search  current discussion  categories  glazes - crazing & crackle 

high silica crazing glazes

updated mon 31 jan 00

 

David Hewitt on wed 26 jan 00

Some recent tests on a cone 6 glaze in oxidation, involving an increase
in quartz, have surprised us in that they showed increased crazing while
conventional calculations of expansion showed a reduction in the
expansion coefficient.
We believe that this is due to the silica not being absorbed totally
into the glaze melt and that the excess changes into cristobalite.
Comments on this would be much appreciated.
Also does this suggest that expansion calculations are not valid when an
over saturation of a glaze with silica applies?

Glaze BG1
Soda feldspar 45.0
China clay 6.0
Bentonite 2.0
Whiting 17.0
Zinc oxide 5.0
Quartz 25.0

This is a shiny, well balanced glaze. Presumably all the silica has been
taken into the glaze melt. It did not craze on stoneware but did show
some crazing on porcelain.

Our analysis of this glaze is as follows:-

K2O .052 Al2O3 .337 SiO2 3.109
Na2O .178 Fe2O3 .004 TiO2 trace
CaO .574
MgO .005
ZnO .191

Ratio Al2O3/SiO2 1/3.11

5 x 10% increments of quartz were
added to this so that the final glaze equalled.

Soda feldspar 45.0
China clay 6.0
Bentonite 2.0
Whiting 17.0
Zinc oxide 5.0
Quartz 75.0

K2O .052 Al2O3 .337 SiO2 5.695
Na2o .178 Fe2O3 .004 TiO2 trace
CaO .574
MgO .005
ZnO .192

Ratio Al2O3/SiO2 1/16.9

The Linear Coefficient of Expansion ( English & Turner x10-6/oC)for
these 6 glazes is in sequence

BG1 = 5.01, 4.56, 4.20, 3.90, 3.64, 3.42.

We were very surprised to see that the degree of crazing increased as
the expansion coefficient went down. The last has the lines so close
together that it looks like a high alkaline raku glaze!!!

Our theory is that the excess silica, that cannot get into the melt,
converts to cristobalite. This will in fact do its beta - alpha
conversion of around 2% shrinkage at 220 deg C. The result is a
shrinkage of the glaze and crazing results from the fact that there is
now more cristobalite in the glaze than in the body. This is something
that we have seen in other tests with high silica too but I've never
done such a direct test to show it up. I'd assumed there was some
explanation such as: Well perhaps that particular glaze had a lot of
alkali in it.

The glaze was next re-fired to cone 9. The amount of crazing may be
lessened slightly by this but nothing very obvious. The overall picture
of increased crazing with increased silica still holds.

--
David Hewitt
David Hewitt Pottery ,
7 Fairfield Road, Caerleon, Newport,
South Wales, NP18 3DQ, UK. Tel:- +44 (0) 1633 420647
FAX:- +44 (0) 870 1617274
Own Web site http://www.dhpot.demon.co.uk
IMC Web site http://digitalfire.com/education/people/hewitt.htm

Lawrence Ewing on wed 26 jan 00

Hi David,

I believe your deductions are correct. The clues are in the Silica/Alumina
ratios. The ratio for the original glaze was 8.32 (for the materials I used
in Matrix calculations) This increased to 15.45 for the last glaze in the
line. At cone 6 the likely ratio for a shiny clear glaze would be around 8
depending on what the flux group looked like. Clear shiny glazes indicate a
melt with no excess oxides remaining undissolved. Of course the overall
levels of both Al2O3 and SiO2 are a factor as well e.g. it is possible to
have a ratio of say 8 but have an overall level of Al2O3 and SiO2 so
excessive that all of these oxides are not dissolved. The opposite can also
occur where these is excessive flux oxides in relation to the levels of
Al2O3 and SiO2.

