search  current discussion  categories  glazes - chemistry

## expansion values

### Tom Buck on thu 22 dec 05

Carole & Rosanne
I note at start that you both asked practically the same
question on the same day, so here are my comments, suitable for you both:

When you do glaze calculations using a well-known program, you will arrive
at a value for the linear coefficient of expansion (COE, otherwise
called coefficient of thermal expansion (CTE).

Some authors cite the COE (CTE) in millionths of a length change
per original length per degree temperature. this will appear as
micrometres/metre, or microinches/inch, or similar length value changes
per degree C (or degre F). The temperature scale chosen does matter. If I
say the COE/CTE value is 7.0 x 10-6 I am using the Celsius (Centigrade)
scale. If I were to use the Fahrenheit scale (as many still do), then I
would divide the 7.0 by 1.8 because the Fahrenheit degree is smaller than
the Celsius degree (hence the F number for the same temperature is
bigger). So in F degrees the 7.0 x 10-6 becomes 3.9 x 10-6.

Finally, some authors use 10-7 values and others use 10-8 value
for example, Ron Roy uses 10-8/degree Fahrenheit when he cites a COE
value. and also, Ron likes COE values from UK for each ingredient, whereas
others, Insight for example, use COE value cited in USA documents. so
there ends up a bit of differences in the numbers for COE but not
significantly different in my view.

Most calculation programs will automatically calculate the COE/CTE
for you as you enter the amounts of each ingredient into the recipe list.

Let us take Rosanne's glaze recipe, namely Ron&John's "Raspberry"
(Page 102 first edition, Mastering Cone 6 Glazes).

Recipe: Raspberry Cone 6 Oxidation
20.0 Whiting
18.0 Nepheline syenite
14.0 Ferro Frit 3134
18.0 OM#4 ball clay
30.0 Flint/silica/quartz

(then the two colourants are cited: Green chrome oxide 0.2 and Tin Oxide
7.5 - but since these ingredeints do not have a significant effect on
COE/CTE, they are ignored by glzcalc programs in arriving at the COE/CTE.

when R&J performed the analysis (using GlazeMaster), they obtained
the following Seger (Flux Unity) Formula:

CaO/MgO 0.80 B2O3 0.15 SiO2 3.03
K2O/Na2O 0.20 Al2O3 0.29 TiO2 0.009
Fe2O3 0.005

The Si/Al ratio = 10.6 to 1 (if B2O3 is taken as a glass-former too, the
ratio becomes 11.0)

R&J did not cite the COE/CTE for this mixture, but they did say
that "this glaze has the highest expansion/contraction characteristics of
any of our glazes..." and warned that it might craze in time on some C6
claybodies.

I did the calculation using Insight, once with UK values and
Fahrenheit, and got 484 x 10-8, and once with USA values and Celsius and
got 7.27 x 10-6 (which equals 404 x 10-8 in Fahrenheit values if I did a
simply division by 1.8; this indicates the differences that can show up in
different programs).

Since I am more familiar with USA/oC values, I will work with
those. Cone 6 clay bodies tend to have an expansion/contraction value of
7.0 x 10-6, and if the glaze COE is slightly above that value, then a good
fit will occur. But there are porcellaineous stoneware bodies for Cone 6
that will have a COE of 6.5 x 10-6 or a bit above that. And that is what
Rosanne is using; so yes, she will find that Raspberry glaze will
craze/crack on this clay body.... 7.3 - 6.5 = 0.8 x 10-6 is too much of a
gap between body and glaze. a fit is likely when the spread is below 0.25
x10-6.

