questions about UMF

[elenab]

Hello, 

The last few days I was reading about UMF online and I understood that H Seger categorized the ingredients into 3 types, fluxes, stabilizers, and glass formers and the mol.e. of the fluxes should be in total equal 1.

My question -  is it the only point of the UMF? I didn't understand how to deal with the numbers of the other two types of ingredients - stabilizers and glass formers. I didn't find any information about the correlation of the numbers of all three types.

If the numbers of stabilizers and glass formers are not in any relation with fluxes, does it mean that any recipe with fluxes mol. e. equaling 1 would be good? 

thank you

[Bill Kielb]

3 hours ago, elenab said:

If the numbers of stabilizers and glass formers are not in any relation with fluxes, does it mean that any recipe with fluxes mol. e. equaling 1 would be good? 

Nope, temperature and apportionment change. Under UMF fluxes add up to one by definition- always. So Glaze calculating software sums the fluxes to equal one and everything else is apportioned accordingly.  The fluxes traditionally have been Lio2, K20, MgO, CaO, SrO, BaO, ZnO. Alkaline metals and Alkaline earths respectively.

You end up with a consistent way to compare glass former's, stabilizers, and even ratios that become meaningful with testing. Setting fluxes equal to one (unity) is a requirement to place things in UMF form

[Min]

3 hours ago, elenab said:

If the numbers of stabilizers and glass formers are not in any relation with fluxes, does it mean that any recipe with fluxes mol. e. equaling 1 would be good? 

 

Nope again.

Any glaze calc, whether done by hand or with software does a Unity Molecular Formula (UMF) with the fluxes totalling 1.0 when looking at a formula in "unity". (you can also look at recipes in RO, R2O3 or non unity)

If you think of this in terms of baking it might be simpler. Say you are making really basic cookies and flour is the stabilizer (alumina in glazes), sugar is the glass former (silica and boron in glazes) and baking powder is the flux (calcium, magnesium, zinc etc in glazes). Say the total weight of the ingredients is 100 grams. Now if you mixed 1 gram of flour + 98 grams of sugar + 1 gram of baking powder its not going to make a decent cookie. Making glazes is very much about ratio's just as making cookies is. You need to have the appropriate ratio of ingredients to make the type of cookie or glaze you are looking for.

The ratio of fluxes to silica and alumina is very important and is partially dependent on the type of glaze you are looking for. Macro Crystalline glazes for example have very little alumina (stabilizer) in them. High gloss durable glazes should have high levels of both alumina and silica. Many matte glazes have high alumina in relationship to the silica level. Boron comes into play in low and midrange glazes so looking at the silica boron to alumina ratio can also be helpful.

Circling back to your question about how do you know how much silica and alumina should be in a glaze this is where targets can be useful. Glazy dropped using limit formulas in favour of using the Stull chart however it's good to have an idea of what targets to aim for or to be able to look at a glaze UMF and see straightway that a glaze is or isn't likely to be what your are seeking by this information. Test tiles are your best source of information but glaze calc can narrow down the field of possible glazes worth testing. Have a look at some glazes you have used that you like and note the silica and alumina levels plus the type of oxides used for the fluxes. After a while you will start to see patterns.

Screenshot below taken from this link with typical cone 6 target ranges. It isn't a be-all and end-all situation though, many of the more interesting glazes don't fall within these targets or limits but it's another tool in the toolbox for looking at glaze chem.

 

Edited May 21 by Min

[elenab]

39 minutes ago, Bill Kielb said:

Nope, temperature and apportionment change. Under UMF fluxes add up to one by definition- always. So Glaze calculating software sums the fluxes to equal one and everything else is apportioned accordingly.  The fluxes traditionally have been Lio2, K20, MgO, CaO, SrO, BaO, ZnO. Alkaline metals and Alkaline earths respectively.

You end up with a consistent way to compare glass former's, stabilizers, and even ratios that become meaningful with testing. Setting fluxes equal to one (unity) is a requirement to place things in UMF form

This! is what I just discovered 5 min ago when I downloaded an app for recipes Potter's Pal, it was showing that in any recipes fluxes are 1 in total, and I understood that as if it was # of fluxes in the formula and the app gets it to 1 proportionally. I didn't catch it before, I watched videos, read articles, etc... Wasn't sure, but you confirmed it! thank you! I have to process it now, then I come back.

