Silicon Wafer Pricing: 8 Cost Drivers and How to Optimise

1. Why no supplier publishes a wafer price list

Silicon wafers are not standardised the way catalogue parts are. A wafer is described by at least eight independent parameters, and each one affects manufacturing cost. Two orders both reading "4-inch silicon wafer" can be entirely different products in cost terms.

On top of that, semiconductor material prices move with industry cycles and with the supply situation of each substrate type. A published price list would be out of date within weeks. The industry standard is therefore send specifications → receive a quote, and this article helps you send specifications that produce a quote matching your real needs.

2. Eight factors that drive wafer price

2.1. Substrate material — the biggest factor

This variable dominates all the others. Silicon is the cheapest material thanks to a mature manufacturing ecosystem and abundant feedstock. Compound semiconductors cost considerably more because crystal growth is harder, slower and yields fewer usable wafers.

2.2. Diameter

Price does not scale linearly with area — it moves in steps. For silicon, the mainstream 4-inch and 6-inch diameters usually offer the best cost per unit area because the lines run at volume. Two-inch sounds cheap but has poor cost per cm², while 8-inch and 12-inch demand higher-grade equipment and tighter control, so they cost noticeably more.

For SiC and InP, larger diameters are scarce and several times more expensive because crystal yield drops sharply with size. Digifund supplies SiC at 2", 3", 4", 6" with an 8" option, and InP at 2", 3" and 4".

2.3. Grade: Prime, Test, Dummy

This is the largest saving most buyers overlook. The three grades differ in surface defect density, flatness and inspection level — not in the underlying material.

• Prime: tightest tolerances, for real device fabrication and high-resolution lithography steps.
• Test: relaxed tolerances, for tool calibration, process trials and repeatability checks.
• Dummy: filler wafers in furnaces, cassette spacers, student handling practice.

2.4. Polishing: single-side or double-side

Double-side polished (DSP) costs more than single-side (SSP) because it adds a full chemical-mechanical polishing step. DSP is only necessary when you genuinely process both faces — through-wafer MEMS structures, transmission optics, or when overall flatness must be very high. For most processes that only touch the top surface, SSP is the correct choice.

2.5. Resistivity and how tight the tolerance is

It is not the resistivity value that drives cost but how narrow the tolerance band is. A request for "1–10 Ω·cm" is easy and cheap to meet, since most of a sliced ingot falls in that range. A request for "4.5–5.5 Ω·cm" forces the maker to sort and discard many wafers, pushing the price up sharply.

The standard bands Digifund stocks for silicon are 0.001–0.005, 0.01–0.02, 1–10 and 1–100 Ω·cm. Picking one of these instead of specifying a custom tolerance saves noticeably on both cost and lead time.

2.6. Thickness and crystal orientation

Standard thicknesses for silicon — 275, 300, 365, 500, 525 and 725 µm — map to each diameter by industry convention and are always cheaper than custom thicknesses. Unusually thin wafers cost more because breakage during processing rises.

On orientation, ⟨100⟩ is the most common and cheapest. ⟨111⟩ is widely used in surface science and epitaxy, while ⟨110⟩ is rarer and usually carries both a premium and a longer lead time.

2.7. Order quantity

Wafers ship in standard 25-piece cassettes. Ordering a full cassette always beats ordering loose pieces, because vacuum packing, freight and import handling are spread across more wafers. If you need 18, ask for the price of 25 before deciding — the gap is usually smaller than expected.

2.8. Lead time

Stock items are always cheaper than made-to-order. If your schedule allows waiting, say so when requesting a quote — the supplier may be able to fold your order into a shared production batch at a meaningfully better price.

3. Five ways to optimise your wafer budget

1. Split demand by project phase. Dummy for handling practice, Test for calibration and trials, Prime only for real device runs. This is the single biggest lever.
2. Stay on standard specifications. Choose stocked resistivity bands, thicknesses and orientations rather than custom ones. Reserve custom for applications that genuinely require it.
3. Only choose DSP if you process both sides. Re-check your process flow — many orders specify DSP out of habit when SSP would do.
4. Size to your equipment, not your wishes. A 6-inch wafer will not fit a 4-inch chuck. Conversely, buying 2-inch when your tools accept 4-inch inflates your cost per unit area.
5. Consolidate annually. If your lab consumes steadily, combining a year of demand into one or two orders instead of monthly dribs cuts both unit price and freight.

4. Specification checklist for an accurate quote

Send all of the items below and you will get an accurate quote on the first pass instead of several rounds of back-and-forth:

5. Four common mistakes when buying wafers

• Ordering Prime for everything. The most expensive mistake and also the most common, especially in newly established labs.
• Comparing price per wafer alone. A cheap wafer with the wrong orientation or resistivity band is money entirely wasted. Compare on identical specification sets.
• Ignoring packaging and shipping conditions. Polished wafers need cleanroom vacuum packing. Saving here usually ends in surface contamination and a ruined lot.
• Not asking about certificates. For research you intend to publish or production that needs traceability, request the lot inspection report when you ask for the quote, not when the goods arrive.

Frequently asked questions