The same is true for Carole Fox's special blue glaze:
30 Neph Sy; 36 FF3134; 14 flint/silica; 20 EPK; 2.0 Iron Oxide Red; 4.0
rutile; and 0.12 Cobalt oxide. when the COE is calculated, I get this
Seger
CaO/MgO 0.52 B2O3 0.46 SiO2 3.76
K2O/Na2O 0.48 Al2O3 0.56 TiO2 0.19
Fe2O3 0.05
Ratio SiO2/Al2O3 = 6.8 COE = 8.0 x 10-6

This special blue is "iffy" under best conditions because the KNaO is too
high (making for good melting but high expansion); and the Al2O3 is way
too high which would make for a very stiff/matt glaze at best. The B2O3 is
there to try and make everything happy with one another. But despite the
B2O3 being in the right range (below 12 mole percent), it is unable to
harmonize the glaze. so, back to the COE for both glazes, they are too
high. what can one do? revise the recipe so to seek a suitable COE.

I spent sometime trying to make a new recipe for Carole but the
four ingredients simply will not meld, and even when I change two of them,
the COE remains too high. Perhaps more effort is needed there. But I
suggest Carole accept Rosanne's new mix, change the colourants, and give
it a try:

Raspberry Cone 6 Oxidation (revised)
22.0 G-200 feldspar
13.0 FF3134
10.5 Whiting
8.5 Wollastonite
14.0 OM#4 ball clay
5.0 EPK
27.0 flint/silica/quartz

add colourants. I will let Carole do the glzcalc and then tell us what the
new ratio is, and what the COE/CTE is expected to be on firing.

and I shall await comments from other glaze calculators,
especially Ron.

good pots. peace Tom.

Tom Buck ) -- primary address.
tel: 905-389-2339 (westend Lake Ontario, province of Ontario, Canada).

### Edouard Bastarache Inc. on fri 23 dec 05

Tom,

according to Smart, there are 4 methods to calculate the CTE (COE),
he prefers Appen's, check this out:

http://www.flickr.com/photos/potier/46606941/

"Here is the chart of the various methods of calculation (Appen, Gan Fu-Si,
Takahashi, Winkelmann and Edouard's). A priori, the data obtained by
Edouard are identical to the method of Winkelmann, but the data of
Winkelmann are valid only between 20 and 100°C, while all the others
are set for the range between 20 to 400°C, which accounts better for the
T.C.E.(CTE) for ceramists or glassmakers."

Appen's method is the more complicated, there are oxides for which
you need a special calculation method to get the "alpha" coefficient.
They are: boron, silica, titanium dioxide, lead oxide.
You may see that if you go to Smart's site. Even if the article is in
French,
numbers can be understood by everyone here, more especially by a
chemical engineer (Hehehehehe)

Now, you see why I have already recommended to the list
to at least calculate the COE of 1 glaze before using a computer
to do it.

Later,

"Ils sont fous ces quebecois"
"They are insane these quebekers"
"Están locos estos quebequeses"
Edouard Bastarache
Irreductible Quebecois
Indomitable Quebeker
Sorel-Tracy
Quebec
edouardb@sorel-tracy.qc.ca
www.sorel-tracy.qc.ca/~edouardb/Welcome.html
http://retrodemonstration.blogspot.com/
http://www.flickr.com/photos/potier/
http://stainlessfre.blogspot.com/

----- Original Message -----
From: "Tom Buck"
To:
Sent: Thursday, December 22, 2005 11:03 PM
Subject: Re: expansion values