[Bill Kielb]

1 minute ago, elenab said:

I have to process it now, then I come back.

It’s pretty easy to do by hand and if you are proficient with excel, there are some spreadsheets that relate it well. Regardless, knowing the concept then comparing glazes under UMF evens the playing field so to speak so one does not have to learn it to use it productively. The comparisons have more meaning and trends are more easily recognized under UMF.

[Dick White]

In the UMF calculation, the underlying notion is that you are counting the number of mols of each of the usual ceramic oxides, and developing a ratio of the number of each in a sample. A mol is simply a chemist's unit of counting. One can have a dozen pencils (12) or a ream of paper (500 sheets). A mol is ~6 x10^23 atoms or molecules (6 bazzilion). After doing all the math on the molecular weights of all the various oxides in a recipe and then reducing the sum of the flux molecules to 1 mol, the number of mols of alumina, silica, boron (the most important of the non-flux oxides) in the recipe sample will also be calculated. For the 1 mol of flux molecules, you can have innumerable ratios of the individual flux oxides within that 1 mol. For example, you might have .35 mols of sodium and .55 mols of calcium and .10 mols of magnesium oxides equaling the 1 mol of total fluxes. Once the 1 mol of fluxes has been set, the number of mols of the other oxides is revealed. For a typical cone 6 glaze, the quantity of alumina oxide will often fall between .25 and .5 mols, the quantity of silica oxide will often fall between 2.5 and 5 mols, and the quantity of boron oxide might be around .15 mols. These ranges are not required amounts, but rather are generally accepted targets for a stable glaze. There might be other reasons you want a glaze that is outside these typical ranges, but knowing how the ratios relate to each other will allow you to adjust a glaze recipe by adding or subtracting particular materials that have the oxides that you are trying to adjust. For example, if you want your recipe to have more sodium oxide in the 1 mol of fluxes, adding whiting (pure calcium carbonate) will not help. But if there is whiting in the recipe, you can subtract some of that to cause the balance of sodium within the 1 mol of flux to go up.

[Callie Beller Diesel]

I wanted Glazy’s target and solve for something, so I sprang for the membership for a year. The paid version lets you overlay the common limit formulas over the Stull chart. It also lets you overlay the Montmollin fuse charts, which is less helpful, but I digress. 

The limit formula overlays are a nice visual. If you’re not at the place where you want to pay a bunch of money for glaze software, I’m happy to screenshot here for educational purposes. 

[elenab]

Thank you, I understand it. Like Min said, it's very much like in recipes, it's all about proportions where the flux is 1. 

But now I have more questions, sorry for that! Dick said, "For a typical cone 6 glaze, the quantity of alumina oxide will often fall between .25 and .5 mols, the quantity of silica oxide will often fall between 2.5 and 5 mols" .  Now here is an example of a glaze.  https://help.glazy.org/downloads/pdf/Finkelnburg-NCECA 2012-PDF.pdf  Why do they say this glaze will melt at 5-10 cones? isn't it cone 3-4 according to their graph, and  what makes them say it will be durable mechanically andchemically:

[Bill Kielb]

I am thinking the confusion arises from the lack of explanation of cones. While their chemistry is established they indicate by bending so not really fully melted. A cone 4 cone ground up and fired to cone 8-10 turns out to be a fully melted glaze. At present there is not a straightforward exact predictor of when a composition will fully melt. Stull did his work at cone 10 I believe (likely closer to cone 11 today). His map was for a very specific flux ratio and his glazes. Katz showed a reasonable correlation at cone 6 and lower further incorporated boron as a means to lowering the melting temperature. Sue McCleod presented a cone six version of this at a2018 NCECA (Google Sue McCleod and Stull. Stulls map does not predict the temperature something melts at. Higher melting temperatures would progress from left to right and upward in relative ranges. Limits are these tested ranges, but are approximate guidance Testing is generally the only way to know with reasonable certainty when something will fully melt.

Advanced knowledge of durability is a Katz thing where certain flux ratios may be an indication of durability.

