Observations on Some Suggestions about Annealing

There are writings from a teacher attempting to make glass fusing simple.  Unfortunately, glass physics and chemistry are very complicated.  Attempting to avoid these complications leads to failures and other difficulties as the practitioner progresses. 

Proper annealing is one of the fundamentals to achieving sound kilnforming results.  Some suggestions have been made by a widely followed person to “simplify” the understanding of the annealing process.  Discussion of the meaning and importance of annealing can be found in many places, including here.  

Annealing temperatures

It has been suggested that the annealing temperatures can be inferred from the CoE of the glass that is being used. Discussion of what CoE is and is not can be found here and here.

Annealing temperatures are not directly related to the expansion coefficient (CoE) of the glass.  This can be shown from the published annealing temperatures for different glasses organised by presumed CoE:

·        “CoE96”: Wisssmach 96 - anneal at 482°C;  Oceanside - anneal at 515°C

·        “COE94”: Artista - anneal at 535°C

·        “CoE 93”: Kokomo - anneal between 507°C and 477°C – average 492°C

·        “CoE 90”: Bullseye - anneal at 482°C; Wissmach90 - anneal at 482°C; Uroboros FX90 - anneal at 525°C

·        “CoE 83”:

o   Pilkington (UK) float - anneal at 540°C;

o   typical USA float - anneal at 548°C;

o   Typical Australian float - anneal between 505°C and 525°C, average 515°C

This shows there is no direct relationship between CoE and annealing temperature.  Do not be tempted to use a CoE number to indicate an annealing temperature.  Go to the manufacturer’s web site to get the correct information.

Temperature equalisation soak

Annealing for any glass can occur over a range of temperatures.  The annealing point is the temperature at which the glass can most quickly be annealed.  However, the glass cannot be annealed if it is not all at the same temperature throughout the substance of the glass.  It has been shown through research done at the Bullseye Glass Company that a temperature difference of more than 5°C will leave stress within the glass piece. To ensure good annealing, adequate time must be given to the temperature equalisation process (annealing). 

From the Bullseye research the following times are required for an adequate anneal soak:

6mm /   1/4"            60 minutes

[9mm /  3/8"           90 minutes]

12mm  / 1/2"          120 minutes

[15mm  /   5/8"       150 minutes]

19mm   / 3/4"         180 minutes

[ ] = interpolated from the Bullseye chart for annealing thick slabs

Anneal Cooling

There are suggestions that a “second anneal” can be used on important pieces.  Other than observing that all pieces are important to the maker, the suggestion should be investigated.  On looking into the idea, it is essentially a second soak at 425°C, which is slightly below the strain point, rather than controlled cool from the anneal soak temperature.

It is reported that the Corning Museum of Glass considers 450°C as the lower strain point – the temperature below which no further relief of strain is possible.  This means that any secondary soak must occur above 450°C rather than the suggested 425°C. Such a soak is unnecessary if the appropriate cooling rates are used. 

Cooling Rate

Except in special circumstances, the cooling rate needs to be controlled as part of the annealing process.  Soaking the glass at the anneal is not the completion of the annealing.  Most practitioners follow the practice of choosing a slow rate of cooling from the annealing soak to some point below the strain point rather than a rapid one with a soak at the strain point temperature.

Annealing is not just the soak time (which is there to equalise the temperature), it is about the rate of the annealing cool too. The rate at which you cool is dependent on the thickness of the glass piece and whether it is all of one thickness or of variable thicknesses.

Even thickness

                                         Cooling rate

Dimension      time (mins)     1^st 55°C   2^nd 55°C

6mm              60                 83°C       150°C

9mm              90                 69°C       125°C

12mm            120                55°C       99°C

15mm            150                37°C       63°C

19mm            180                25°C       27°C

The “first 55°C” and the “second 55°C” refer to the temperature range below the chosen annealing temperature.  So, if you choose to anneal at 515°C, the “first 55°C” is from 515°C to 460°C and the “second 55°C is from 460°C to 405°C.  If you choose 482°C as the annealing temperature, the “first 55°C” is from 482°C to 427°C and the “second 55°C from 427°C to 372°C.