> Carole & Rosanne
> I note at start that you both asked practically the same
> question on the same day, so here are my comments, suitable for you both:
>
> When you do glaze calculations using a well-known program, you will arrive
> at a value for the linear coefficient of expansion (COE, otherwise
> called coefficient of thermal expansion (CTE).
>
> Some authors cite the COE (CTE) in millionths of a length change
> per original length per degree temperature. this will appear as
> micrometres/metre, or microinches/inch, or similar length value changes
> per degree C (or degre F). The temperature scale chosen does matter. If I
> say the COE/CTE value is 7.0 x 10-6 I am using the Celsius (Centigrade)
> scale. If I were to use the Fahrenheit scale (as many still do), then I
> would divide the 7.0 by 1.8 because the Fahrenheit degree is smaller than
> the Celsius degree (hence the F number for the same temperature is
> bigger). So in F degrees the 7.0 x 10-6 becomes 3.9 x 10-6.
>
> Finally, some authors use 10-7 values and others use 10-8 value
> for example, Ron Roy uses 10-8/degree Fahrenheit when he cites a COE
> value. and also, Ron likes COE values from UK for each ingredient, whereas
> others, Insight for example, use COE value cited in USA documents. so
> there ends up a bit of differences in the numbers for COE but not
> significantly different in my view.
>
> Most calculation programs will automatically calculate the COE/CTE
> for you as you enter the amounts of each ingredient into the recipe list.
>
> Let us take Rosanne's glaze recipe, namely Ron&John's "Raspberry"
> (Page 102 first edition, Mastering Cone 6 Glazes).
>
> Recipe: Raspberry Cone 6 Oxidation
> 20.0 Whiting
> 18.0 Nepheline syenite
> 14.0 Ferro Frit 3134
> 18.0 OM#4 ball clay
> 30.0 Flint/silica/quartz
>
> (then the two colourants are cited: Green chrome oxide 0.2 and Tin Oxide
> 7.5 - but since these ingredeints do not have a significant effect on
> COE/CTE, they are ignored by glzcalc programs in arriving at the COE/CTE.
>
> when R&J performed the analysis (using GlazeMaster), they obtained
> the following Seger (Flux Unity) Formula:
>
> CaO/MgO 0.80 B2O3 0.15 SiO2 3.03
> K2O/Na2O 0.20 Al2O3 0.29 TiO2 0.009
> Fe2O3 0.005
>
> The Si/Al ratio = 10.6 to 1 (if B2O3 is taken as a glass-former too, the
> ratio becomes 11.0)
>
> R&J did not cite the COE/CTE for this mixture, but they did say
> that "this glaze has the highest expansion/contraction characteristics of
> any of our glazes..." and warned that it might craze in time on some C6
> claybodies.
>
> I did the calculation using Insight, once with UK values and
> Fahrenheit, and got 484 x 10-8, and once with USA values and Celsius and
> got 7.27 x 10-6 (which equals 404 x 10-8 in Fahrenheit values if I did a
> simply division by 1.8; this indicates the differences that can show up in
> different programs).
>
> Since I am more familiar with USA/oC values, I will work with
> those. Cone 6 clay bodies tend to have an expansion/contraction value of
> 7.0 x 10-6, and if the glaze COE is slightly above that value, then a good
> fit will occur. But there are porcellaineous stoneware bodies for Cone 6
> that will have a COE of 6.5 x 10-6 or a bit above that. And that is what
> Rosanne is using; so yes, she will find that Raspberry glaze will
> craze/crack on this clay body.... 7.3 - 6.5 = 0.8 x 10-6 is too much of a
> gap between body and glaze. a fit is likely when the spread is below 0.25
> x10-6.
>
> The same is true for Carole Fox's special blue glaze:
> 30 Neph Sy; 36 FF3134; 14 flint/silica; 20 EPK; 2.0 Iron Oxide Red; 4.0
> rutile; and 0.12 Cobalt oxide. when the COE is calculated, I get this
> Seger
> CaO/MgO 0.52 B2O3 0.46 SiO2 3.76
> K2O/Na2O 0.48 Al2O3 0.56 TiO2 0.19
> Fe2O3 0.05
> Ratio SiO2/Al2O3 = 6.8 COE = 8.0 x 10-6
>
> This special blue is "iffy" under best conditions because the KNaO is too
> high (making for good melting but high expansion); and the Al2O3 is way
> too high which would make for a very stiff/matt glaze at best. The B2O3 is
> there to try and make everything happy with one another. But despite the
> B2O3 being in the right range (below 12 mole percent), it is unable to
> harmonize the glaze. so, back to the COE for both glazes, they are too
> high. what can one do? revise the recipe so to seek a suitable COE.
>
> I spent sometime trying to make a new recipe for Carole but the
> four ingredients simply will not meld, and even when I change two of them,
> the COE remains too high. Perhaps more effort is needed there. But I
> suggest Carole accept Rosanne's new mix, change the colourants, and give
> it a try:
>
> Raspberry Cone 6 Oxidation (revised)
> 22.0 G-200 feldspar
> 13.0 FF3134
> 10.5 Whiting
> 8.5 Wollastonite
> 14.0 OM#4 ball clay
> 5.0 EPK
> 27.0 flint/silica/quartz
>
> add colourants. I will let Carole do the glzcalc and then tell us what the
> new ratio is, and what the COE/CTE is expected to be on firing.
>
> and I shall await comments from other glaze calculators,
> especially Ron.
>
> good pots. peace Tom.
>
> Tom Buck ) -- primary address.
> tel: 905-389-2339 (westend Lake Ontario, province of Ontario, Canada).
> mailing address: 373 East 43rd Street, Hamilton ON L8T 3E1 Canada
>
> ______________________________________________________________________________
> 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
>

