Guides / Glaze chemistry 101

Glaze chemistry 101: UMF for working potters
If you've Googled "UMF" you've probably bounced off a wall of percentages. Here's what it actually does — and why it's the most useful tool you'll ever ignore.
A glaze recipe written as percentages — 25% feldspar, 20% silica, 15% kaolin… — looks like a cooking recipe. It isn't. Two recipes with very different percentage lists can fire to nearly identical surfaces if their oxide chemistry is similar. And two recipes with almost identical percentages can fire completely differently if you swap a flint for a frit, or a Custer for a G-200.
Percentages describe what you put in the bucket. UMF describes what ends up in the glass. That's the entire pitch.
What UMF actually is
UMF stands for Unity Molecular Formula. It's a way of expressing a glaze recipe as molar ratios of oxides, normalized so the fluxes add up to 1.0 (the "unity").
That's the whole math, and you don't even need to do it. Modern tools (kilnmuse, Glazy, Insight) compute UMF automatically from your raw materials and their published chemical analyses. What you actually need is to read the result. The numbers come out in three families.
The three oxide families
Every oxide in glaze chemistry plays one of three roles. Hermann Seger formalized this grouping in the 1880s, and 140 years later it still works:

Fluxes
What melts the glaze
The lowest-melting oxides — they soften and dissolve the others. K₂O, Na₂O, CaO, MgO, ZnO, Li₂O, BaO, PbO. In UMF they sum to 1.0 by convention.
Examples: Feldspars (KNaO via K₂O+Na₂O), whiting (CaO), dolomite (CaO+MgO), zinc oxide (ZnO), wood ash (mostly K₂O+CaO)
Stabilizers
What holds the glaze on the piece
Refractory oxides that resist the melt and give the glaze structure. Almost always Al₂O₃, occasionally B₂O₃. Without enough, the glaze runs off; with too much, it stays unfused.
Examples: Kaolins (EPK, Calcined), ball clays, alumina hydrate. Boron contributes via frits and gerstley borate.
Glass formers
What turns into the glassy structure on cooling
Almost entirely SiO₂ (silica/quartz). The actual glass network. Typical glazes have 2–6× more SiO₂ than Al₂O₃.
Examples: Flint, silica sand (325 mesh), quartz. Frits also contribute pre-melted SiO₂.
The two ratios that matter
You don't need to memorize all the oxides. Two ratios tell you nearly everything about a glaze's structural behavior:

SiO₂ : Al₂O₃
silica / alumina
Typical range: 3 : 1 to 10 : 1 (typical glazes)
The classic one. Low ratio (3–5) means matte and refractory; high ratio (8–10) means glossy and runny. The Stull chart plots glazes on this axis specifically. See our Stull chart guide for what the zones mean visually.
R₂O : RO
alkali / alkaline-earth
Typical range: 0.3 : 0.7 to 0.5 : 0.5 (durable glazes)
Tells you how the fluxes are distributed. Alkali-heavy (K₂O, Na₂O dominant) glazes are bright but expand a lot — they craze. Alkaline- earth-heavy (CaO, MgO dominant) glazes are tougher and fit clay bodies better. A 0.4 / 0.6 split is a common target for foodware.
A worked example
Here's a simple cone-6 satin recipe by percentages:

Recipe (by weight)
- Custer Feldspar40%
- Whiting (CaCO₃)14%
- EPK Kaolin20%
- Silica 32520%
- Talc6%
- Total100%
Run it through UMF and you get something like:
Unity Molecular Formula
- ROK₂O0.18
- RONa₂O0.04
- ROCaO0.62
- ROMgO0.16
- R₂O₃Al₂O₃0.42
- RO₂SiO₂3.18
Two ratios pop out:
· SiO₂ : Al₂O₃ = 3.18 : 0.42 ≈ 7.6 : 1 — comfortably in the bright/satin band on a Stull chart. Should fire glossy-ish to satin at cone 6.
· R₂O : RO = 0.22 : 0.78— strongly alkaline-earth biased (the whiting + talc dominate). Good for durability and crazing resistance, but might come out a touch matte from the CaO+MgO loading. That tracks with "satin" in the recipe name.
From those two numbers alone, an experienced glaze chemist could predict the surface, the food-safety profile, and the likely failure modes — without ever firing a tile. That's the point of UMF.
What UMF doesn't tell you
UMF is structural chemistry. It says nothing about color (you can have a copper red, a celadon, and a tenmoku at the exact same UMF — the difference is iron and copper loading on top). It says nothing about specific surface effectslike crystalline growth, metallic luster, or shino flashing — those depend on cooling schedule, atmosphere, and minor minerals that don't show up at the oxide level.
And critically: UMF assumes your published material analyses are accurate. They're close — but every batch of feldspar, every kaolin source, every frit lot has small variations. A cautious potter still test-tiles. UMF tells you whether the recipe is structurally sound; the kiln tells you whether your batch is.
Why this matters in practice
The day-to-day value of UMF isn't writing recipes from scratch — most potters work from established recipes. It's substituting safely. You ran out of Custer; can you use G-200? UMF answers in seconds: G-200 has slightly more Na₂O and less K₂O, so you'd expect very slight surface differences but no structural change. Substitution OK, watch for craze creep.
Or: your supplier dropped Gerstley Borate. UMF lets you build a replacement (frit + whiting + kaolin) that matches the original oxide profile within ±10%, which is the realistic tolerance for ceramic chemistry.
kilnmuse runs UMF on every recipe it generates, and the Seger UMF card on a recipe page shows you the full breakdown — both ratios, all oxides, with a flag if anything sits outside the cone's limit formula.
Try it
Generate a recipe and see its full UMF breakdown in seconds.
Upload an inspiration photo, and kilnmuse formulates the recipe, calculates UMF + the SiO₂:Al₂O₃ and R₂O:RO ratios, plots it on a Stull chart, and runs Limit Formulas for your target cone. Free preview, no signup needed for the first try.
Try kilnmuse→References
- Hermann Seger, collected works on glaze chemistry (1880s) — the foundational classification of fluxes, stabilizers, and glass formers.
- Glazy.org — modern UMF calculator and oxide reference data (CC BY-NC-SA 4.0).
- Digitalfire — material chemistry database and limit formula data by cone.