Reducing Annealing Time for Float Glass

Credit: Bullseye Glass Co.

Annealing float glass seems take a long time. The annealing point (Tg) is higher than most fusing glasses, although float glass is part of the family of soda lime glass. This group of glasses should be cooled slowly from annealing temperature to 427ºC/800ºF and below to reduce risks of thermal shock.  This makes a greater temperature range over which to  anneal float than fusing glasses, consequently it extends the cooling time and increases energy expenses.

It does not have to be this way.  Annealing of glass takes place over a range.  This range extends below the published annealing point (Tg).  This is the temperature at which equalisation can most quickly take place, but it is not as energy efficient as starting in the lower range.  Annealing points (Tg) vary between manufacturers, but these are some of them:

Pilkington Optiwhite              559ºC/1039ºF

Pilkington Optifloat               548ºC/1019ºF

USA float (typical)                548ºC/1019ºF

Australian float (average)      550ºC/1022ºF

The annealing range extends to a practical 38ºC/68ºF below the Tg temperature.  Annealing at a lower temperature can be as effective at the lower portion of the range as at the Tg.  Using a lower annealing soak temperature reduces the temperature range of the first cooling stage by as much as 38ºC/68ºF, and reduces the cooling time without increasing risks of breaking.  It also creates a denser glass according to scientific research.  Denser glass is arguably a stronger glass.

This means that the annealing of float glass can take place at the following reduced temperatures:

Pilkington Optiwhite              521ºC/971ºF

Pilkington Optifloat               510ºC/900ºF

USA float (typical)                510ºC/900ºF

Australian float (average)      512ºC/954ºF

 

This reduces the first cooling stage for 12mm/0.5” Pilkington Optiwhite from 2 hours 24 minutes to 1 hour 43 minutes.  Forty-one minutes may not seem much but in electricity costs is significant.  Also using the Bullseye concept of a three stage cooling, further savings can be made.  Their research shows the second cooling stage to 371ºC/700ºF can be increased by 1.8 times the first cooling rate, saving further time and energy.  The chart which shows these rates is Annealing Thick Slabs -  Celsius and - Fahrenheit.

More information on annealing is available in the ebook Annealing: Concepts, Principles and Practice

Annealing float glass at the lower part of the annealing range reduces the time and cost of firings.

Scheduling with the Bullseye Annealing Chart

This post is about adapting the Bullseye chart Annealing Thick Slabs to write a schedule for any soda lime glass as used in kilnforming.

I frequently recommend that people should use the Bullseye chart for Annealing Thick Slabs in Celsius  and Fahrenheit.  This chart applies to glass from 6mm to 200mm (0.25” to 8”).

“Why should the Bullseye annealing chart be used instead of some other source?  I don’t use Bullseye.”

My answer is that the information in the chart is the most thoroughly researched set of tables for fusing compatible glass that is currently available.  This means that the soak times and rates for the thicknesses can be relied upon.

“How can it be used for glass other than Bullseye?”  

The rates and times given in the chart work for any soda lime glass, even float. It is only some of the temperatures that need to be changed.

"How do I do that?"  

My usual response is: substitute the annealing temperature for your glass into the one given in the Bullseye table.

 "It’s only half a schedule."

That is so.  The heating of glass is so dependent on layup, size, style, process, and purpose of the piece.  This makes it exceedingly difficult to suggest a generally applicable firing schedule.  People find this out after using already set schedules for a while. What works for one layup does not for another.

Devising a Schedule for the Heat Up

There is no recommendation from the chart on heat up.  You have to write your own schedule for the first ramps.  I can give some general advice on some of the things you need to be aware of while composing your schedule.

The essential element to note is that the Bullseye chart is based on evenly thick pieces of glass.  Tack fusing different thicknesses of glass across the piece, requires more caution. The practical process is to fire as for thicker pieces.  The amount of additional thickness is determined by the profile being used.  The calculation for addition depends on the final profile.  The calculation for thickness is as follows:

• Contour fusing - multiply the thickest part by 1.5. 
• Tack fusing - multiply the thickest part by 2. 
• Sharp tack or sinter - multiply the thickest part by 2.5.

The end cooling rate for the appropriate thickness is a guide for the first ramp rate of your schedule.  For example, the final rate for an evenly thick piece 19mm/0.75” is 150ºC/270ºF.  This could be used as the rate for the first ramp. 

Bob Leatherbarrow has noted that most breaks occur below 260ºC/500ºF.  If there are multiple concerns, more caution can be used for the starting ramp rate.  My testing shows that using a rate of two thirds the final rate of cooling with a 20 minute soak is cautious.  In this example of a 19mm piece it would be 100ºC/180ºF per hour.

Even though for thinner pieces the rates given are much faster, be careful.  It is not advisable to raise the temperature faster than 330ºC/600ºF per hour to care for both the glass and the kiln shelf.

