CoE as the Determinant of Temperature Characteristics

Many people are under the impression that CoE can tell you a wide number of things about fusing glass. 

What does CoE really mean?

The first thing to note is the meaning of CoE.  Its proper name is the coefficient of linear expansion.  It tells you nothing certain about the expansion in volume, which can be as or more important than the horizontal expansion. 

It is an average determined between 20°C and 300°C.  This is fine for materials that have a crystalline structure. Glass does not.  Glass behaves quite differently at higher temperatures. 

It may have an average expansion of 96 from 20°C-300°C – although there is no information on the variation within that range – but may have an expansion of 500 just above the annealing point. 

The critical temperatures for glass are between the annealing and strain points.  One curious aspect to the expansion of glass is that the rate of expansion decreases around the annealing point.  The amount of this change is variable from one glass composition to another.

The CoE of a manufacturer’s glass is an average of the range which is produced.  Spectrum has stated that their CoE of their fusing compatible glass is a 10 point range.  Bullseye has indicated that their CoE range is up to 5 points. These kind of ranges can be expected in every manufacturer’s compatible glass.

CoE does not tell us anything about viscosity, which has a bigger influence on compatibility than CoE alone. 

Comparison of CoE and Temperature

Among the things people assume CoE determines is the critical temperatures of the strain, annealing and softening points of various glasses.

Unfortunately, CoE does not necessarily tell you fusing or annealing temperatures. 

“CoE 83”

Most float glass is assumed to be around CoE 83.  The characteristics depend on which company is making the glass and where it is being made.

Pilkington float made in the UK has an annealing point of 540°C and a softening point (normally the slump point) of 720°C.

Typical USA float anneals at 548°C and has a softening point of 615°C.

Typical Australian float has a CoE of 84 and anneals in the range 505°C -525°C.

“CoE 90”

Uroboros FX90 has an annealing point of 525°C compared to Bullseye at 482°C, and Wissmach 90 anneal of 510°C. 

Wissmach 90 has a full fuse temperature of 777°C compared to Bullseye's 804 - 816°C.   

There is a float glass with a CoE of 90 that anneals at 540°C and fuses at 835°C.

Bullseye has a slump temperature of 630°C-677°C and Wissmach’s 90 slumps between 649°C and 677°C, slightly higher.

“CoE 93”

Kokomo with an average CoE of 93 has an annealing range of 507°C to 477°C. Kokomo slumps around 565°C

“CoE 94”

Artista with a CoE of 94 has an annealing point of 535°C and a full

fuse of 835°C, almost the same as float with a Coe of 83. 

“CoE96”

Wissmach 96 anneals at 482°C with a full fuse of 777°C and a slump temperature of 688°C.

Spectrum96 and its successor Oceanside Compatible anneals at 510°C and full fuses at 796°C.

Conclusion

In short, CoE does not tell you the temperature characteristics of the glass. These are determined by several factors of which viscosity is the most important. More information can be gained from this post or from your own testing and observation as noted in this post.

CoE and Temperatures

CoE as a Determinant of Temperature Characteristics

What CoE Really Tells Us

The wide spread and erroneous use of CoE to indicate compatibility (it does not) seems to have led to the belief that CoE tells us about other things relating to the characteristics of fusing glasses.  It is important to know what CoE means.  

First it is an average of linear expansion for each °C change between 0°C and 300°C.  This is fine for metals with regular behaviour, but not for glasseous materials where we are more interested in the 400°C to 600°C range.  Measurements there have shown very different results than at the lower temperatures at which CoLE (coefficient of linear expansion) are measured.  In kiln forming we are also interested in volume changes and CoE tells us nothing about that.

Unfortunately, CoE does not tell you fusing or annealing temperatures. 

And not even relative temperatures.  

Some examples: 

• Uroboros FX90 has an annealing point of 525C compared to Bullseye (516/482C), and to the Wissmach 90 anneal of 510C. 
• Wissmach 90 has a fuse temperature of 777C compared to Bullseye's 804C.  
• Another example is Kokomo with an average CoE of 93 which has an annealing range of 507-477C and slumps around 565C. 
• There is a float glass of a CoE of 90 that anneals at 540C and fuses at 835C.  
• Artista (which is no longer made, except in clear) had a Coe of 94 with an annealing point of 535C and fuse of 835C, almost the same as float with a Coe of 83. 

These examples show that CoE can not tell you the temperature characteristics of the glass. These are determined by a number of factors of which viscosity is the most important. More information can be gained from this post on the characteristics of some glasses, or from testing and observation as noted in this post .

