Vitreous China Glaze Batches: Why Your Third Container Won't Match the First
Three containers into a sanitary ware programme, the buyer sends me a photo. A close-up of two toilet cisterns, side by side, both 3/5.8L dual flush, same SKU, same PO template. One is a warm bone white, the other a slightly cooler, greener white. Same order. Two containers apart. The distributor in Düsseldorf has already mixed them on the showroom floor.
Nobody changed anything. That is the annoying part. The clay is the same, the mould is the same, the kiln is the same kiln. What changed is the glaze lot (釉料批次), the day the pieces went through the spray booth, and where the car sat inside the tunnel kiln.
White is a recipe, not a colour
A vitreous china glaze is roughly a mix of frit, quartz, feldspar, kaolin and zircon. The zircon is what makes it opaque and what makes it “white”. Change the frit supplier, change the zircon grade, or let the mill load a slightly different batch and the fired result shifts. On piece A it reads as 88.5 lightness on the colormeter, on piece B it reads 88.0. That is a ΔE of maybe 0.7 — nothing, you would say.
But put 120 toilets from car batch 1 next to 120 from car batch 2 in a 20 square metre bathroom with 4000K downlights and a ceramic tile floor, and 0.7 ΔE becomes visible on the edges. 1.5 ΔE becomes the thing the customer emails you about. In my experience anything over about 1.0 ΔE between containers will get flagged in a retail or hotel project, and over 2.0 it will get rejected outright if the specification is written properly.
The four things that actually move the colour
Glaze batch. A 20-tonne ball mill batch, properly dispersed, should be internally consistent. The problem is the transition between batches. Most factories do not drain a mill completely and make a fresh start; they top up. So the first 200 pieces of “new” glaze are actually 30% old and 70% new, or the other way around.
Spray thickness. This is the big one and everyone knows it and nobody measures it. A zircon opacified glaze at 0.35 mm versus 0.50 mm fires to visibly different whiteness because the light scattering changes. When I do a factory visit I ask for the wet film gauge reading sheet for the last week. About one factory in three can produce it. The ones that can are the ones I trust with a two-year colour commitment.
Kiln car position. Tunnel kilns are not isothermal. The top of a car sits 15-25 °C warmer than the bottom in many designs, and the side nearest the burners is hotter still. If the plant loads 12-piece cars and only the middle four positions get used for the key SKU, fine. If they load every position to hit an urgent order, the outer pieces fire slightly over and read warmer. That is where the matching problems in container three come from.
Firing cycle changes. Factories switch cycles for load reasons: pull a longer cycle for thick-wall pieces, shorten it for thin one-piece toilets. Cycle changes fire the same glaze differently. This happens most often after the Spring Festival restart, when the plant is ramping and pushing volume.
What to put in the PO instead of “white”
Do not write “glossy white”. It means nothing and the factory will agree to it. Write a master sample number and a tolerance against it. Something like:
- Reference standard: retained master sample plate, 100 × 100 mm, fired from production glaze, signed and dated by both parties, three per container ship date.
- Colour tolerance: ΔE ≤ 1.0 against the master under D65, measured on the vertical face of the bowl, three random pieces per pallet, minimum nine readings per container.
- Glaze film thickness: 0.35-0.50 mm wet, recorded per shift.
- All pieces of one SKU within a single container must come from cars fired on the same cycle.
- Rejection: any piece over ΔE 1.5 is treated as a non-conforming lot for the whole container, not per piece.
The last line matters. If the factory knows one bad car means a whole container claim, they will manage car loading properly. If the penalty is per-piece replacement, they will just ship and let you find 40 rejects at the destination.
The sample you keep is worth more than the spec
Ask the plant to fire six 100 × 100 mm glaze tiles from the actual production slip and spray line, not from the lab. Have them fired in the same kiln, on a normal production car, in a normal position. Sign them, date them, half for you, half for them. Pay for them if they ask, it will be under 30 USD.
Then, every six months, request a new set. Compare. If the drift is walking in one direction — say the supplier has moved to a cheaper zircon — you will see it coming three containers out instead of finding out in a distributor's warehouse.
One more thing. If you are buying a full bathroom package where the toilet, basin and bathtub all need to match, understand that the tub is almost certainly a different material (acrylic or steel enamel) and will never match perfectly. Set a realistic expectation there or your client will compare the tub to the cistern under a window and lose his mind. Nobody wins that conversation.