Once the soak at 260ºC//500ºF is finished, the ramp to the bubble squeeze should maintain the previous rate.  It should not be speeded up.  The glass is still in the brittle phase.

After the bubble squeeze you can use a ramp rate to the top temperature of up to 330C/600F.   AFAP rates to top temperature are not advisable.  It is difficult to maintain control of the overshoots in temperature that are created by rapid rates.  

The top temperature should be such as to achieve the result in 10 minutes to avoid problems that can occur with extended soaks at top temperature.

In the example of an evenly thick 19mm/0.75” piece a heat up full fuse schedule like this could be used:

• 150ºC/270ºF to 566ºC/1052ºF for 0 minutes
• 50C/90F to 643C/1191F for 30 minutes
• 333ºC/600ºF to 804ºC/1479ºF for 10 minutes

 

If a more cautious approach to the heat up is desired, this might be the kind of schedule used:

 

• 100ºC/180ºF to 260ºC/500ºF for 20 minutes
• 100ºC/180ºF to 566ºC/1052ºF for 0 minutes
• 50C/90F to 643ºC/1191ºF for 30 minutes
• 333ºC/600ºF to 804ºC/1479ºF for 10 minutes

This approach is applicable to all fusing glasses.

 

Adapting the Bullseye Annealing Chart

After writing the first part of the schedule, you can continue to apply the annealing information from the Bullseye chart.  The first part of the anneal cooling starts with dropping the temperature as fast as possible to the annealing temperature.

The method for making the chart applicable to the annealing is a matter of substitution of the temperature.  All the other temperatures and rates apply to all fusing glasses.

Use the annealing temperature from your source as the target annealing  temperature in place of the Bullseye one.  The annealing soak times are important to equalise the temperature within the glass to an acceptable level (ΔT=5ºC).  The annealing soak time is related to the calculated thickness of the piece.  This measurement is done in the same way as devising the appropriate rate for heat up. 

Applying the Cooing Rates

Then apply the rates and temperatures as given in the chart.  The three stage cooling is important.  The gradually increasing rates keep the temperature differentials within acceptable bounds with the most rapid and safe rates.

The temperatures and rates remain the same for all soda lime glasses – the range of glass currently used in fusing, including float glass.  The soak time for the calculated thickness of your glass piece will be the same as in the Bullseye chart.  

This means that the first cooling stage will be to 427ºC/800ºF.  The second stage will be from 427ºC/800ºF to 371ºC/700˚F.  And the final stage will be from 371ºC/700˚F to room temperature.

I will repeat, because it is so important, that the thickness to be used for the anneal soak and cooling rates for your schedule relates to the profile you desire.  A fuse with even thickness across the whole piece can use the times, temperatures, and rates as given in the chart as adapted for your glass.  The thicknesses to use are for:

Contour fusing - multiply the thickest part by 1.5. 

Tack fusing - multiply the thickest part by 2. 

Sharp tack or sinter - multiply the thickest part by 2.5.

An annealing cool schedule for 19mm/0.75" Oceanside glass is like this:

• AFAP to 510˚C/ 951˚F for 3:00 hours
• 25˚C/45˚F to 427˚C/800˚F for 0 time
• 45˚C/81˚F to 371˚C/700˚F for 0 time
• 150˚C/270˚F to room temperature, off.

Many will wish to turn off the kiln as early as possible.  This is not part of best kilnforming practice.  If you still wish to do this, the turn off temperature must be related to the thickness and nature of the piece.  To turn off safely, you need to know the cooling characteristics of your kiln.  This can be determined by observing the temperature against time and then calculating the kiln’s natural cooling rate. And then applying that information to cooling the kiln.

 

The best source for devising schedules is the Bullseye chart for Annealing Thick Slabs.  It is well researched and is applicable with little work to develop appropriate schedules for all the fusing glasses currently in use.

 

 

Float annealing

As a result of various memory failures, I've done a bit of searching on the annealing of float glass.  There are now various compositions of float glass and with different coatings for various applications.

This leads to a variety of annealing points for Pilkington float glasses. The search led to various hard to find documents, which indicate a range of annealing temperatures between 548°C and 559°C. This is not a huge range, so anywhere between 548°C and 560°C can be taken as the annealing point. Pilkington indicate that optifloat has an annealing point of 548°C

The strain point seems to be mostly between 525°C and 530°C for all the varieties.  This indicates the temperature equalisation soak should not be less than 535°C.

The conclusion seems to be that annealing should have a temperature equalisation soak between 550°C and 535°C. It will not matter much where you choose, but remember that the closer to the strain point you do the temperature equalisation, the longer the soak should be.  The length of soak at 535°C can be determined by use of the Bullseye chart for Annealing Thick Slabs. This gives the times and rates for the anneal cooling of glass by thickness.  The temperatures need to be changed, but otherwise the information can be applied.