My understanding of the issue is that the further away from a 'balanced'
glaze (clear - shiny) one moves the less likely the actual Coefficient of
Expansion is to be an accurate indicator of crazing. What it does do for us
is to indicate which direction our changes are taking us. If we want to
reduce crazing in a glaze which has a high SiO2/Al2O3 ratio then we should
perhaps be reducing the SiO2 in order to bring the glaze back into balance
while keeping an eye on the levels of Al2O3 and SiO2 as indicated by the
Limits for that temperature.

Regards,

Lawrence Ewing

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

email: lewing@clear.net.nz

MATRIX GLAZE CALCULATION SOFTWARE:
http://home.clear.net.nz/pages/Matrix2000

GLAZETEACH:
http://home.clear.net.nz/pages/GlazeTeach

MATRIX TUTORIALS:
http://home.clear.net.nz/pages/lewing/MatrixTutorials

MATRIX ADDITIONAL MATERIALS RESOURCE:
http://home.clear.net.nz/pages/Matrix2000/MatrialsWeb/default.htm

-----Original Message-----
From: Ceramic Arts Discussion List [mailto:CLAYART@LSV.UKY.EDU]On Behalf
Of David Hewitt
Sent: Thursday, January 27, 2000 9:51 AM
To: CLAYART@LSV.UKY.EDU
Subject: HIGH SILICA CRAZING GLAZES


----------------------------Original message----------------------------
Some recent tests on a cone 6 glaze in oxidation, involving an increase
in quartz, have surprised us in that they showed increased crazing while
conventional calculations of expansion showed a reduction in the
expansion coefficient.
We believe that this is due to the silica not being absorbed totally
into the glaze melt and that the excess changes into cristobalite.
Comments on this would be much appreciated.
Also does this suggest that expansion calculations are not valid when an
over saturation of a glaze with silica applies?

Glaze BG1
Soda feldspar 45.0
China clay 6.0
Bentonite 2.0
Whiting 17.0
Zinc oxide 5.0
Quartz 25.0

This is a shiny, well balanced glaze. Presumably all the silica has been
taken into the glaze melt. It did not craze on stoneware but did show
some crazing on porcelain.

Our analysis of this glaze is as follows:-

K2O .052 Al2O3 .337 SiO2 3.109
Na2O .178 Fe2O3 .004 TiO2 trace
CaO .574
MgO .005
ZnO .191

Ratio Al2O3/SiO2 1/3.11

5 x 10% increments of quartz were
added to this so that the final glaze equalled.

Soda feldspar 45.0
China clay 6.0
Bentonite 2.0
Whiting 17.0
Zinc oxide 5.0
Quartz 75.0

K2O .052 Al2O3 .337 SiO2 5.695
Na2o .178 Fe2O3 .004 TiO2 trace
CaO .574
MgO .005
ZnO .192

Ratio Al2O3/SiO2 1/16.9

The Linear Coefficient of Expansion ( English & Turner x10-6/oC)for
these 6 glazes is in sequence

BG1 = 5.01, 4.56, 4.20, 3.90, 3.64, 3.42.

We were very surprised to see that the degree of crazing increased as
the expansion coefficient went down. The last has the lines so close
together that it looks like a high alkaline raku glaze!!!

Our theory is that the excess silica, that cannot get into the melt,
converts to cristobalite. This will in fact do its beta - alpha
conversion of around 2% shrinkage at 220 deg C. The result is a
shrinkage of the glaze and crazing results from the fact that there is
now more cristobalite in the glaze than in the body. This is something
that we have seen in other tests with high silica too but I've never
done such a direct test to show it up. I'd assumed there was some
explanation such as: Well perhaps that particular glaze had a lot of
alkali in it.

The glaze was next re-fired to cone 9. The amount of crazing may be
lessened slightly by this but nothing very obvious. The overall picture
of increased crazing with increased silica still holds.

--
David Hewitt
David Hewitt Pottery ,
7 Fairfield Road, Caerleon, Newport,
South Wales, NP18 3DQ, UK. Tel:- +44 (0) 1633 420647
FAX:- +44 (0) 870 1617274
Own Web site http://www.dhpot.demon.co.uk
IMC Web site http://digitalfire.com/education/people/hewitt.htm

Cindy Strnad on thu 27 jan 00

David,

After trying out your glaze in Insight, I think your expansion is still too
high for porcelain. It figures out to about what I aim for in stoneware.

