search  current discussion  categories  glazes - misc 

glaze question (mole percent, limit formulas)

updated fri 9 mar 01

 

Paul Lewing on mon 26 feb 01


John did, indeed, have some very good things to say about this subject,
but I'd like to comment on his assertion that the weight percentage
figure is useless.
I don't find this to be true. One of the things I most wanted
calculation software for was to predict, if I could, which glaze recipes
would work for me and which would not, without testing them. By "work"
I just mean that they will melt to a useable surface, not that they
would be beautiful glazes. Anyway, I find the weight % numbers to be as
useful for this as the unity formula ones. When I see a new recipe, I
compare it to several sets of parameters, (one of which is a standard
set of limit formulas) and one of those sets is the wt% for SiO2. I
find this to be a more accurate predicter than moles, in that I have
many recipes for which the silica is outside the limits in unity form,
but none that fall outside what I have observed to be the acceptable
parameters for wt%.
Also, whether or not the wt% numbers are useful, they are used by people
for several things. Tony H. pointed out that only potters use the flux
unity system. Most scientist, like geologists and chemists, use wt%.
So if you get an analysis for a material from a geologist, it will be in
wt% form. This is one of the big reasons why all calculation programs
(I think) let you change back and forth and allow you to input analyses
in either form.
Several calculation programs have numbers that attempt to predict the
surface or melt from some set of analysis numbers, and I believe they
use the wt% figures. I'm pretty sure the surface tension index in
Matrix and the viscosity index in Glaze Chem are both derived from them.
In both of those instances, the index numbers are not an actual
measurement of anything, but rather a ranking of the relative surface
tension or viscosity of the oxides. Lawrence E. and Bob W., am I right
about this? I've devised a crude viscosity number system for myself,
based on teh wt% #'s, and I find it to be very useful, though not
foolproof.
Probably everybody who has read this far in this ramble already knows
this, but I'm going to take a shot at explaining the difference between
molar percentage and flux unity. Someone please correct me if I'm
wrong, but this is how I understand it, in probably oversimplified
terms.
Both of these systems are basically counting molecules, they just
arrange them in different sets. Now, we can take any column of numbers
and express it as percentages. It's one of the basic math operations
you do in the Seger method. In the flux unity system, only the numbers
in the flux, or RO column, are made to add up to 1, and the numbers for
the other two are expressed in comparison to that number. In the molar
% system, all the numbers from all three columns are made to add up to
1. It's like "everything unity" instead of just flux unity. It makes
it easier to compare the amounts of oxides from different columns, but
I'm not sure it gives any really different information. On the other
hand, none of my programs do mole%, so I'm not used to working with it.
In my understanding, neither flux unity nor mole% weights the oxides for
fluxing power, it's just that mole% includes stuff like boron on a more
equal footing with the more normal fluxes. Again, someone please
correct me if I'm wrong.
It does seem to me that looking at both sets of numbers is a good idea,
just because they give a different picture of what's happening in a
glaze. To illustrate this, look at a high-lithium or a high-lead glaze
both ways. A glaze with .50 moles of Li, which would be a lot, might
only have 3% by weight of Li, because it's so light. Lead is the
opposite because it's so heavy. It's like if you have 3 ping-pong balls
and 3 bowling balls and you want to express how much you have of each.
Well, 50% of the number of balls is bowling ball, but maybe 99% of the
wieght is. Both statistics are useful, but for different things.
This is exactly the kind of thing I'm hoping will be discussed in my
breakout session at NCECA. I'm also hoping to shut up and listen to some
more knowledgeable people than I. See all you glaze-guru wannabes
there!
Paul Lewing, Seattle,
finally done with the Seattle Home Show- 88 hours of show in 9 days, but
I think I have a job again for the next year.

Tony Hansen on mon 26 feb 01


John gave some insightful views on this. It made me do some
thinking also.

>>I am working on a stain base glaze and trying to compare it
>>to Tony 's recipe in Magic Of Fire and I understand the need
>>for high CaO. It seems that the value of CaO in Tony's glaze
>>is over the limit tables.

The reasons for this are outlined on the web pages where
these base glazes are documented in detail.