I think the simple answer is cones fully melt 4-5 cones higher than their rating. I don’t think the author clearly represented this in the cited link.

I would also suggest that much of pottery is understanding the variability and by experience important trends. It is somewhat very exact and scientific but verified through testing.

Edited May 23 by Bill Kielb

[Hulk]

The proposed glaze (slide 15, Finkelnburg's NCECA presentation):

0.3 R2O+0.7 RO : 0.35 Al2O3 : 3.5 SiO2
One mole of flux (combined):
     0.3 alkali oxides (R2O)
     0.7 alkaline earth oxides (RO)
0.35 moles of alumina (Al2O3)
3.5 moles of silica (SiO2)

As for the wide range, if the glaze is stiff enough at the higher cones (doesn't run off the ware) and doesn't exhibit other over fired problems, it's a go?
The durability, my understanding is that durability is largely a function of the silica and alumina amounts and ratio when fully melted.

The proposed glaze doesn't specify any oxides other than alumina and silica.
My guess would be that its behaviour would vary significantly depending on the RO and R2O selections and ratios.

Edit: The numbers across the top of the Stull graph are "ordinals" - not cones.
The graph is generated at a specified cone...
Each spot on the graph can be located by its letter(vertical) and number (horizontal), per the transcript of Stull's original presentation.

Edited May 23 by Hulk
My mistake

[Bill Kielb]

6 minutes ago, Hulk said:

 

It spots on the Stull graph just under cone 4.

Stulls graph does not show melting point in cones

Edited May 23 by Bill Kielb

[Min]

@elenab, the chart you posted in your post above isn't for glazes, it's the formulas for making cones. 

Edited May 23 by Min

[Callie Beller Diesel]

It’s also worth noting that there’s some wiggle room in how “melt” is defined. Materials fuze and then ooze over a range of temperatures. For instance, some glazes are matte if they’re underfired, but are properly a gloss with a bit more heat work. Yet others develop matte surfaces with additional heat work, because the assorted components need time to form a crystalline matrix. There can also be noticeable changes in colours or other surface qualities if a glaze is “over” or “under” fired. And what one person finds aesthetically pleasing someone else might hate. 

Leaving aside durability issues with the first example, the desired finish point of a glaze is determined in part by how you want the finished piece to look. 

[PeterH]

2 hours ago, Callie Beller Diesel said:

Leaving aside durability issues with the first example, the desired finish point of a glaze is determined in part by how you want the finished piece to look. 

... and some glazes can be influenced by crystallization during the cooling cycle (with little or no change in heat-work). 

Such as adding

Super Cool! Slow Cooling in an Electric Kiln
https://ceramicartsnetwork.org/daily/article/Super-Cool-Slow-Cooling-in-an-Electric-Kiln

Dixie Teal. Left: Fired on a medium speed cone 6 program, resulting in a glossier, darker color. Right: Fired to the cone 6 program on page 14, resulting in a lighter, speckled glaze

I suspect that:
- the UMF doesn't help a great deal here. 
- but UMF is almost always more informative than a glaze ingredient list
- except in the few cases where it isn't, e.g. the relative merits of cobalt oxide and carbonate as a colorant (same UMF but potentially different speckling).

[davidh4976]

Great discussion!

Is there any published information that compares the effectiveness of the fluxes? For instance, in adjusting the melt of a glaze, in fluxing power, is there data that says BaO equals x times K2O? Or something like that?

[Callie Beller Diesel]

Yes, but it can be a gross oversimplification to think of just swapping out one flux for another just to get silica and alumina to melt.

Info like that is mostly important for making sure you’re not using too much of a given thing, or making sure you’re using the right proportions of it to fire at the temperature you want. And even then, you have to consider that info in context of how it reacts with the other materials. They’re not interchangeable like that. 
 

Subbing Ba for K was probably an arbitrary choice, but I’ll use it anyways to show the kinds of considerations that also come into play, even though this is an unlikely scenario.
 

For starters, just on the numbers side, you’d be altering the flux ratio because the 2 are in different categories. This will alter things like the colour response and the fit of the glaze. And possibly some aspects of durability. 