### Ron Roy on wed 28 dec 05

Hi Tom - good work - always a good sign when you show up with your
experience and knowledge.

I would have subbed in some Spodumene to get the expansion down to 425 - my
normal response to this type of problem - thinking it will give the most
accurate reproduction of the glaze under most conditions.

My second approach would be to increase the Boron at the expense of KNaO -
Boron having the lowest expansion (below 12%) and KNaO having the highest
expansion. I'm not as sure about how the 2nd method would affect certain
colours but it would be viable with some for sure.

As in most of my recommendations - I prefer to give two revisions so they
can be used in a line blend and can help establish a working calculated
expansion number for the clays they are tested on. In this case - because I
believe the solution is not far away - using the original as part of the
line blend.

By the way - when reporting crazing it is helpful to describe what it looks
like and when it happens - which will help with the solution of the
problem.

If the glazes is already finely crazed when unloaded then a more drastic
solution is called for.
If the craze pattern is large then it will be easier to correct.
If no craze when unloaded but crazes in a day - then the solution will be
even easier.

So in this case I would do the revision with Spodumene.

Whiting - 20.5
Spodumene - 9.0 (I calculated with Aussi Spod)
Neph Sy - 6.0
F3134 - 14.0
OM#4 - 19.5
Silica - 31.0
Total - 100.0
Ratio - 10,58 (original is 10.53)
Expansion - 418 (org is 483)

So line blending 500 grams of the original with 500 grams of the spod
revision will show where the crazing stops - drop down to the next lowest
expansion test to allow for delayed crazing - and test that version keeping
in mind that clay/glaze fit problems are more likely to show up on larger
work.

In this case - chrome/tin reds are touchy - and revisions may not get the
desired colour - just part of the deal with some types of glazes. Again - a
line blend will be helpful is assessing just when the colour starts to
change.

RR

RR

> I spent sometime trying to make a new recipe for Carole but the
>four ingredients simply will not meld, and even when I change two of them,
>the COE remains too high. Perhaps more effort is needed there. But I
>suggest Carole accept Rosanne's new mix, change the colourants, and give
>it a try:
>
>Raspberry Cone 6 Oxidation (revised)
> 22.0 G-200 feldspar
> 13.0 FF3134
> 10.5 Whiting
> 8.5 Wollastonite
> 14.0 OM#4 ball clay
> 5.0 EPK
> 27.0 flint/silica/quartz
>
>add colourants. I will let Carole do the glzcalc and then tell us what the
>new ratio is, and what the COE/CTE is expected to be on firing.
>
> and I shall await comments from other glaze calculators,
>especially Ron.
>
>good pots. peace Tom.

Ron Roy
RR#4