>>I'm not sure if I see a way to connect the limits I followed
>>before to this new way, probably because there is no
>>connection. Is the "good" approach to balance the Dr Eppler
>>mole% with the limit tables. For example: use mole% to
>>target the cone target and Al:Si ratio, but keep the Li2O
>>within it's limits?

If you use Mole% you need to use Mole% targets (limits). If
you use unity formua, then use limits intended for unity
formulas. etc. There is not problem with using both, INSIGHT
will display a glaze is both formats.

>>One of the things that makes this glaze stuff so confusing
>>is I hear or read something from Ron then something else
>>from someone else that seems to conflict. When I suspect
>>that given the context they probably all agree, it's just up
>>to me to sort out the context.

This is an 'inexact science' and it seems that one advisor
will often say "yes, but you forgot about...". We are
learning. It seems that limit charts are not 'law', but
they are a convenient 'play pen' for beginnners.

>>I'm working on some glazes and have a question. I've been
>>studing the mole% along side the RO unity. If I'm reading
>>correctly this is the newer approach to limits.

>Mole percentages can be just as informative as unity
>formulas. However be absolutely sure you are looking a mole
>percentages. I know some of the glaze calculation
>programs...they don't make that clear. Eppler's preference
>for mole percentages stems from the fact that they are a
>more useful way to look at the composition for earthenware
>or low fire glazes.

INSIGHT 5.3 shows 'm%' after each number on reports. It just
shows '%' on the screen, but the calculation type picklist
shows "mole%". We will look at this again and see if it can
be made more clear. This is a complicated topic, but John is
right, Mole% is a reponse to the fact that fluxes vary in
'strength' and at lower and middle temperatures many
so-called fluxes are inert or weak. The seger method groups
them all together as fluxes, in many cases this just isn't
true.

The Seger unity model does not work well at lower
temperatures. Some oxides that are powerful fluxes at high
temperatures are refractory in low fire.

Oxides have a much more individual presence than the Seger
method tends to recognize. Their contributions to
particular properties often are not linear according to
concentration. Thus a more complex understanding of
concentration vs. effect is needed.

Oxide interplay producing characteristics attributable to
the group is not recognized by the Seger system.

Boron is both a glass and a flux and the logic for its
employment at various temperature ranges differs. It does
not 'plug into' a Seger formula very well.

Mole% is simply a calculation of the percentage of oxide
molecules by number (as opposed to an analysis which
compares their weights).

>Weight percentages are just about useless--in my opinion,
>worse than useless. they can really confuse you--when you
>are thinking about glaze stability. They can be helpful when
>you are looking at two glazes that are very similar and want
>to compare one to the other.

Seger formulas really shine when looking at one glaze in
isolation and comparing with limits. However we have been
rethinking the whole concept of limits (see article at
http://www.digitalfire.com/magic/limits.htm for example) & feel
that 'Targets' is a better word choice. We have also changed
the name of the Limits dialog to the Advisor dialog. This is
now intended to help users understand the oxide or material
secrets of any fired effect (not just the creation of a
durable and corrosion resistant glass). This dialog thus
supports target recipes in addition to formulas (eg. we have
included a few basic clay body targets). A crystalline glaze
at cone 6 would thus have a target recipe that considers
only visual appearance. A cone 6 functional ware liner glaze
would have a target recipe that focusses on durability and
leach resistance. That means there are a lot of target
recipes to be written and clearly documented. We are
creating the tools to do this. We feel that this whole idea
of targets will replace the trade in glaze recipes.

However, due to the problems mentioned above, many ceramic
engineers do infact use percentage analyses to formulate
glazes. Formulas don't clearly indicate how much there is of
an oxide, they show only relative amounts. However many
engineers want to think in terms of how much there is. For
example, they might be targetting a specific percentage of a
coloring or opacifying oxide, or they might be trying to
hold the amount of a given oxide below a certain percentage
to maintain a fragile mechanism. The ceramic materials
industry is built on percentage analyses, many salesman and
even technicians in industry have no idea what a unity
formula is. The page http://digitalfire.com/magic/basics.htm
explains the values of formulas and analyses (we have not
updated it yet to include Mole%).