Cindy Strnad
Earthen Vessels Pottery
Custer, SD

Frank Bosco on thu 27 jan 00

Your problem could be due to the fact that 'fused silica' in the amorphous
state has very little coefficient of expansion. However, if you still have
crystalline quartz undesolved in your glaze, then I suppose you could run
into trouble. For cristobolite to form takes a good deal of soaking above
2100 F.

Jeff Lawrence on fri 28 jan 00

> Frank Bosco wrote:
>Your problem could be due to the fact that 'fused silica' in the amorphous
>state has very little coefficient of expansion. However, if you still have
>crystalline quartz undesolved in your glaze, then I suppose you could run
>into trouble. For cristobolite to form takes a good deal of soaking above
>2100 F.

Hello David, Lawrence, Cindy and Frank,

I've observed the same crazing David noted in a glaze made mainly (70%) with
local volcanic tuff. This base material is about 75% amorphous silica

This glaze (no recipe on hand, sorry) crazed slightly. To cure this, I added
5% and 10% silica and, as in David's test, made the crazing more intense. I
was firing only to cone 6 over about 12 hours, conditions under which I
understand that little cristobalite will form. I, too, tried to cure the
crazing by higher firing -- to cones 8, 9, and 10 -- with no perceptible
change in the crazing. The clay was a cone 6 whitish stoneware.

If I test, I'll post the details. I'll try tinkering with the alumina:silica
ratio.

Interesting stuff. THanks for posting it.

Best,
Jeff

Jeffrey Lawrence 505 753 5913
Sun Dagger Design 505 753 8074 fax
Route 3 Box 220 (mile marker 185 on 285) jml@sundagger.com
Espanola, NM 87532

Frank Bosco on fri 28 jan 00

Then again, there's always the 'magic bullet' boron which can always be
counted on to produce a melt plus reduce crazing taboot!

Frank Bosco
Professor of Ceramic Art
Rhode Island School of Design

Michael Banks on fri 28 jan 00

David,

Cristobalite could be the source of the increased crazing, but incompletely
absorbed quartz is just as likely, or maybe more likely. Your glaze must
have some undissolved quartz grains present which are initiating crazing by
the abrupt beta/alpha contraction at 573 degrees C. Undissolved small
quartz nuclei may be concentrated at the base of the glaze layer (where they
do the most damage) and none present at the top surface. This will give the
appearance of a glossy mature glaze.

You don't mention the mesh size of the added quartz. Try 400 mesh or a
brand of amorphous silica and see if the crazing reduces. This would be
diagnostic of dissloution problems.

Michael Banks,
Nelson,
New Zealand

Janet Kaiser on sun 30 jan 00

I know I am exceedingly dim, but I DO TRY to learn from this sort of
discussion... But I do not understand what folk are saying about this.

It was my understanding that the most important characteristics of various
silica formations, is the thermal expansion of the various modifications and
the volume changes which take place at different temperatures. It is between
220 and 250 degrees Celsius that a sudden expansion takes place whether
quartz, tridymite or cristobalite. This is caused by a conversion of the
silica crystal into another crystal having the same chemical composition but
a larger volume / smaller density.
Another important conversion takes place at 600 deg. C.

Conversions happen slowly if the temperature rise or fall takes place
slowly. If there are large silica grains and the temperature changes
abruptly then the volume changes so fast so it causes stresses and
strains...

Surely the crazing is therefore due to either the speed it is fired or large
silica grains? Or is this just me being simplistic again?

Janet Kaiser
The Chapel of Art: Home of The International Potters' Path
Criccieth LL52 0EA, GB-Wales, UK
WEBSITE: http://www.the-coa.org.uk
EMAIL: postbox@the-coa.org.uk
We have server problems. If you cannot access
The CoA web page, please try again later. Thanks!