The softening point seems to be 725°C for all the glasses. This is a good low temperature for slumping.

Float Annealing Temperatures

Float glass annealing temperatures vary quite a bit from one manufacturer to another; and even within one manufacturer’s product line.

Comparisons of various float glasses

Some companies are more informative that others.  Pilkington are one of the more open of European glass manufacturers on various bits of information.

Pilkington Float

CoLE 83 *^10-5

Softening point:  715°C

annealing point:  548°C

strain point: 511C

Pilkington Optiwhite ™

Softening point:  ca. 732°C

annealing point:  ca. 559°C

strain point:  ca. 526°C

There is a difference of 11C between two of the Pilkington product lines for the annealing points.  The softening and strain points are slightly wider.

Glaverbel, a Belgian company, restricts their information to CoLE and the softening point.

CoLE 91 * 10^-5

Softening point: 600°C

Saint-Gobain, a French company, shows some more of the variation in the product lines, although they do not give specific annealing points for the different products.

CoLE 90 * 10^-5

annealing range:  520 - 550°C

Low E glass

softening – 840°C

strain - 617°C

R glass (sound reducing)

softening – 986°C

strain - 736°C

D glass (decorative)

softening point – 769°C

Compatibility

Even this small sample of float glasses shows there is a significant difference between manufacturers for the softening, annealing and strain points.  This means that, unless you are sure of the glass merchant’s source of glass, you will need to test each batch of glass for compatibility with previous batches, if you are combining from different suppliers.

I included the CoLE numbers (which all the manufacturers specified as an average change in length for each degree C increase in temperature from 0 to 300°C) to show the variation and to challenge anyone to find Bullseye and Saint-Gobain or Glaverbel compatible with each other.  My experience has shown that the Optul coloured frit and confetti is more likely to be compatible with Pilkington than the other two.

Annealing

I have been beginning my annealing of float glass at 525°C.  This little bit of literature research shows that my annealing soak should be starting higher, possibly at 540°C, certainly no lower than 530°C.  Other areas of the world may find their float glass has significantly different annealing ranges.

Tin Bloom

Using float glass sometimes produces partial clouding as though devitrification were present. Although float glass is prone to devitrification, not all the cloudy film on the surface is due to devitrification.

Float glass, which these days, is almost all clear smooth glass, gets its name from the process of floating the glass on molten tin. The tin in compression gives an apparent devitrification effect which is called tin bloom.

it is different from devitrification, to which float glass is particularly subject. Devitrification sprays and solutions will not have an effect on this surface defect called tin bloom. 

When the tin layer is stretched, it does not create a tin bloom on the surface.  Therefore, it is important to have a means to detect which is the tin surface.  Always fire the glass with the tin in the same relative location to each other.  I.e., on several layers of glass have all the tin side down or all up, but not mixed. 

This example of a test by Glass Art by Margot shows the tin bloom on the outer portions of the platter where the tin side was up, causing the tin too be compressed and show.  The flatter portion of the piece did not show this tin bloom as there was not the same extent of compression. You can visit the description of the experiment here.


When forming the glass (slumping, draping, kiln carving) make sure the tin sides will be stretched rather than compressed.  Of course, you can take advantage of the tin bloom by controlling the compression of the tin layers.

Float Glass

A reported 90% of the world's flat glass is produced by the float glass process invented in the 1950's by Sir Alastair Pilkington of Pilkington Glass. Molten glass is “floated” onto one end of a molten tin bath. The glass is supported by the tin, and levels out as it spreads along the bath, giving a smooth face to both sides. The glass cools as it travels over the molten tin and leaves the tin bath in a continuous ribbon. The glass is then annealed by cooling in a lehr. The finished product has near-perfect parallel surfaces.

An important characteristic of the glass is that a very small amount of the tin is embedded into the glass on the side it touched. The tin side is easier to make into a mirror and is softer and easier to scratch.

Float glass is produced in standard metric thicknesses of 2, 3, 4, 5, 6, 8, 10, 12, 15, 19 and 22 mm. Molten glass floating on tin in a nitrogen/hydrogen atmosphere will spread out to a thickness of about 6 mm and stop due to surface tension. Thinner glass is made by stretching the glass while it floats on the tin and cools. Similarly, thicker glass is pushed back and not permitted to expand as it cools on the tin.

More information on float glass in the kiln is here.

Types of Glass

Glass Types by manufacturing method

There are several ways of categorising glass and this overview of glass types looks at the way the glass is manufactured.

Crown Glass

Crown glass is the oldest method of producing sheet glass and continued to be used until the 19^th century.  This method consisted of blowing a very large bubble of glass.  It was then spun rapidly over a pit until the bubble collapsed into a disc that ranged from 1500mm to 1800mm diameter.  