Tack fused/ uneven thickness

If your piece is tack fused, you need to treat the annealing rate and soak as though it were twice the actual total thickness. This gives the following times and rates:

Tack fused

Dimension (mm)                                Cooling rate

Actual     Calculated       time (mins)     1^st 55°C   2^nd 55°C

6            12                 120                55°C       99°C

9            18                 150                37°C       63°C

12          25                 180                25°C       27°C

15          30                 300                37°C       63°C

18          38                 360                7°C         12°C

Contour fusing required firing as though the piece were 1.5 times thicker.  Sharp tack or laminating requires 2.5 times the the actual thickness.

Fusing on the floor of the kiln

There is a further possible complication if you are doing your fusing on the kiln floor, or a shelf resting on the floor of the kiln.  In this case you need to use the times and rates for glass that is at least 3mm thicker than the piece actually is. 

Thus, a flat 6mm piece on a shelf on the floor would use the times and rates for 9mm: anneal soak for 90 minutes, anneal cool at 69°C to 427°C and then at 124°C to 371°C.  It would be safest if you continued to control the cooling to room temperature at no more than 400°C per hour.

But if it were a tack fused piece of a total of 6mm you would use the times and rates for 18mm.  This is using the rates for twice the total thickness plus the additional 3mm for being on the base of the kiln.  This gives the times and rates as being an anneal soak of 360 minutes and cooling rates of 7°C to 427°C and 12°C to 370, followed by 40°C per hour to room temperature.  Any quicker rates should be tested for residual stress before use.

Source for the annealing and cooling of fused glass

These times and rates are based on the table derived from Bullseye research, which is published and available on the Bullseye site.   It is applicable to all fusing glass with adjustments for differing annealing soak temperatures.

Annealing over multiple firings

It has been recommended by this widely followed person that the annealing soak should be extended each time subsequent to the first firing.  I am uncertain about the reasoning behind this suggestion. But the reasons for discounting it are related to adequate annealing and what is done between firings.

If the annealing is adequate for the first firing, it will be adequate for subsequent firings unless you have made significant alterations to the piece.  If you have added another layer to a full fused piece, for example and are using a tack fuse, you will need to anneal for longer, because the style and thickness have been altered.  Not because it is a second firing.  If you are slumping a fired piece, the annealing does not need to be any different than the original firing.

The only time the annealing needs to be altered is when you have significantly changed the thickness of the piece, or the style of fusing (mainly tacking additional items to the full fused piece).  This is when you need to look at the schedules you are planning to use to ensure your heat up is slow enough, that your annealing soak is long enough, and the cool slow enough for the altered conditions.

Determining the annealing point of unknown glass

You don’t have to guess at the annealing temperature for an unknown glass.  You can test for it.  It is known as the slump point test.

This test gives the softening point of the glass and from that the annealing point can be calculated.  This test removes the guess work from choosing a temperature at which to perform the anneal soak. The anneal temperature is important to the result of the firing.  This alone makes this test to give certainty about the annealing temperature worthwhile.

You can anneal soak at the calculated temperature, or at 30°C below it to reduce the anneal cool time.  This is because the annealing can occur over a range of temperatures.  The annealing occurs slowly at the top and bottom of the range. But is at least risk of "fixing in" the stress of an uneven distribution of temperature during the cool when the annealing is done at the lower end of the range.

Do not be fooled into thinking that CoE determines annealing temperatures.  Use published tables, especially the Bullseye table Annealing for Thick Slabs to determine soak times and cooling rates.  Use the standard test for determining the softening and annealing points of unknown glasses.

Further information is available in the ebook Low Temperature Kiln Forming.

Effects of Dams on Scheduling

 I recently made a statement about the effects of various dam materials on the 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.  I found I was wrong.

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.

 

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.

 

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 is cooler than the material being compared and hotter with negative numbers.

 

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 annealing at similar rates – generally 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.