Mole% ignores LOI as do formulas, it just looks at the
oxides that makeup the fired glass. The INSIGHT Advisor
dialog contains a few examples of target formulas from
Richard Eppler and references are based on Mole%. These will
give you a feel for how the system is used. However we need
to develop 'interpretation skills' using Mole%. If you don't
already know these it might be best to leave them alone for
now.

The INSIGHT Advisor dialog has target formulas, target Mole
Percents and even target recipes (target recipes are for
clay bodies). Thus if you want to make a particular glaze or
body, you start with the target and adjust as needed to
adapt it to your situation. The target should thus have
notes that explain the presence of each component and how
adjusting it will affect the fired product. Targets should
thus be 'middle-of-the-road' mixtures.

====================================================
T o n y H a n s e n thansen@digitalfire.com
D I G I T A L F I R E C O R P O R A T I O N
http://digitalfire.com Calculation/Database Software for Ceramic Industry

Ron Roy on tue 27 feb 01


Paul has this almost right,

Unity formula has fluxing adding to 1.0

% mols has the same mol count but all add to 100 - when anything adds to
100 we can read every part as percent.

The usefulness of mol % - some fluxes in our limits are not fluxes at lower
temps so the picture is clouded. With mol % you can count what is really
fluxing at 04 and compare to the other - refractories.

RR


>Both of these systems are basically counting molecules, they just
>arrange them in different sets. Now, we can take any column of numbers
>and express it as percentages. It's one of the basic math operations
>you do in the Seger method. In the flux unity system, only the numbers
>in the flux, or RO column, are made to add up to 1, and the numbers for
>the other two are expressed in comparison to that number. In the molar
>% system, all the numbers from all three columns are made to add up to
>1. It's like "everything unity" instead of just flux unity. It makes
>it easier to compare the amounts of oxides from different columns, but
>I'm not sure it gives any really different information. On the other
>hand, none of my programs do mole%, so I'm not used to working with it.
>In my understanding, neither flux unity nor mole% weights the oxides for
>fluxing power, it's just that mole% includes stuff like boron on a more
>equal footing with the more normal fluxes. Again, someone please

Ron Roy
RR# 4
15084 Little Lake Rd..
Brighton,
Ontario, Canada
KOK 1H0
Residence 613-475-9544
Studio 613-475-3715
Fax 613-475-3513

Barney Adams on tue 27 feb 01


I think what I am looking for with analysis is close to the same as Paul's. I want
to predict
if the soup will melt. From playing with the mole% I seem to be able to use more
CaO
and Strontium (wish I could get a good boron/strontium frit) and in turn get more
Al2O3
and SiO2. These two things are what I want in a glaze. If I'm getting the right
message
from all the guru's melting the maximum amount of SiO2 and Al2O3 that gives me the
ratio for the type of glaze I want is the first priority. If I have it right Boron
seems to
fill in on the SiO2 part for forming glass and helps melt the Al2O3? Then What ever
I
come up with has to fit my clay bodies expansion. PieceOCake. I'm ramping up for
some
glaze tests to see how the mole% vs limits melt. Insight is nice, but you need to
mix up glazes
to see what it means.

The problem with both limits and mole% guidelines is there are many sets of glaze
formulas
that don't seem to be fit within their consideration. Many glazes seem to dance
around the edges of
the guidlines (Shinos, stoney matts, crystal). I guess that's where testing comes
into play.

What's the best way to test durability. Scratch at the glaze with some steel tool
and examine under a loop?
That's how I've been doing it.

Barney

John Hesselberth on tue 27 feb 01


Ron Roy wrote:

>The usefulness of mol % - some fluxes in our limits are not fluxes at lower
>temps so the picture is clouded. With mol % you can count what is really
>fluxing at 04 and compare to the other - refractories.


This is a good point. If I could redesign things my way, I would have a
"relative fluxing power" for each flux at each of 3 temperatures--say 06,
6 and 10. So, for example, calcia might be 0.2 at cone 06, 1.0 at cone 6
and 1.2 at cone 10. Soda might be 1.0 at cone 06, 0.6 at cone 6 and 0.5
at cone 10. I'm just making numbers up, but you get the idea. Calcia is
a more effective flux at high temperatures; soda is more effective at low
tempertatures. Then if you multiplied "relative fluxing power" times mole
percent for each flux and added them up, it might tell you how much
silica + alumuna that flux combination would melt at each temperature.
Then Paul would be happy and I would be too.