CoE does not tell you much about compatibility either, since viscosity is more important in determining compatibility.  CoE needs to be adjusted and varied in the glass making process to balance the viscosity of the glass.  Viscosity is described here .

This post and its links describes why Coe is not a synonym for compatibility. 


What CoE REALLY tells us is that we look for simple answers, even when the conditions are complex.  

Fixing a Broken Piece

This conversation is reproduced by permission (with some editing out of extraneous information). It is presented as an example of how conducting a critique of your schedule can have dramatic effects on the results of your firing. 

This is the piece as it came out of the kiln.

Picture credit: Ike Garson

You may have seen the photo I posted of a large copper blue streaky piece that has cracked right across. …  I’m wondering if it would be better trying to bring the 2 pieces together instead of opening up the 2 pieces and inserting frit. I was thinking of firing it with a tack or contour schedule.

This is the crack that developed later through the frit and single layer centre.

Picture credit: Ike Garson

I have 4 questions:

A.   Even if I manage to fix it, do you think that fissure line will always be too weak and liable to break off at any point?

The strength of the joint will be dependent on the firing conditions.  To make it strong, the temperature should go to full fuse.  Tack fusing will leave the joint more visible and weaker.  To stop the joint rounding during heat up, you will need to dam the piece tightly to stop the normal expansion of the glass and ensure the glass is forced together during the higher temperatures.

B.     I have some large pieces of clear confetti. Would it benefit using them to bridge the 2 sections from below?

Anything you put on the bottom will have distinct outlines and visibility.  The temperature on the bottom can be 10C or more different from the top surface, which is why you can get crisp lines with the flip and fire technique.

C.    Would clear powder hide the crack or would it always be visible after firing?

Any additions to the top may be less visible, but adding clear powder makes the join more obvious.  You need to use powder of the same colour as the sheet glass.  Since you are using a streaky glass, you can’t use coloured power either as it is very difficult to imitate the steaks even with powders of the same colours. 

More information was given indicating the first contour fuse schedule in Celsius:

1. 260 730 00.20
2. FULL 515 00.60
3. 260 150 End

This is the contour schedule I have used many times successfully but never for a piece during this week.

My critique of the schedule. 

Segment 1.

• ·    It is too fast for the small distance to the side of the kiln. 
• ·    It is too fast for a piece of varying thicknesses. Most expansion breaks occur above 300˚C, so a soak at ca.260˚C will help ensure the glass maintains an even temperature, especially with large differences in thickness. Then you can advance more quickly. 
• ·   There is no bubble squeeze.
• ·   The top temperature seems low for a good tack, or the soak is a bit short.  Long soaks allow the glass molecules to bind at the atomic level firmly. This is the principle used in pate de verre.
• ·   It definitely needs to be on fibre paper covered with thinfire to allow air out.

Segment 2.

• ·   The soak at 515˚C is better done at 482˚C for Bullseye.
• ·   My tests have shown that contour firing a piece like this at rates and holds for 1.5 times the height of the piece is necessary for good results.

Segment 3.

• ·   Also, my tests have shown that a three-stage cooling provides the best result.  Slow cooling keeps the glass within the 5°C difference required for avoiding stress.
• ·   Annealing at the bottom end of the range combined with an appropriate length of soak and slow cooling gives a denser glass than soaking at the middle of the annealing range. 
• ·   The best cooling comes from a three-stage cooling process.  This involves a slow rate for the first 55C, a rate of 1.8 times this for the second 55C, and a rate of 3 times this for the final cool to room temperature.

These points mean that I would recommend you fire for at least 10mm thick.  This recommendation is for a new piece, not a repair. In this repair case and for the conditions, I would choose 12mm as being more cautious. My schedule would look something like:

1. 120˚C to 260˚C, 20’
2. 300˚C to top temperature, 10’
3. Full to 482˚C, 120’
4. 20˚C to 427˚C,0’
5. 36˚C to 370˚C, 0’
6. 120˚C to room temperature, off

The anneal soak is for a piece 12mm thick.  The cool rates are for 21mm thick.  This is to compensate for the nearness of the glass to the edge of the kiln.  It will help to ensure the glass does not have excess stress locked into the piece during the cooling.

D. Do you think this schedule would work [for a repair]? It's adapted from a standard tack schedule.. 

1. 222 677 00.30
2. 222 515 00.40
3. FULL 482 01.30
4. 63 371 ENDS

Critique of the schedule.

Segment 1. 

• ·   Too fast given earlier difficulties. 
• ·   Too low for good adhesion unless you use about 10 hours soak. 
• ·   Even at sintering temperature (690°C) you would need 2 hours.  But at sintering temperature you do not alter the surface 

Segment 2. 

• ·   Too slow a cool from top temperature and risks devitrification. Should be FULL.
• ·   You do not need the soak at 515˚C.  It only delays the annealing process.  It seems this idea of soaking at the upper portion of the annealing range was introduced by Spectrum over 2 decades ago. 
• ·   Any advantage that might be achieved by the higher soak is cancelled by the FULL rate to the annealing soak. 
• ·   Go straight to the anneal soak. 

Segment 3. 

• ·   You need a more controlled 3 stage cooling to get the best result.

My schedule for repair would look something like this:

1. 120˚C to 540˚C, 10’
2. 300˚C to 780˚C, 10’
3. Full to 482˚C, 210’
4. 20˚C to 427˚C,0’
5. 36˚C to 370˚C, 0’
6. 120˚C to room temperature, off

I am making the assumption that 780˚C is full fuse in your kiln.  Anything less than full fuse will certainly show the crack.

 

A Look at Causes.

• ·  The piece is far enough away from the elements.  It is not on the floor. These are not the causes.
• ·  It is very near the sides of the kiln.  These are always cooler than the centre. There is always a risk of breaking in this case.  Very slow rates are needed. 
• ·  There is a 3.5 times difference in thickness within the piece. This also requires slow rates.
• ·  If the break were to have been on the heat up these elements of uneven heating, and rapid rates are a problem.  But the break occurred after the cool down. So, the annealing soak and cool is a problem. 
• ·  I have suggested some alterations to the schedules to address these things.

 

Fixing for Yourself

• ·   Dam it tightly to avoid expansion within the glass as it heats.  This holds the join together and causes the glass to gain a little height. 
• ·   Place the piece on 1mm or thicker fibre paper topped with thinfire.  This will help avoid a bubble forming in the clear.
• ·   I have suggested a schedule which is slower to ensure no further breaks.  It is slow to the strain point and fast after that. 
• ·   It needs to be a full fuse to fully join the two pieces and ensure it is sound.
• ·   The cool to annealing should be FULL.  Eliminate the soak in the upper annealing range. The effects of the time spent there is nullified by the rapid rate to the main annealing soak. 
• ·   Anneal as for 12mm, but with slower cool rates (for 21mm) to ensure there are no stresses built into the piece by the nearness of the glass to the edge of the kiln.
• ·   These methods and schedules will make it a strong whole.  But the join will still show on the bottom. 
• ·   After fixing, if you are still not satisfied, break it up for incorporation in other projects.

Finally, and unfortunately, I do not think it can be satisfactorily repaired for a client.  The crack will show on the back. You will know it is a repair, rather than a whole. And that will reflect on your feeling about the piece, and possibly your reputation.

 

Conclusion

The commission was successfully re-made from scratch by the artist using some of my suggestions on scheduling. This is the resulting piece.

 

Picture credit: Ike Garson

 

Careful analysis of the conditions around a break are important to making a successful piece in the future. Many factors were considered, but the focus became the schedule. Analysis of each step of the schedule led to changes that resulted in a successful piece with the original vision and new materials.

Writing Slumping Schedules

 

Slumping Schedules

When slumping fired pieces, it is most often appropriate to use a slow ramp rate to avoid too rapid expansion of the glass that might lead to a break. Most glass breaks on the ramp up are above 300°C/573°F. It is in this range that there is a rapid expansion of ceramic. This means a slow rate is protective for both glass and ceramic moulds.

This slow first ramp rate is followed by the rate determined as appropriate for profile and thickness. The table below gives rates and times for different profiles that are 6mm/0.25” thick. Of course, the slumping temperature will be altered for the glass according to the manufacturer’s stated range. The nature of the mould will also have a big effect on temperature and time. The soak times at the slump soak are those appropriate for the mould. The annealing soaks are related to the profile of the glass.

Slumping Schedules by Profile (Celsius) 6mm thick

Flat Fuse and Contour Tack

Actual thickness	Ramp 1 rate to 260°C	Soak time (min)	Ramp 2 rate	Slumping  temp. for mould *	Soak time (min)	Anneal as for contour:
6	240	20	240		30	9mm

Rounded Tack

Actual thickness	Ramp 1 rate to 260°C	Soak time (min)	Ramp 2 rate	Slumping  temp. for mould *	Soak time (min)	Anneal as for round tack:
6	150	20	150		30	9mm

Sharp Tack

Actual thickness	Ramp 1 rate to 260°C	Soak time (min)	Ramp 2 rate	Slumping  temp. for mould *	Soak time (min)	Anneal as for sharp tack:
6	120	20	120		30	9mm

 

Slumping Schedules by Profile (Fahrenheit) .025" thick

Flat Fuse and Contour Tack

Actual thickness	Ramp 1 rate to 500°F	Soak time (min)	Ramp 2 rate	Slumping  temp. for mould *	Soak time (min)	Anneal as for:
0.250”	432	20	432		30	0.375”

Rounded Tack

Actual thickness	Ramp 1 rate to 500°F	Soak time (min)	Ramp 2 rate	Slumping  temp. for mould *	Soak time (min)	Anneal for:
0.250”	270	20	270		30	0.375”

Sharp Tack

Actual thickness	Ramp 1 rate to 500°F	Soak time (min)	Ramp 2 rate	Slumping  temp. for mould *	Soak time (min)	Anneal for:
0.250”	216	20	216		30	0.375”

 

Rates

It is most often best to use a slow ramp rate to at least 500°C/933°F. This avoids the risk of inducing a too rapid differential expansion within the glass as it heats up. Experiments about the first ramp rate have shown firing as for two layers thicker than indicated by the profile schedule provides the best results. 

The rates for the anneal soak and cool are those that are one layer thicker than determined by the schedule for the profile. This has been shown by experimentation to give the best annealing result – i.e., least stress.

Temperatures

The slumping temperature needs to be altered for two factors:

• ·        the glass according to the manufacturer’s stated range, and
• ·        the nature of the mould.

Many manufacturers are giving recommended temperatures and times for slumping in their moulds. An example is the Bullseye “Quick Tip” which gives suggested temperatures and times for various sizes and natures of moulds that can form the basis of your independent scheduling of slumps. The rates are normally for flat uniformly thick pieces. This will need alteration for tack profile pieces.

Take note of the soak time in these recommendations. If it is less than 10 minutes, it is possible to reduce the temperature by about 10°C/18°F by using a 30-minute soak. This will reduce marking on the back of the glass.

Soaks / Holds

Slumping schedules tend to be more imprecise than many other operations in kilnforming because of variations in moulds and what is placed on them. This, consequently, makes observation of the slump more important. It is needed from a point below the target temperature – say 22˚C/40°F – to ensure the slump is stopped when it is complete, or extended if not. The controller manual will give the information on how to do both of these operations. In general, schedule slower ramp rates for thicker pieces in combination with the half hour soak. This means for each thickness greater than 6mm, the top temperature can be reduced slightly and still achieve a full slump.

The schedules here are applicable for pieces up to 9mm actual thickness.

Slumping of thicker pieces needs to apply the underlying scheduling method:

• ·        Apply the rate for two layers thicker for the advance to 260°C/500°F.
• ·        Continue the next ramp rate as for two layers thicker than calculated up to the slumping temperature.
• ·        For annealing, also select the rates and times for one layer thicker than indicated by the profile.

For example:

• ·        Rounded Tack of Bullseye, 12mm/0.5” thickness
• ·        Schedule for 25mm/1” (2 times multiplier)

Celsius schedule for up to 9mm actual thickness:

Segment >	1	2	3	4	5	6	7
Rate	150	150	ASAP	15	27	90	off
Temp	260	Top	482	427	370	RT
Time(mins)	20	30	240	0	0	0

and in Fahrenheit:

Segment >	1	2	3	4	5	6	7
Rate	270	270	ASAP	27	49	162	off
Temp	500	Top	900	800	700	RT
Time(mins)	20	30	240	0	0	0

 

A further example:

• ·        Sharp Tack of Bullseye, 0.5” thickness
• ·        Schedule for 31mm/1.25” (2.5 times multiplier)

Celsius schedule for up to 9mm actual thickness:

Segment >	1	2	3	4	5	6	7
Rate	78	78	ASAP	11	20	65	off
Temp	260	Top	482	427	370	RT
Time(mins)	20	30	300	0	0	0

and in Fahrenheit:

Segment >	1	2	3	4	5	6	7
Rate	140	140	ASAP	20	36	117	off
Temp	500	Top	900	800	700	RT
Time(mins)	20	30	300	0	0	0

 

These examples show that considerable differences in scheduling are needed for different tack profiles. It also shows longer annealing soaks and slower cooling rates are required for sharp than rounded tack pieces.

More information is given in the e-book Low TemperatureKilnforming. 

* Of course, the slumping temperature will be altered for the glass according to the manufacturer’s stated range. The nature of the mould will also have a big effect on temperature and time.