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 glass by the dams to give unequal temperatures across the glass with the risk of inadequate annealing.  I suggest the soak should be extended to that for glass of 6mm thicker than actual to account for this.

·        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.

Scheduling Effects 

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 by 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 perform the best of the three tested materials.

 

Removing Shelves for Slumping

There are those who advocate removing the kiln shelf(s) before slumping.  The advantages claimed include:

Better heat distribution around mould.  The shelf acts as a heat sink. During the firing the shelf absorbs heat and during the cooling the heat is released, so slowing the cool down. 

Additional height. For kilns with little head room, greater height is provided by this practice.

Observations

My observations on this practice lead me to some questions about the necessity, desirability and in some cases the practicality of it.

Elevation of mould above the shelf

This is a widely recommended practice.  I haven’t found the need, but many people do.  One of the points of this is to allow increased air circulation around the mould and under the bottom.  Another is to let air out from under the bottom of the mould to avoid creating air pockets between the mould and the glass.

If the elevation of the mould allows air circulation, what is the necessity to remove the shelf?  There is air circulation around the bottom of the shelf and of the mould. If the mould is placed on the floor of the kiln, the mould will still need to be raised from the bed of the kiln to allow air circulation under the mould. Of course, if the kiln does not have enough space for the height of the mould, it will be necessary to remove the shelf, but not for circulation purposes.

There is also the fact that the floor of the kiln is most often made of refractory bricks even if the walls and top are of refractory fibre.  This also is a heat sink.  I don’t see the advantage of removing the shelf to avoid a heat sink when the base of the kiln works in holding heat in the same way as the shelf.

Difficulty of removing shelves from some kilns

It is difficult to remove shelves from many kilns.  This can be avoidance of damage to the thermocouple; difficulty of getting fingers around the shelf; weight; size; or even depth of the kiln.  It is impractical to remove the shelves from kilns of this nature.  It is still possible to get a good slump in these kilns.

Uneven cooling of the glass

Research shows long soaks lead to a cooler bottom of the glass than top during the anneal – sometimes greater than the +/- 5°C for adequate annealing.  This is a consequence of the fact that the hot air above the glass is not balanced by the same amount of heat below the glass.  So, there may be good arguments for retaining that heat sink of a shelf under the mould to more evenly balance the cooling of the upper and lower surfaces of the glass during the anneal soak and cool.

Height

I don’t have any argument that when extra height is needed, as removing the shelf will provide some.

Some consideration needs to be given on whether to remove the kiln shelf when slumping.  Research implies that increased cooling of the bottom of the glass may go outside the parameters for the even cooling of the glass.

Altering Annealing Temperatures

Sometimes  it is discovered that a kiln is firing hotter than other kilns, and you need to alter your process temperatures from the generally presented ones.  That your kiln is firing hotter than others is when you recognise the tack fusing profile of your tack fused piece is rounder than expected. 

Altering process temperature and soak times

There are two things you can do.

1)  Reduce the time at the temperature.  If the recommended schedule has the process work being done at 780°C for 15 minutes and the glass is too rounded or more like a contour fuse, you can reduce the soak time to 5 minutes, depending on how over-done the pieces are. 

2)  If the reduction in soak at process temperature does not work, then you can begin to reduce the process temperature.  Often only 5°C with a 10-minute soak is enough.  For some kilns it may be as much as 20°C again with a 10-minute soak.

Remember that the speed at which you advance to the process temperature will have an effect.  The slower you go the lower the temperature can be.  The faster you go, generally the higher the temperature needs to be.  There several factors combining to determine which is the right process temperature and soak.  Experimentation and record keeping are required to find just the right combination.

Annealing temperatures in a “hot” kiln

If your kiln fires hot, you do not need to alter the annealing soak temperature.  I have seen the recommendation that when you need to reduce the process temperature you also need to reduce the annealing temperature by the same amount.  This is not so for several reasons.

The first is that reducing the temperature of the annealing soak runs the risk of trying to anneal below the acceptable range.  These are a few paragraphs to explain.