Ahhhh, so many experiments to run; so little time.

John

"The life so short, the craft so long to learn." Hippocrates, 5th cent.
B.C.

Paul Lewing on tue 27 feb 01


Ron Roy wrote:
>
> Paul has this almost right,
Thanks, Ron. I almost appreciate your support.


> The usefulness of mol % - some fluxes in our limits are not fluxes at lower
> temps so the picture is clouded. With mol % you can count what is really
> fluxing at 04 and compare to the other - refractories.
Some programs (Insight for sure, and maybe others) allow you to decide
which oxides are participating in unity and which are not. This is
mostly to allow you to include boron in the fluxes if you wish. But
wouldn't it serve the same purpose to just not include in unity the
oxides that are not fluxes at that low temperature, e.g. Mg? Perhaps if
you were doing low-fire you'd include boron in unity, but not magnesium.
Although I can see where that would be much clearer in mole% form.

Paul Lewing, Seattle

Lawrence Ewing on wed 28 feb 01


Hi Paul,

>Several calculation programs have numbers that attempt to predict the
>surface or melt from some set of analysis numbers, and I believe they
>use the wt% figures. I'm pretty sure the surface tension index in
>Matrix and the viscosity index in Glaze Chem are both derived from them.
>In both of those instances, the index numbers are not an actual
>Measurement of anything, but rather a ranking of the relative surface
>tension or viscosity of the oxides. Lawrence E. and Bob W., am I right
about this?

Yes Matrix calculates its surface tension using wt% figures figures.

Wt% values are interesting particularly in relation to SiO2 and Al2O3. some
years ago when working with ian Currie's first book 'Stoneware Glazes - a
Systematic Approach' I plotted the corner glazes of all of his Biaxial grids
on 2 Al Si graphs using on the one hand the unity formula values and on the
other Wt% values. The results were really interesting.

Cheers,

Lawrence Ewing

Senior Lecturer
Ceramics Department
School of Art
Otago Polytechnic
Dunedin
New Zealand

email: lewing@clear.net.nz

phone +64 03 472 8801

MATRIX GLAZE CALCULATION SOFTWARE:
http://www.Matrix2000.co.nz

GLAZETEACH:
http://www.Matrix2000.co.nz/GlazeTeach

MATRIX TUTORIALS:
http://www.Matrix2000.co.nz/MatrixTutorials

MATRIX ADDITIONAL MATERIALS RESOURCE:
http://www.Matrix2000.co.nz/MatrialsWeb/default.htm

Paul Lewing on thu 1 mar 01


John Hesselberth wrote:
If I could redesign things my way, I would have a
> "relative fluxing power" for each flux at each of 3 temperatures--say 06,
> 6 and 10. So, for example, calcia might be 0.2 at cone 06, 1.0 at cone 6
> and 1.2 at cone 10. Soda might be 1.0 at cone 06, 0.6 at cone 6 and 0.5
> at cone 10. I'm just making numbers up, but you get the idea. Calcia is
> a more effective flux at high temperatures; soda is more effective at low
> tempertatures. Then if you multiplied "relative fluxing power" times mole
> percent for each flux and added them up, it might tell you how much
> silica + alumuna that flux combination would melt at each temperature.
> Then Paul would be happy and I would be too.