There are also no sources of potassium that I can think of that are separate from at least a little bit of sodium. Most often it comes from feldspar, and if you remove the feldspar, you’re also taking out silica and alumina that have to be made up somewhere else in the formula. And the sodium, assuming you want to keep all other ratios the same. That material substitution is going to affect how this glaze behaves not only in the kiln, but in the bucket. How a glaze applies to a piece affects how it melts, and how it looks in the end. We’ve all seen glazes that look different when thin vs thick. Things like how the different sized or shapes of particles also affect how a glaze melts, and can affect things like micro bubbles or pinholes. Or lack thereof. 

Then you’ve got individual material properties. Ba used in excess (think 20% by weight) actually does have a very lovely and interesting colour response. When used like that, it makes matte glazes that break and pool 2 different colours (think pink and blue combos) in high fire reduction. However some of those glazes are so soft you can scratch them with a butter knife. Plus they leach. That’s an extreme example, and a few percent to compliment other fluxes still gives nice flow, can clear bubbles and still offer vibrant colours. 
 

If you’ve made it this far into my answer without nodding off, I believe there are some charts in Hamer and Hamer’s The Potter’s Dictionary of Materials and Techniques. It’s well worth the price tag if you want to get into more of the chemistry. 

[davidh4976]

On 5/25/2024 at 10:39 AM, Callie Beller Diesel said:

If you’ve made it this far into my answer without nodding off, I believe there are some charts in Hamer and Hamer’s The Potter’s Dictionary of Materials and Techniques. It’s well worth the price tag if you want to get into more of the chemistry. 

I've pulled my copy off the shelf and read the entries for each oxide and flux in general. It only makes me again realize how complicated some of this can be! Thanks for pointing in that direction.

[Min]

On 5/25/2024 at 9:39 AM, Callie Beller Diesel said:

There are also no sources of potassium that I can think of that are separate from at least a little bit of sodium.

psst, potassium carb aka pearl ash, but it's really only useful for frit makers as it's water soluble.

 

[Dick White]

And when diving deep into the UMF rabbit hole, several other things become apparent. Boron is not a flux. Fluxes create a eutectic with the silica and alumina to cause them to melt at lower temperatures. Boron is actually a stabilizer (R2O3, like alumina) but melts by itself at such a low temperature that it needs no assistance from a flux. It brings the other oxides into the melt simply because it has already melted. Also, the other fluxes each impart unique properties to the glaze aside from their melty-ness. Lithium is a more powerful flux, but too much messes with expansion (it can both craze and shiver). Shifting the balance of sodium vs. potassium can result in different color responses from some colorants (e.g., copper). So much of glaze chemistry cannot be demonstrated simply by the numbers - experience and testing is how we sort out many of the variables.

[PeterH]

10 hours ago, Dick White said:

And when diving deep into the UMF rabbit hole, several other things become apparent. Boron is not a flux.

Related references:
boron in glazes
https://ceramicartsnetwork.org/docs/default-source/uploadedfiles/wp-content/uploads/2008/10/tf-boroninglazes-0912.pdf
Glass Former or Flux?
Boron is very basic in function and it performs in a predictable manner. However, boron usage in the US for the last 50 plus years has been dominated by Gerstley borate, a raw material that is prized for its unpredictability. This has left at least a couple generations of ceramists with varying levels of confusion about the true nature of boron and a collection of recipes that don’t work without this particular material.
...
Boron and UMF
The function of silica, alumina, and fluxes are well understood in the unity molecular formula (UMF) or Seger formula, thanks to the work of R.T. Stull. However, boron, as a bit player, did not receive any attention in his work. All high-temperature glazes require silica, alumina, an alkaline earth, and an alkali. Boron is not required for a cone 10 glaze. At Alfred University, Dr. Bill Carty and I have put a lot of effort into defining how to best utilize boron from a UMF
perspective. We believe that one can predict the amount of boron needed to make a glaze melt based on the desired firing temperature.

And also ...
Reducing the Firing Temperature of a Glaze From Cone 10 to 6
https://www.digitalfire.com/article/reducing+the+firing+temperature+of+a+glaze+from+cone+10+to+6
Method 2: Add the Magic Oxide
 

 

[Hulk]

and

Borate (digitalfire.com)
Borates in Ceramics, Enamels, and Glazes | U.S. Borax