Coe and Annealing

If you have changed CoE (i.e., the manufacturer), then the annealing temperature is different. If you don't correct that, it's never going to work quite right.

 

I have several problems with this statement.

CoE does not determine the manufacturer. There are several manufacturers who claim to manufacture fusing glass to the same CoE.

No manufacturer makes to one CoE. All manufacturers have to vary the CoE of a particular glass to balance its viscosity. The CoE is a dependent variable. It depends on what the viscosity of the colour is. Spectrum at one point stated their System96 glass had a 10-point variation in CoE number. Oceanside will be no different. Bullseye have stated a 5-point difference. Other manufacturers have not stated their variations.

No manufacturer can guarantee compatibility with another’s. This is because the ingredients to make a fusing range of glass varies from one manufacturer to another. These variations can make the glass incompatible. To determine if you can combine two glasses from different manufacturers you need to do the compatibility testing yourself. The CoE number does not determine the temperature characteristics of the glass either. 

Annealing 

Having got my disagreements with the statement out of the way, I can go on to looking at differing annealing temperatures. There is a difference between annealing point and annealing temperature.

Annealing Range

Annealing occurs over a relatively small range between the softening point at the higher end to the strain point at the lower end of the range. The softening point is the temperature, above which the glass is so plastic that it cannot be annealed. The strain point is the temperature at which the glass becomes so solid than no annealing can occur below it.

Annealing Point

The annealing point is mathematically determined as the point at which the glass most quickly relieves the stresses within it. That temperature is determined by the viscosity of the glass. It is known as the glass transition point, and is expressed as Tg. In practice there are advantages in annealing at or below the published annealing point.

A soak above the annealing point is of no effect. Any equalisation of temperature that occurs on that soak is negated by the drop to the annealing point. It is better to spend the cumulative soak/hold time at the (lower) annealing temperature.

Annealing Temperature

The average annealing point for Bullseye is 516°C/962°F. Different formulations of their fusing compatible glass have different Tg temperatures. Research showed the best results for their thick glass is 482°C/900°F. Other research in academic institutions has shown that annealing at the lower part of the range provides a denser and stronger finished glass piece. This applies to thick as well as thin glass.

Bullseye has chosen to use a temperature 34°C/61°F below the average annealing point, based on their research. This is still about 7°C/13°F above the strain point. This approach can be applied to any fusing glass.

The strain point is approximately 43°C/78°F below the mathematically determined annealing point. If you know the annealing point you can choose to anneal – i.e., equalise the temperature of your glass – up to 30°C/54°F below that. 

This has a practical demonstration. Wissmach for some years designated 510°C/950°F as the annealing point for W96. A few years ago, they changed their recommended annealing temperature to be 482°C/900°F. The annealing results are good at both temperatures. The difference is that the annealing soak is for a in longer time at the lower than at the higher temperature. But it still provides a shorter annealing cool.

Firing with different anneal points

This apparent diversion - into annealing ranges - shows that it is possible to anneal glass with slightly different glass transition points at the same temperature. You may compromise a little for one glass or the other. You will also use longer times at the annealing temperature.

The annealing soak of Oceanside and Wissmach96 could both be at 482°C/900°F. Or, if it felt safer, it can be an average of the two. The average of the difference would make the annealing soak at 496°C/926°F. You would use a longer soak at this temperature than at the higher one. The safest would be to hold for an hour instead of 30 minutes for 6mm/0.25” of glass.

However, if the annealing point differs greatly, it is much more difficult. For example, float glass with an annealing point of 540°C/1005°F would be difficult to fit in the same firing with most fusing glass because of the wide range of official annealing points.

 

It is possible to anneal different glass at the same time if the annealing points are not widely different. Compromises need to be made.

 

Longer Anneal on Each Firing

 Need for an Extension of Anneal Soak on Subsequent Firings

 Sometimes people recommend extending the length of the anneal soak each time the piece is fired. If nothing significant is added, there is no reason to extend the anneal soak.  If the piece can be fired as fast as the previous firing, the annealing will be the same, not longer.

 The physics and chemistry of annealing glass are the same for re-fired glass - without additions - as they were for the first. Extending the annealing soak seems to be more about reassurance of the kilnformer than a necessity.