John, no reason why we can't both be happy. This is the idea behind the
viscosity number I've worked out for myself, and also the viscosity
number in Glaze Chem and the surface tension number in Matrix. They
start with the rankings in Hamer & Hamer for those subjects, then (in
the case of viscosity, eg) assign a 1 to the least active flux, a 2 to
the next least active, and the highest number to the most active flux.
These numbers are then multiplied by the value produced by the analysis,
then added up to produce a viscosity index. However, both Glaze Chem
and Martix use the wt% figures. It would be interesting to see if using
the mole% figures gave a more accurate prediction of melt.
Rick Malmgren once told me of a professor in Finland who had worked out
a very sophisticated system for doing this. Instead of just assigning a
whole number to each flux based on a ranking, he had worked out a real
relative fluxing power based on melting temperature. So instead of the
least active flux being a 1 and the next a 2, the next one might be a
1.8, or a 3.2 or something. He also did it by assigning a positive
number to the fluxes and a negative number to the alumina, and a 0 to
the silica. And I don't know whether he used wt% or mole%. I didn't
investigate further because he wanted money for the disc that contained
the spreadsheet, which I would have done, except it was PC and I have a
Mac.
In his book, Cullen Parmalee says that such a system can never be truly
accurate because it doesn't take into account particle size of materials
or the fact that getting an oxide from one material may produce a more
active melt than getting it from another.
Happy now?
Paul Lewing, Seattle,
Shaken but not stirred. I ran for the backyard during yesterday's
earthquake and the ground felt like pudding shaking. No damage, though.
I would imagin there was a whole lot of crashing at Seattle Pottery
Supply, though. It's in the area of town that is built on fill.

Ron Roy on sun 4 mar 01


Paul makes a good point here - yes it would be appropriate to eliminate the
middle temp fluxes if you are working on glazes below 1100C.

It is also true that - no matter what system you use - you will eventually
be able to forecast results - simply because you will learn from your
experiments.

Percent (not mole %) is also useful - and I do use it almost every day. %
of Boron is useful when trying to predict fit - over 12% and things change.
Trying to duplicate a clay body and sourcing iron for instance - from other
materials - % is more useful than mols.

I see no reason to have to favour one system over the other - perhaps it
appears to be too complicated for those starting out and perhaps that is a
consideration but I say use as much as you can and you will be better off
for it.

RR


>> The usefulness of mol % - some fluxes in our limits are not fluxes at lower
>> temps so the picture is clouded. With mol % you can count what is really
>> fluxing at 04 and compare to the other - refractories.
>Some programs (Insight for sure, and maybe others) allow you to decide
>which oxides are participating in unity and which are not. This is
>mostly to allow you to include boron in the fluxes if you wish. But
>wouldn't it serve the same purpose to just not include in unity the
>oxides that are not fluxes at that low temperature, e.g. Mg? Perhaps if
>you were doing low-fire you'd include boron in unity, but not magnesium.
>Although I can see where that would be much clearer in mole% form.
>
>Paul Lewing, Seattle
>
>______________________________________________________________________________
>Send postings to clayart@lsv.ceramics.org
>
>You may look at the archives for the list or change your subscription
>settings from http://www.ceramics.org/clayart/
>
>Moderator of the list is Mel Jacobson who may be reached at melpots@pclink.com.

Ron Roy
RR# 4
15084 Little Lake Rd..
Brighton,
Ontario, Canada
KOK 1H0
Residence 613-475-9544
Studio 613-475-3715
Fax 613-475-3513

Hank Murrow on tue 6 mar 01


John wrote;
>
>Hank, I'm not completely following you. How would ionic potential explain the
>fact that calcium is a very active flux at cone 6 and above, but not a
>flux at all at cone 06. On the other hand sodium and potassium are
>active at all temperatures. And of course lead is, in a way, the
>opposite of calcium. It works well at low temperatures and begins to
>lose its effectiveness at high temperatures because it vaporizes. If
>ionic potential were related in some way to fluxing power it would have
>to be adjusted for temperature.


John;

I don't have an answer for that. As I said, I only work at C/10 in
OX and RED, but I hope if some of us try to work the Ionic Potential into
our mental models, we may make additional progress towards a more complete
understanding of the melt. I welcome any insights generated along these
lines.
Regards, Hank

Hank Murrow on tue 6 mar 01


>John Hesselberth wrote:
> If I could redesign things my way, I would have a
>> "relative fluxing power" for each flux at each of 3 temperatures--say 06,
>> 6 and 10. So, for example, calcia might be 0.2 at cone 06, 1.0 at cone 6
>> and 1.2 at cone 10. Soda might be 1.0 at cone 06, 0.6 at cone 6 and 0.5
>> at cone 10. I'm just making numbers up, but you get the idea. Calcia is
>> a more effective flux at high temperatures; soda is more effective at low
>> tempertatures. Then if you multiplied "relative fluxing power" times mole
>> percent for each flux and added them up, it might tell you how much
>> silica + alumuna that flux combination would melt at each temperature.
>> Then Paul would be happy and I would be too.

AndPaul replied:

>John, no reason why we can't both be happy. This is the idea behind the
>viscosity number I've worked out for myself, and also the viscosity
>number in Glaze Chem and the surface tension number in Matrix. They
>start with the rankings in Hamer & Hamer for those subjects, then (in
>the case of viscosity, eg) assign a 1 to the least active flux, a 2 to
>the next least active, and the highest number to the most active flux.
>These numbers are then multiplied by the value produced by the analysis,
>then added up to produce a viscosity index. However, both Glaze Chem
>and Martix use the wt% figures. It would be interesting to see if using
>the mole% figures gave a more accurate prediction of melt.

And Hank wonders;

Why don't we use the Ionic potential as an indicator of relative
fluxing power? On page 302 in Cardew's, "Pioneer Pottery" there is an
illuminating exposition on Ionic Potential and its relation to fluxing
power. Linus pauling had the original idea of dividing the valency of an
ion by its radius to give what he called the Ionic Potential. The lower the
number for Ionic Potential, the higher the fluxing power. In addition,
elements having Ionic Potentials greater than around 7 are shown to be
glass formers. Here are the common elements used in ceramic glazes and
their corresponding values for Ionic Potential:

Potassium_______ 0.75
Sodium__________ 1.00
Lithium_________ 1.30
Barium__________ 1.40
Strontium_______ 1.80
Calcium_________ 2.00
Zinc____________ 2.40
Reduced Iron____ 2.40 *Note that reduced iron is a flux, while
Magnesium_______ 2.60 oxidized iron acts more like alumina,
Oxidized Iron___ 4.50 zirconium, tin, and Titanium.
Zirconium_______ 4.60
Alumina_________ 5.30
Tin_____________ 5.40
Titanium________ 5.80
Silicon_________ 10.25
Phosphorus______ 14.30
Boron___________ 15.00

Glaze experimenters who use the percent method of testing may feel
that Lithium is a more aggressive flux than Potassium because this method
does not recognize mole weight; while those who use the mole method will
see that Ionic Potential calibrates their melting power pretty accurately.
Those wanting to explore this possibility should look up Linus Pauling's
book, "The Nature of the Chemical Bond", which details the reasoning for
his now completely accepted notions. David Stannard first introduced me to
these concepts in the mid sixties, and they have proven very fruitfull in
forming an accurate view of the melt, though I confess that I work only at
C/10. I offer this to open the discussion to those interested.

Cheers! Hank in Eugene

John Hesselberth on tue 6 mar 01


Hank Murrow wrote:

> Glaze experimenters who use the percent method of testing may feel
>that Lithium is a more aggressive flux than Potassium because this method
>does not recognize mole weight; while those who use the mole method will
>see that Ionic Potential calibrates their melting power pretty accurately.
>Those wanting to explore this possibility should look up Linus Pauling's
>book, "The Nature of the Chemical Bond", which details the reasoning for
>his now completely accepted notions. David Stannard first introduced me to
>these concepts in the mid sixties, and they have proven very fruitfull in
>forming an accurate view of the melt, though I confess that I work only at
>C/10. I offer this to open the discussion to those interested.

Hi Hank,

I'm not completely following you. How would ionic potential explain the
fact that calcium is a very active flux at cone 6 and above, but not a
flux at all at cone 06. On the other hand sodium and potassium are
active at all temperatures. And of course lead is, in a way, the
opposite of calcium. It works well at low temperatures and begins to
lose its effectiveness at high temperatures because it vaporizes. If
ionic potential were related in some way to fluxing power it would have
to be adjusted for temperature.

Regards, John

"The life so short, the craft so long to learn." Chaucer's translation of
Hippocrates, 5th cent. B.C.

Paul Lewing on wed 7 mar 01


Interesting, Hank!
Three comments: first, where would lead be?
Second, looks like, once again, boron doesn't fit the system.
Three, why don't you make up a spreadsheet or just do the math and see
how accurate a predicter of melt this is. Try not just your usable
matte and gloss glazes, but some that don't melt at all, and some that
run right off the pots.
I'd be real interested to see how well it worked.
Paul Lewing, Seattle