Annealing occurs over a range.  The annealing point is the temperature at which annealing can most quickly occur.  But there is a range during which annealing can occur.  It is generally around 43°C either side of the annealing point.

If you follow the recommendations to anneal in the lower end of the annealing range, it is possible that you will start the annealing soak at too low a temperature by reducing the annealing soak temperature in line with the reduction of the top temperature.

The second is that the temperature measurement is of the air, not the glass.  On cooling, the glass is hotter than the air temperature in the kiln.  The recommendations for the annealing temperature take that into account.  So, reducing the temperature risks straying outside the annealing range.

Example of the annealing of a tack fused piece comparing temperatures of the air to the under tack stack and exposed base during the anneal soak and first cool

You should note that if you are using the Bullseye recommendations to do the anneal soak at 482°C, you already are in the lower end of the annealing range.  The average annealing point of Bullseye remains at 516°C. This new recommendation for the annealing soak is 34°C below the annealing point and any reduction of more than 9°C will put your anneal soak outside the annealing range, meaning that your anneal will be inadequate, no matter how long you soak there.

The third element relates to the annealing range.  The anneal soak can occur anywhere within that range. But the practical measure is to soak at, or below, the annealing point.  If your kiln fires hot, you do not need to alter the annealing soak temperature.  It will not matter if the glass is in fact hotter at the annealing soak than in some other kilns. 

It does not matter, because the soak at the annealing point, or lower in the range, is to equalise the temperature throughout the glass piece. The annealing point is not some magic number or temperature that sees to producing a sound piece of glass.  The soak at annealing point is to equalise the temperature to + or - 5°C within the glass.  This is referred to by the technically minded as Delta T = 5°C, or in symbols as Δ T = 5°C.  Bullseye has published a table that gives practical information on the length of soak required for this temperature equalisation for different thicknesses.

Once the temperature is equalised within these limits, you can begin the anneal cool.  This is an essential part of annealing and is designed to maintain the equality of temperature differentials during the cooling.  The rate of cooling is directly related to the length of the temperature equalisation soak required for the piece which in turn is related to the thickness of the piece.  This forms the fourth reason that starting the anneal soak slightly higher than recommendations, will not affect the annealing process adversely. The first slow cool is essential to achieving a sound piece as it maintains this small differential in temperature during the early part of the cooling into the brittle phase of the glass.

Annealing Temperatures in a Cool Kiln

Exactly the same reasoning process is applied to both hot and cool firing kilns.  You do not need to alter the anneal soak, even though it means you will start the temperature equalisation at a slightly lower temperature than the published schedules.  This is because you have to increase the top temperature to get the effect you want and so would also be annealing in a cooler kiln.  Since you are measuring the air temperature, the glass temperature will be above the air temperature and will still be in the safe annealing range.

Summary

The reasons annealing temperatures do not need to be altered if you kiln fires hot or cool are related to:

·        annealing range

·        air temperature measurements

·        rate of the anneal cool

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

Long Annealing Times

I’ve noticed a lot of people are using long annealing times for full or contour fused glass of 6mm thickness.  Whether this comes from the advice for tack fused pieces to be annealed as though there were twice the total thickness of the piece, or not, is uncertain. 

Temperature Equalisation

This term employs the concept of what happens at the annealing point chosen for your glass.  The soak at the annealing point - or even at some degrees below - is the process of ensuring the glass is all at the same temperature before proceeding to the annealing cool.

A “flattish” fused piece of 6mm needs only a half hour when the temperature equalisation soak is at the annealing temperature.  When using a temperature equalisation soak about 30°C-35°C below the annealing temperature – as Bullseye, and now Wissmach, does – you may need an hour.

Tack fusing is much more difficult to anneal properly than a full or contour fuse.  The general rule of thumb has become that you must schedule the firing as though it is twice the actual maximum height. 