----------------------------Original message----------------------------
> Frank Bosco wrote:
>Your problem could be due to the fact that 'fused silica' in the amorphous
>state has very little coefficient of expansion. However, if you still have
>crystalline quartz undesolved in your glaze, then I suppose you could run
>into trouble. For cristobolite to form takes a good deal of soaking above
>2100 F.

Ron Roy on sun 30 jan 00

Hi David,

I am away from home till Feb 11 and do not have access to my
dilatometer charts till then - but I do know I have a chart of a
glaze with some free quartz in it - I know that because the quartz
inversion shows up. That would explain the crazing you are seeing. I
also know you will never see any cristobalite in a glaze - there is
is just too much flux around for that to happen.

Ill check this out when I get back and compare the numbers I have
with your recipe.

RR



>----------------------------Original message----------------------------
>Some recent tests on a cone 6 glaze in oxidation, involving an increase
>in quartz, have surprised us in that they showed increased crazing while
>conventional calculations of expansion showed a reduction in the
>expansion coefficient.
>We believe that this is due to the silica not being absorbed totally
>into the glaze melt and that the excess changes into cristobalite.
>Comments on this would be much appreciated.
>Also does this suggest that expansion calculations are not valid when an
>over saturation of a glaze with silica applies?
>
>Glaze BG1
>Soda feldspar 45.0
>China clay 6.0
>Bentonite 2.0
>Whiting 17.0
>Zinc oxide 5.0
>Quartz 25.0
>
>This is a shiny, well balanced glaze. Presumably all the silica has been
>taken into the glaze melt. It did not craze on stoneware but did show
>some crazing on porcelain.
>
>Our analysis of this glaze is as follows:-
>
>K2O .052 Al2O3 .337 SiO2 3.109
>Na2O .178 Fe2O3 .004 TiO2 trace
>CaO .574
>MgO .005
>ZnO .191
>
>Ratio Al2O3/SiO2 1/3.11
>
>5 x 10% increments of quartz were
>added to this so that the final glaze equalled.
>
>Soda feldspar 45.0
>China clay 6.0
>Bentonite 2.0
>Whiting 17.0
>Zinc oxide 5.0
>Quartz 75.0
>
>K2O .052 Al2O3 .337 SiO2 5.695
>Na2o .178 Fe2O3 .004 TiO2 trace
>CaO .574
>MgO .005
>ZnO .192
>
>Ratio Al2O3/SiO2 1/16.9
>
>The Linear Coefficient of Expansion ( English & Turner x10-6/oC)for
>these 6 glazes is in sequence
>
>BG1 = 5.01, 4.56, 4.20, 3.90, 3.64, 3.42.
>
>We were very surprised to see that the degree of crazing increased as
>the expansion coefficient went down. The last has the lines so close
>together that it looks like a high alkaline raku glaze!!!
>
>Our theory is that the excess silica, that cannot get into the melt,
>converts to cristobalite. This will in fact do its beta - alpha
>conversion of around 2% shrinkage at 220 deg C. The result is a
>shrinkage of the glaze and crazing results from the fact that there is
>now more cristobalite in the glaze than in the body. This is something
>that we have seen in other tests with high silica too but I've never
>done such a direct test to show it up. I'd assumed there was some
>explanation such as: Well perhaps that particular glaze had a lot of
>alkali in it.
>
>The glaze was next re-fired to cone 9. The amount of crazing may be
>lessened slightly by this but nothing very obvious. The overall picture
>of increased crazing with increased silica still holds.
>
>--
>David Hewitt
>David Hewitt Pottery ,
>7 Fairfield Road, Caerleon, Newport,
>South Wales, NP18 3DQ, UK. Tel:- +44 (0) 1633 420647
>FAX:- +44 (0) 870 1617274
>Own Web site http://www.dhpot.demon.co.uk
>IMC Web site http://digitalfire.com/education/people/hewitt.htm

Ron Roy

93 Pegasus Trail,
Scarborough, Ontario
Canada. M1G 3N8

Tel: 416-439-2621
Fax: 416-438-7849