 Bullseye research has shown that it IS possible to over anneal, locking in stress. If concerned about inadequate annealing, it is best to reduce the cooling rate. Especially over the first 55°C below the anneal soak temperature.  The testing and recording that I have done for a book on tack fusing shows that any differences in the glass - at the end of the anneal soak - will be relieved in that first 55°C/100°F. The remainder of the cool to 370°C/700°F can be about 1.8 times faster, and the final cool ramp can be 3 times faster than the 2nd stage cool. 

 I have observed that a three stage anneal cool is important to successfully anneal a piece. This has been reinforced by the temperature recordings of many firings. Often at the end of the anneal soak there is a little more than the desired 5°C/10°F difference in temperature across the piece. The recordings show this is relieved during the slow firs stage cool and maintained over the next two cool stages. If the kiln is cooling more slowly than the schedule, no electricity is used.  No kiln time is lost.

An example of the first cooling stage

 The first stage cool is key to a successful stress-free result.  If there are concerns about inadequate annealing, two things are important.  Be sure the right length of soak is chosen for the piece being fired.  Second, reduce the speed of cooling by the rates for a piece at least 3mm thicker.  These rates are available from the Bullseye chart for annealing thick slabs.

 The rates are applicable to other than Bullseye glass.  Only the temperatures need to be changed.

 If no significant changes (other than powder, wafer or stringers) are made to the glass before the second firing, no lengthening of the annealing is necessary.

 More detailed information is available in the e-book: Low Temperature Kilnforming.

Re-firing

A frequently asked question is “how many times can I re-fire my piece?”

This is difficult to answer as it relates to the kind of glass and the firing conditions.

Kind of glass

Float glass is prone to devitrification. This often begins to appear on the second firing. Some times it may be possible to get a second firing without it showing. Sandblasting the surface after getting devitrification will enable another firing at least.

Art glass is so variable that each piece needs to be tested.

Fusing glasses are formulated for at least two firings, and experience shows may be fired many of times. The number will depend on the colours and whether they are opalescent. Transparent colours on the cool side of the spectrum seem to accept more firings than the hot colours. Both of these accept more firings than opalescent glasses do.

Firing conditions

Temperature

The higher the temperature pieces are fired at, the fewer re-firings are possible. So if multiple firings are planned, you should do each firing at the lowest possible temperature to get your result. This may mean that you have relatively long soaks for each firing. The final firing can be the one where the temperature is taken to the highest point.

Annealing

You do have to be careful about the annealing of pieces which have been fired multiple times. A number of people recommend longer annealing soaks. However, I find that the standard anneal soak for the thickness is enough. What is required is cooling rates directly related to the anneal soak.  This is a three-stage cooling as described in the Bullseye chart Annealing Thick Slabs.  The slump firing can be annealed at  the standard. 

Slumping

In general, slumping is at a low enough temperature to avoid any creation of additional stress through glass changes at its plastic temperatures.  But any time you heat the glass to a temperature above the annealing point, you must anneal again at least as slowly as in the previous firing. Any thing faster puts the piece at risk of inadequate annealing.  Of course, having put all this work and kiln time into the piece, the safest is to use the cooling rate as for a piece one layer thicker.  My research has shown that this gives the least evidence of stress.

Testing

Testing for stress after each firing will be necessary to determine if there is an increase in the stress within the piece. In the early stages of multiple firings, you can slow the annealing and if that shows reduced stress, it will determine your previous annealing schedule was inadequate. When reducing the rate of annealing does not reduce the stress, it is time to stop firing this piece at fusing temperatures.

Revised 6 May 2023

The Importance of Three-Stage Cooling

It is common to think of cooling after annealing as a simple single cool rate to an intermediate temperature between annealing and room temperatures before turning off. This most often works well for full fused pieces up to 6mm/0.25. But as the pieces become thicker or more complex, the need for more controlled cooling becomes necessary.

 The aim of annealing is to get the glass to be the same temperature throughout its substance during the annealing soak. This is called the ΔT (delta T).  This difference has been shown to be 5°C to avoid high levels of stress.  Therefore, ΔT=5°C/10°F.  This difference in temperature needs to be achieved during the annealing soak and maintained during the cool.

 The object of controlled cooling is to maintain this small difference in temperature. It needs to be maintained throughout the cool to avoid inducing excessive stress in the glass, even if the stress is only temporary.  

 As the thickness or complexity of the piece grows, the annealing soak needs to be longer and the cool slower. The first cool is critical to the production of stress-free fused glass. That is the fastest rate that can be used in a single or multiple stage cooling. If you use that rate all the way to 370°C/700°F you will need at least 1.3 times longer to get to that temperature than if you used the first two parts of a 3-stage cool. This time saving becomes greater as complexity and thickness demand slower cool rates. It is not only time that is saved.

 The risk of breaks from rapid cooling after the anneal soak and to 370°C/700°F increases with more complex and thicker pieces. Although the stress induced by rapid cooing below the strain point is temporary, it can be great enough momentarily to break the glass. This is so even if the glass meets the ΔT=5°C/10° during the annealing soak.

  

Examples may help understand the cooling requirements of glass that it thicker, or tack or contour fused.

Example 1

A 12mm/0.5” full fused piece needs a two-hour annealing soak, followed by three cooling rates of 55°C/100°F per hour, 99°C/180°F hour and finally 300°C/540°F per hour. The first rate is for the first 55°C/100°F, the second rate for the next 55°C/100°F, and the final rate is to room temperature.

 What happens here is instructive as to the reasons for soaks and cool rates. In this recorded example the ΔT at the start of the anneal is 7°C/12.6°F. During the soak, the ΔT reduces to as little as 2°C, but ends with a ΔT=3°C. The 55°C/100°F cool rate over the first 55°C/100°F enables the ΔT to remain between 3°C and 4°C.  The second cool over the next 55°C/100°F maintains this ΔT of 3°C to 4°C. During the final cool the ΔT varies from 5°C to 1°C.

 

An example of the variation in ΔT during the first 55C/100F of cooling

Example 2

A rounded tack fuse of 1-base and 2-layer stacks gives a total of 9mm/0.375”. Research has shown that you need to schedule for twice the actual thickness for rounded tack fusing - so for 19mm/0.75”.

This requires an anneal soak of 150 minutes, and a first cool of 20°C/36°F. The second cool rate can be increased to 36°C/65°F. The final rate can be at 120°C/216°F per hour to room temperature.

 The ΔT at the beginning of annealing was 7°C/12.6°F and at the end of a 2-hour soak was a ΔT of 1°C/2°F. The first cool ramp was 20°C/36°F per hour and gave a variance of between 2°C/3.6° and 0°. The final cool produced variances of up to 6°C/11°F, ending at 88°C/190°F with a ΔT=2°C.

 The first two stages of cooling save 1.27 hours of cooling time over a single stage cooling of 20°C/36°F to 371°C/700°F. It still keeps the glass within that ΔT=5°C. More importantly, the third stage cooling is able to keep the variance to between 6°C and down to 2°C.

 The natural (unpowered) cooling rate of my 50cm/19.5” kiln at 370°C/700°F is 240°C/432°F per hour. It settles to the 120°C/216°F per hour only at 200°C/392°F. This is a fairly typical cooling rate for medium sized kilns. This rapid cooling at 370°C/700°F creates a greater risk of breakage than the controlled cool.

 

An example of the ΔT during the second 55C/100F of cooling

Example 3

A sharp tack or sintered piece with two base layers and two tack layer stacks on top requires firing as though 30mm/1.18”.

 This needs a 4-hour soak during which the ΔT varied from 8°C to 4°C. The first cooling rate was at 7°C/12.6°F and gave a ΔT variance of 4°C to 2°C. The second cooling rate of 12°C/22°F produced variances of 3°C to 1°C by 370°C/700°F. The final cool of 40°C/72°F per hour gave differences ranging from 5°C to 0° at 110°C/230°F.

 Note that the test kiln’s natural cooling rate does not achieve the third cooling rate until 140°C/284°F.  This shows that turning off the kiln at 370°C/700°F produces a high risk of breakage for thick and complicated pieces.  In addition, the two stage cooling rates saves 3.27 hours of cooling time.

An example of the ΔT during the final stage of cooling to Room Temperature

 The temperature differentials below the strain point can exceed the ΔT=5. The stresses induced are temporary according to scientists. But they can be great enough to break the glass during the cooling. It follows that the anneal soak may have been adequate, but the cool was so fast that excess stress was induced by the differential contraction rates. This stress being temporary, implies that testing for stress in a broken piece may not show any. The momentary excess stress will have been relieved upon cooling completely to room temperature.  (IMI-NFG Course on Processing in Glass, by Mathieu Hubert, PhD. 2015 , p.9.)

 

More information on cooling is given in the book LowTemperature Kilnforming; an Evidence-Based Approach to Scheduling.

Annealing Soaks are Related to Cool Rates

Good annealing is important to the success of each firing of a piece. 

This is generally agreed.

 

I do not understand the reasoning of those who use long anneal soaks followed by quick cool rates and early shut offs. I don't understand because reasons are not given. Or the reasons are in the realm of kiln fairies and other mythical beings.

The length of the annealing soak can be determined from established sources. The Bullseye table for annealing thick slabs gives the recommended soak times for evenly thick slabs of glass from 6mm/0.25” to 200mm/8.0”. Use that to determine the annealing soak time.

The soak times do not need lengthening except for pieces of uneven thicknesses. The ebook Low Temperature Kilnforming gives the calculations for variations in thickness and degree of tack. Generally, they are 1.5 for contour; 2 for rounded tack; and 2.5 for sharp/angular tack. Excessively long soaks are not desirable. This is additional evidence that long soaks and quick cools create problems.

The Relationship Between Soak and Cool Rate

Use of the Bullseye table shows that there are cool rates associated with the soak times. These rates for the length of annealing soak need to be used, as they are based on research, rather than fingers in the air or mythical beings.

My experiments have shown the need to control the cooling rates to at least 50C before shutting off. The end of an adequate annealing soak has the glass within 5°C/10°F of each other part (the ΔT=5). The slow cool for the first 55°C/100°F below is important to avoid exceeding that maximum differential. The rate for the next 55°C/100°F is faster and can allow a wider ΔT, as the stresses are temporary. But they can be great enough at any point to break the glass during fast cools. Therefore, the rates associated with the annealing soaks cannot be exceeded safely.

Do not just use "what works" for others. Use information based on research. The only company publishing research is Bullseye. Their research is applicable to all fusing glass with the appropriate temperature adjustments. 

If you use long annealing soaks and quick cool rates or ones that stop at about 370°C/700°F, you risk breakages of your glass. There is no reason to take that risk.  Also long cools from annealing to 370°C take longer than the staged cooling recommended by the Bullseye research.

 

More information is available in the ebook Low Temperature Kilnforming.

 

 

Annealing Tack Fused Pieces

"I'd like advice about annealing. I'm about to start a series that are to be wall hangings. The outside 100mm is only 3mm thick and the centre is 6mm and occasionally 9mm thick. They are going to be A3 and A2 size. I intend to tack fuse. I'm happy to experiment with the tack fuse temperature (I think about 677°C). How long should I anneal it? That's my question."

Determining the Tack Temperature

The high end of slumping, and the low end of sintering is 677°C/1252°F. Unless your kiln fires very hot, this is not hot enough for a tack fuse. Make some small-scale mock-ups in clear. Schedule the kiln to a full fuse on a rate not more than 300°C/540°F per hour. Peek into the kiln at 10°C/18°F intervals from 677°C/1252°F upwards. When you see the profile you want, note the temperature. When scheduling the tack firing, reduce the target temperature by 5C and add a 10-15 minute soak to get approximately the same result as you observed in the test firing.

Scheduling to Avoid Dog Boning

You have a border of 100mm/4.0” that is only one layer thick. This has the risk of becoming irregular at the corners compared to the sides (dog boning). To avoid dog boning of the 3mm base, the lowest temperature you can use is important. This is the main reason for the peeking – to determine the temperature at which dog boning begins. It is not only the degree of rounding of the edges you are looking for, but also the degree of retraction of the sides of the piece. When you note the beginning of the dog boning, you have reached just beyond the temperature to avoid that.

You will of course have to set the mock-up in such a way that you can see at least one side through the peep hole. The front will not give you accurate information, but if the side is in your sight line, you will see when it begins to deform. This peeking will keep you occupied for about 3/4 hour. Make sure you have gridded paper and pencil to hand to record information in between peeking.

It may be that the glass has not begun to round when the dog boning starts. In this case you will need to make the border larger and cut the glass back to straight lines. 20-30mm/0.75-1.125” extra all around will make it easy to trim the excess cleanly.

Annealing

I do not know the degree of tack you are aiming to achieve. It is important to the scheduling of the anneal. A sharp tack profile will require annealing for longer than a contour profile for your thicknesses. These suggestions assume the total height is 9mm/0.375”. If it higher, the soak and cooling times and rates will be longer and slower.

A sharp tack profile will need:

• Annealing for 270 minutes (4.5 hours) with a cool rate of:

First 55°C/100°F cool at 13°C/23°F per hour.

Second 55°C/100°F cool at 23°C/41°F per hour.

Final cool at 78°C/140°F per hour to room temperature.

A rounded tack profile will need:

• Annealing for 180 minutes (3 hours) with a cool rate of:
• First 55°C/100°F cool at 25°C /45°F per hour
• Second 55°C/100°F cool at 45°C /81°F per hour
• Final cool at 150°C /270°F per hour to room temperature.

A contour tack profile will need:

• Annealing for 120 minutes (2 hours) with a cool rate of:
• First 55°C/120°F cool at 55°C /99°F per hour
• Second 55°C/100°F cool at 99°C /178°F per hour
• Final cool at 330°C /216°F per hour to room temperature.

More detailed information is in my eBook Low Temperature Kilnforming, An Evidence-Based approach to Scheduling. It is available from VerrierStudio on Etsy or through Bullseye. 

It is not cheap, but at 300pp worth it (in my opinion!). It discusses the three profiles of tack fusing - sharp, rounded, contour. It also deals with slumping, sintering, freeze and fuse, and bas relief or texture mould firings. The method for determining schedules is outlined and specific sample schedules are listed.

Annealing Requirements for Shaped Pieces.

 Experiments related to slumping show that shaped items such as slumped, textured and kiln carved glass need annealing for at least one layer thicker than they are. The annealing for one layer thicker than the calculated thickness provides the most stress-free result for the finished product. Annealing for the calculated thickness does not always produce a stress-free result.  

Full Fuse

 This indicates that an evenly thick 6mm thick piece will get the best result from an anneal as for 9mm.

Texture Moulds

 A piece of glass on a texture mould with 3mm or more differences in height requires careful annealing. The more defined/sharper the texture, the greater care will be required. A 6mm blank on a mould with 3mm variation taken to a well-defined texture needs to be annealed as though it were 18mm thick. A sharp tack requires annealing as for a piece 2.5 times its actual thickness plus another 3mm.  This gives the 18mm/0.75” thickness annealing requirement for the 6mm thick piece.

Kiln Carvings

 The same kind of calculation applies to kiln carved items as for sharply textured pieces. Pieces with less sharply defined profiles can be treated as one of the more common tack fused profiles.

Credit: Vitreus-art.co.uk

Tack Fusings

 A rounded tack fused piece of a 6mm base with 3mm tack elements that is being slumped will need annealing as for 21mm.  Twice the actual thickness plus 3mm giving the annealing requirement as for 21mm/0.827”.

 A contour tack of the same dimensions as given in the first example will require annealing as for 19mm/0.75”. The annealing requirement when slumping is for 1.5 times the thickness plus another 3mm.

In General

 The general approach to annealing shaped pieces is to calculate the thickness for the anneal and add one layer more to get a good anneal for slumped and other formed pieces. 

 The research and the reasoning behind this approach is given in LowTemperature Kilnforming, An Evidenced-Based Guide to Scheduling available from the Etsy shop VerrierStudio and from Bullseye. 

Effects of Dam Materials on Scheduling

 I once made a statement about the effects of various dam materials on scheduling. This was based on my understanding of the density of three common refractory materials used in kilnforming – ceramic shelves, vermiculite board and fibre board. I decided to test these statements.  This showed I was wrong in my assumptions.

I set up a test of the heat gain and loss of the three materials. This was done without any glass involved to eliminate the influence of the glass on the behaviour of the dams. The dam materials were laid on the kiln shelf with thermocouples between. These were connected to a data logger to record the temperatures.

Test Setup

 The thicknesses of the dams may be relevant. The vermiculite and fibre boards were 25mm thick. The ceramic dam material was 13mm thick.

The schedule used was a slightly modified one for 6mm:

• 300°C/hr to 800°C for 10 minutes
• Full to 482°C for 60 minutes
• 83°C to 427, no soak
• 150°C to 370°C, no soak
• 400°C to 100°C, end

 

The data retrieved from the data recording is shown by the following graphs.

Temperature profile of the air, ceramic, fibre, and vermiculite during the firing.

Highlights:

• The dam materials all perform similarly.
• This graph shows the dams have significant differences from the air temperature – up to 190°C – during the first ramp of 300°C/hr. (in this case).
• There is the curious fall in the dams’ temperatures during the anneal soak. This was replicated in additional tests. I do not currently know the reasons for this.
• The dams remain cooler than the air temperature until midway during the second cool when (in this kiln) the natural cooling rate takes over.
• From the second cool to the finish, the dams remain hotter than the air temperature.

 Some more information is given by looking at the temperature differentials (ΔT) between the materials and the air. This graph is to assist in investigating how significantly different the materials are.

This graph is initially confusing as positive numbers indicate the temperature of the first is cooler than the material it is compared with, and hotter when in negative numbers.

 

A= air; C=ceramic; F=fibre board; V=vermiculite

Temperature variations between air and dams

 As an assistance to relating the ΔT to the air temperature some relevant data points are given. The data points relate to the numbers running along the bottom of the graph.

 Data Point       Event

• 1            Start of anneal soak.
• 30          Start of 1^st cool (482°C)
• 45          Start of 2^nd cool (427°C)
• 65          Start of final cool (370°C)
• 89          1^st 55°C of final cool (315°C)
• 306         100°C

 

At the data points:

• At the start of anneal soak the ΔT between the dams is 16°C with the ceramic shelf temperature being 18°C hotter than the air.
• At the end of the anneal soak of an hour, the air temperature is 20°C higher, although the ΔT between the dams has reduced to 12°C.
• At the end of the 1^st cool the ΔT between the dams has reduced to 9°C and the ΔT with the air is 3°C.
• At approximately 450°C the air temperature becomes less than the dams.
• At 370°C the hottest dams are approximately 17°C hotter than the air.  The ΔT between the dams is 10°C.

 More generally:

• The air temperature tends to be between 17°C hotter and 17°C cooler than the ceramic dams during the anneal soak and cool.  The difference gradually decreases to around 8°C at about 120°C.
• Ceramic and fibre dams loose heat after the annealing soak at similar rates – having a ΔT between 4°C and 1°C, with a peak difference of 9°C at the start of the second cool. This means the heat retention characteristics of ceramic strips and fibre board are very close.
• Between the annealing soak and about 300°C the vermiculite is between 12°C and 9°C hotter than the same thickness of fibre.  Vermiculite both gains and loses heat more slowly than the ceramic or fibre dams do. This means that vermiculite is the most heat retentive of the three materials.
• Vermiculite remains hotter than ceramic from the start of the second cool. This variance is up to 9°C and decreases to 3°C by 100°C.
• Fibre board is cooler than ceramic dams until the final cool starts, when there is little variance.  At the start of the second cool there is about 15°C between the two.
• Vermiculite remains cooler than fibre dams throughout the cooling process. This ranges from about 12°C at the start of the first cool to about 3°C at 100°C.

Since we cannot see more than the air temperature on our controllers it is useful to compare air and dam temperatures. The same data points apply as the graph comparing differences between materials.

 

Ceramic-Vermiculite; Ceramic-Fibre Board; Vermiculite-Fibre Board; Ceramic-Air Temperature
This graph shows the temperature differences throughout the cooling of various materials.

• During the annealing soak, the air temperature is greater than the dam temperatures. The fibre and vermiculite boards remain at similar temperatures and the ceramic dam is the coolest.
• The three dam materials even out with the air temperature at the start of the second cool.
• Through the second and final cools, vermiculite dams remain hotter than the air temperature – between about 24°C at start of the final cool and 9°C at 100°C.
• The ceramic and fibre dams are close in temperature difference to the air from the start of the final cool. Their ΔTs are 17°C at the start of the final cool and 6°C at 100°C.

Conclusions

• Dams will have little effect during the heat up of open face dammed glass.  The slight difference will be at the interface of the glass and the dams where there will be a slight cooling effect on the glass. Therefore, a slightly longer top soak or a slightly higher top temperature may be useful.
• The continued fall in the dams’ temperature during the anneal soak indicates that this soak should be extended to ensure heat is not being drained from the edges of the glass by the dams. There is the risk of creating unequal temperatures across the glass.
• The ability of ceramic and fibre dams to absorb and dissipate heat more quickly indicates that they are better materials for dams than vermiculite board. The slightly better retention of heat at the annealing soak, indicates that ceramic is a good choice when annealing is critical.        
• These tests were fired as for 6mm/0.25” glass and so show the greatest differences. Firing for thicker glass will use longer soaks and slower cool rates. These will allow the dams to perform more closely to the glass temperature during annealing and cool.

Based on these observations, I have come to some conclusions about the effect of dams on scheduling.

• There is no significant effect caused by dams during the heat up, so scheduling of the heat up can be as for the thickness of the glass.
• The lag in temperature rise of the dams indicates a slightly longer soak at the top temperature (with a minor risk of devitrification), or a higher temperature of, say 10°C, can be used.
• The (strange) continued cooling of the dams during the annealing soak indicates that extending the soak time to that for a piece 6mm thicker than actual is advisable.
• The cool rates can continue to be as for the actual thickness, as the dam temperatures follow the air temperature with little deviation below the end of the first cool.
• Ceramic dams of 13mm/ 0.5” perform better than 25mm/1.0” vermiculite and fibre board. 
• However, in further tests of 25mm/1.0” thick ceramic dams performed similarly to the same thickness of vermiculite. So, 25mm/1.0” fibre board the best when choosing between the three materials of the same thickness. But 25mm ceramic strips are not common, nor are they needed for strength or weight.
• The performance of the three dam materials tested do not show enough difference in temperature variation to have significant affects on the annealing and cooling at times and rates appropriate to the thickness of the glass.
• It is the thermal insulation properties of the dam material, rather than the density that has the greatest influence on performance as a dam material.

 

 

Scientific Notes on Annealing

 The course from which this information is taken is based on float glass.  This is a soda lime glass just as fusing glass is.  The general observations – although not the temperatures – can be applied to fusing glasses.  This is a paraphrase of the course. It relates these observations to kilnforming.  The course is IMI-NFG Course on Processing in Glass, by Mathieu Hubert, PhD. 2015 

 

Viscosity vs. Temperature for a borosilicate glass
Graph credit: Schott

Viscosity Influence on Annealing

 Viscosity increases with reduction in temperature.  So high viscosity (low temps) cannot release stress; low viscosity (high temperature) cannot maintain shape – it will deform.  The range of viscosity is small.  The viscosity must not be so high that the stress cannot be relieved, nor must it be so low that the glass is unable to retain its shape. (p.6).  This indicates there is an inverse relationship between temperature and viscosity.  This is something we experience each time we fire. 

 The mathematical definition for strain point - high viscosity - is 10^14.5 Poise.   And the annealing point as 10^13.4 Poise, where if the glass is all the same temperature, the stress can be relieved in about 15 minutes.  (p.7-8)  

 As kilnformers we talk of the annealing range in terms of temperature, because that is what we can measure. The annealing occurs within a small range of viscosity. This has a relation to temperature that is not the same for all glass compositions. 

 The definition of the annealing as the range of viscosity at which annealing can occur is important.  

 First, the viscosity value remains the same over many types and styles of glass.  The temperature required to achieve that viscosity varies, leading to different annealing temperatures for different glass. 

 Second, there is a range of viscosity - and therefore temperature - during which annealing can occur.  The annealing point is 10^13.4 Poise, at which viscosity the stresses in glass can most quickly be relieved (generally within 15 minutes for 3mm glass).  However, the stress can be relieved at greater viscosities up to almost the strain point - 10^14.5 Poise. (p.8).  At higher temperatures, the glass becomes more flexible and cannot relieve stress.  At lower temperatures (beyond a certain point) it becomes so stiff that stress cannot be relieved.  Again, those temperatures are determined by the viscosity of the glass.

 

Annealing Soaks

 Annealing can take place at different points within the range.  Bullseye chose some years ago to recommend annealing at a higher viscosity, i.e., a lower temperature.  This has also been applied by Wissmach in their documentation although initially the published annealing point was almost 30°C higher. 

 The closer to the strain point that annealing is conducted, the longer it will take to relieve the stress.  Annealing at the strain point is possible, but it is impractical.  Apparently, it would take at least 15 hours for a 6mm thick piece (p.8). 

 However, the trade off in annealing a few degrees above the strain point – requiring longer annealing soaks – is reducing the amount of time required by the annealing cool, especially for thicker or more difficult items.

 A further advantage to annealing at lower temperatures and slower rates is that it results in a denser glass – one with lower volume (p.3). Arguably, a denser glass is a stronger one.

 

Annealing Cool

 After annealing, the glass should be cooled slowly and uniformly to avoid formation of internal stresses due to temperature differentials within the glass.  Stresses that are unrelieved above the strain point are permanent.  Stresses induced during cooling below the strain point are temporary, unless they are too great.  To avoid permanent stress, the cooling should be slow between anneal soak and strain point (p.9).  Although glass can be cooled more quickly below the strain point, care must be taken that the temperature differentials within the glass are not so great as to cause breaks due to uneven contraction.