E.g., if you have a two-layer piece with other pieces distributed around this flat base, you have a 9mm thick piece in total height. Scheduling for this piece as though it is 18mm thick requires longer soaks and slower cools.  In this case, you should schedule for a three-hour soak at temperature equalisation, whether using the annealing point or a lower temperature. 

Other kilnformers have found that firing tack fused pieces as though they are one and a half times the actual maximum thickness provides a perfectly adequate result.  It will be up to each person to decide which approach they take.  Once the current lock down is over, I am going to do some work on the actual requirements.

Annealing Cool

People seem to cool a tack fused piece at a rate suitable for thinner pieces. It appears that the general practice is to use rates suitable for a for 12mm thick piece regardless of the calculated thickness. 

But the annealing cool is still too rapid for a tack fuse.  If you need an anneal soak for 19mm, you need to cool for that thickness too. An annealing cool for a 19mm piece is one half of the rate of that for a 12mm piece.  It is during this too rapid cool that stress can be induced, especially on a tack fused piece.

The temperature equalisation soak and anneal cool can take as much as twice - or more – as long as a flat piece of the same total height.  This needs to be allowed for in the scheduling of the firing.


The length of anneal soak and the rate of anneal cool both need to be related to the appropriate thickness.  There is the Bullseye chart for annealing thick slabs the rates of which, can be applied to any fusing glass.  You only need adjustment to the temperatures for your glass, if not Bullseye.

Further information is available in the ebook Low Temperature Kiln Forming.

Extending the Annealing Soak

Assertion

There is a common assertion that you need to extend the annealing - or temperature equalisation soak - on each firing after the first.

The rationale for this is never fully explained.  Possibly it comes from the fact that you need to reduce the rate of advance on already fused pieces over the unfused lay-up. It is also possible the rationale is that since you need to slow the rate of advance, so you need to extend the anneal soak.

However, if this is the rationale, it is rarely followed through on the anneal cool in the detailed schedules in these instances.  Generally, the cool on extended soaks is the same as on the first anneal, but extending the anneal should apply to the cool too, according to this kind of rationale.

Facts

At normal kilnforming temperatures, the anneal for the determined thickness (allowing an adjustment of actual thickness for tack fused pieces) is suitable even for multiple firings.

Once you go up into the temperatures for melts, there is reason for more caution. There is the risk that the high temperatures – especially for hot coloured opals – may induce a little incompatibility.  A longer soak for these at a lower temperature may be considered desirable.  Even so, this does not need to be extended at each subsequent firing. 

You are annealing at each firing for the thickness – actual or calculated – of the piece, not for the number of firings.

Extending the length of the anneal on each subsequent firing is not necessary unless additional thickness or complexity has been added.

Further information is available in the ebook: Low Temperature Kiln Forming.

Scheduling to Room Temperature

Why Schedule the kiln to room temperature? The kiln will cool slowly enough at the final stages.

How do you know?

Relatively large thick pieces need slow rates of cooling below 370°C.  Complex tack fused pieces require slow cooling rates as well as the long annealing soaks. These required rates of cooling may be slower than your kiln’s unpowered rate of cooling.

This means you need to know the natural cooling rate of your kiln from 370°C down to room temperature to be sure you are cooling at a suitable rate. The method described in this blog post gives you information on how to calculate the natural cooling rate of your kiln.

I program my firings to about twice room temperature. Yes, the kiln does not turn on much during that time,  but when I crack my kiln open to speed the cooling, the switching on of the relay tells me I am cooling faster than programmed, and I can reduce the size of the opening to avoid too rapid cooling of the piece.

The following chart is a way to assist in recording your kiln’s unassisted cooling temperatures against time to give you the natural cooling rate at various temperatures.

Natural Cooling Rate of the Kiln

Kiln Name:					Cooling Rate
observ'n	Time (hr:min)	Temperature	Difference		rate/min	rate/hr
1	:		Time (mins)	Temp.	=temp/mins	.=temp/min*60
2	:
3	:
4	:
5	:
6	:
7	:
8	:
9	:
10	:
11	:
12	:
13	:
14	:
15	:
16	:
17	:
18	:
19	:
20	: