; optics customers care more about transmission and wavefront quality. Inspection therefore reports results by intended use, so one set of metrics is never used to judge two very different applications.
On delivery, every batch ships with a data sheet and batch record, double-bagged in the controlled area, and retained samples can be supplied if a customer wants third-party re-inspection. Record-keeping follows the parent company system, so the question of which ingot a given slice came from always has an answer.
Take the 310 × 310 mm sheet as an example: the same piece carries more patterned units on a PSS line, which cuts the number of exposure and development changeovers. For a line manager this gain does not show up in any single process parameter but in line tact time and the number of manual interventions — the kind of improvement you notice only in use.
After-sales and technical support follow the parent company model: process questions go straight to engineers, with on-site review of slices and parameters when needed. For a process-driven product like slicing, a datasheet alone rarely solves the actual problem — a point the company keeps making internally.
On quality control, the expansion added no new inspection items; instead the existing ones moved from sampling to piece-by-piece. Thickness, TTV, flatness, roughness and transmission are recorded per piece, with a full visual check under controlled lighting, and the batch number traces which ingot, which cut and which polishing carrier a slice came from.
The qualification also drew on the parent company's multi-wire cutting experience: wire selection, tension curves and coolant strategy were already documented, saving the time of starting from scratch. Equipment experience does not transfer directly to material, however — the numbers still had to be re-established in this plant's own lapping and polishing sections.
Across the industry, sapphire demand is spreading from lighting into consumer electronics and automotive optics: covers for watch and phone lenses, protective windows for vehicle cameras, all trading larger formats for lower cost per part. For a slicing house, size and tolerance have to move together; getting bigger without getting more precise means nothing.
Next, the company plans to grow sheet size further and to qualify thinner formats — again gated by downstream readiness and measured data rather than numbers on a datasheet. Until then, 6 inch and 310 × 310 mm mark the boundary of what this plant commits to; anything beyond it is stated plainly at the enquiry stage.
Back to news