 Annealing cool factors for flat pieces are about three times that for cylinders and five times that for spheres (p.26). Or the other way around – spheres can be annealed in one fifth the time, and cylinders in one third of the time as flat glass of the same volume.   This indicates how much more difficult it is to anneal in kilnforming than in glass blowing.

 The industrial cooling rate for float glass of 4mm is 6 times the rate for 10mm although only 2.5 times the difference in thickness (p.27). This indicates that the thicker the glass, the slower the rate of cooling should be.  But also, that there is not a linear correlation between cooling rate and thickness.

 Glass with no stress has a uniform refractive index.  Stresses produce differences in the refractive index which are shown up by the use of polarised light filters.

Source: IMI-NFG Course on Processing in Glass, by Mathieu Hubert, PhD. 2015 (available online www.lehigh.edu/imi).

https://www.lehigh.edu/imi/teched/GlassProcess/Lectures/Lecture09_Hubert_Annealing%20and%20Tempering.pdf

Firing Small Pieces

 Do you run small pieces of glass through the whole cycle or just bring it up to your degree posted and cool down?

 

Picture credit: Eva Glass Design

It would appear easy to ignore the need to anneal small pieces.  They can anneal with short heat soaks.  In industry the anneal of sheet glass is 15 minutes for 4mm/0.019” glass.   In kilnforming the 30ºC - 40ºC/54ºF – 72ºF below the annealing point is where annealing is effective.  If you are certain that the natural cooling rate of your kiln is more than 15 minutes for that temperature range, you can simply turn off after top temperature.

However, it is not a good practice unless you intend to confine you kilnforming to small pieces.  All glass needs to be annealed to be sound.  Small pieces may need only 15 minutes and often that can be achieved with the natural cooling rate of your kiln.  But pieces of 6mm/0.25” thick and over 100mm/4” in any direction need to be annealed with longer soaks and slower cools.  This is done with a hold of the amount of time appropriate to your glass and layup.  There is an excellent table from Bullseye that gives the hold times and rates for cooling glass of different calculated thickness. 

Using an annealing soak and a cooling cycle for every firing is a good practice.  This gets you into a habit, so that you do not skimp on the anneal and cool for larger, thicker, or tack fused pieces.  If your kiln cools more slowly than you have scheduled, that's ok.  The kiln does not use any electricity to heat the elements.  No additional electricity cost or wear on the kiln occurs.

Notes on Polarised Light Filters

Polarised light filters are used to detect stress in a non-destructive testing method in kilnforming.  The use of the filters is described in this blog. To produce consistent reliable results, there are certain conditions.

 The light source needs to be diffused in such a way that it is even across the viewing area.  An intense, single point light makes it difficult to determine the relative intensity of apparent stress. Another tip is that you can use your phone or tablet as a source of diffused light and as the bottom filter.  It emits polarised light, meaning only a top filter is needed.

Stress halos from broken and fused bottles

 It is important that the glass being tested is of the same temperature throughout to get a meaningful result.  This was emphasised to me when I was running a series of tests. I got in a hurry to test for stress to be able to start the next trial quickly.  I began to notice inconsistencies in the amount of stress I recorded for results of the series of tests.  Going back to the stressed test pieces, showed different stress levels when they were cold from when they were warm.

 The conclusion is that the glass to be tested for stress must be the same temperature throughout.  Even if it is only slightly warm, the apparent stress will be exaggerated.  It may be that the testing can only be done 24 hours after removed from the kiln.

 Stress will be more evident at points and corners.  The light will be brighter at highly stressed points, and even at extreme stress exhibit a rainbow effect.  More generalised stress is evident in a lighter halo.

Stress points in a drawing square illustrating the concentrated stress at corners

 It is much more difficult to check for stress in opaque areas of a piece.  If there are transparent areas, the stress will show there, although the stress may originate in the opaque ones. To be aware of potential stress in the combination of opaque glass, strip tests must be conducted on samples of the glasses. 

 Remember to include an annealing test too, as the stress test does not distinguish the type of stress.  If the annealing test shows stress, the annealing was inadequate. It is of course, possible that the glass is stressed because of incompatibility.  But the only way to determine that is to fire another test with a longer soak at annealing.