This gave the thinnest and least marked glass at the outer portion of the disc.  The centre was the thickest and became known as the bullseye.  The glass was cut to provide the best use of the disc.  This limited the size of panes to what could be cut from the disc.  Diamond shapes were often cut from the remainder and the central bullseye was used in less expensive glazing.

Corning Museum of Glass

Cylinder Glass

Cylinder Glass is a handmade process that includes broad sheet glass. It was widely used from the 17^th to the 19^th century, and now is limited to a few manufacturers.  

"Among the Glass Workers" Harry Fenn, 1871

An elongated bubble was blown.  The top and bottom of the bubble are broken off and annealed.  Later the cylinder is placed in the lehr for reheating.  It is scored and when it breaks open along the score, the glass is flattened. Characteristically, it has a gradation of thickness with thicker edges where the top and bottom of the cylinder were cut off.

From IdoStuff

Flashed Glass

A development in cylinder glass was to make the bubble of two colours, with the dark colour gathered first and then encased in clear (or sometimes other pale colours) and blown into a cylinder.  This made dense colours more transparent and enabled more detail through abrading and etching.

Drawn Glass

Industrialisation of glass production began with the development of drawn glass.  This method of mass production of window glass was invented and developed by Emile Fourcault in Belgium. Full scale production began in the early 1900’s.  

The glass is drawn upwards from a vat of molten glass until it cools enough to be cut into sheets at the top of the tower.  The process is subject to slight variations in thickness due to uneven cooling and gravity. It enabled much larger panes of glass without the astragals that are common in Georgian and later houses.  It was the most common method of producing window glass until the 1950’s.

Table Glass
Table glass is the process of putting molten glass onto a flat surface (the table) and rolling the glass flat.  This has been used from the latter part of the 19^th century to the present.  It enables textures to be pressed into the glass from the rolling cylinder.  It is easier to produce streaky and wispy glass by combining different colours on the table. 

Kokomo Glass Co.

This can be done as single sheets or further mechanised to roll out long ribbons of glass.  This is now mostly referred to as machine or hand rolled glass depending on the amount of mechanisation.

Float Glass

The glass that we now rely on for large clear windows began with the development of experiments by Alastair Pilkington and the company named after him.  This consisted of floating near molten glass on molten tin, hence the name, float glass.  This has been the standard method of glass for windows since the 1950’s.

Slumping Glass that is not Tested Compatible

Is it Possible?

It is possible to slump unknown glass. This glass might be art glass, window glass, bottles, or any other glass whose characteristics are unknown by you.  There are some suggestions about the characteristics of some glasses in this post that can be used as a starting point.

Preparation of the Glass

Prepare the edges to their final finish before slumping.  This because the slumping temperature will not be enough to alter the finish of the edge significantly.  This preparation can be done with diamond hand pads, or wet and dry sandpapers.  Start with a relatively coarse grit. You may wish to do the initial shaping on your grinder. This will be between 80 and 100 grit.  Continuing with a 200 grit and working your way through 400 and then 600 grit will give you an edge that will become shiny during the slumping.

Cleaning

Clean thoroughly.  This is especially important when using glass that is not formulated for fusing.  Devitrification is more likely on these glasses.  Water with a drop of dishwashing liquid can be enough unless your water has high mineral content.  Then distilled water or a purpose made glass cleaner such as Bohle or Spartan should be substituted.  Finish with a polish to dry with clean paper towels. More here. 

Firing the Slump

Fire up slowly.  You should advance at about 100°C to 150°C per hour.  Set your top temperature around 630°C for a simple slump, for soda lime stained glass.  For bottle or window glass you will need a temperature closer to 720°C although the also are soda lime glasses.

It is best to start with simple curves, as there are fewer difficulties in determining what the glass is doing.  It will help you to learn the characteristics of the glass before you tackle the difficult stuff, such as compound curves or texture moulds.

Observation

It is necessary to observe the progress of the slump as you do not yet know the slumping temperature.  You want to know when the glass begins to deform so that you do not over fire.  Start watching the glass at about 10 minute intervals from about 580°C for stained glass and 680°C for window and bottle glass.  There is not much light in the kiln at these temperatures, so an external light is useful.  You can also observe the reflections of the elements on the glass.  When the image of the elements begins to curve, you know the glass is beginning to bend.

Altering the Schedule

Soak for at least 30 mins at the temperature when the glass begins to visibly drop. This may or may not be long enough.  Continue checking at 5-10 minute intervals to know when the slump is complete.  If the glass is completely slumped before the soak time is finished, advance to the next segment.  If not fully slumped, you need to extend the soak time. This means that you need to know how to alter your schedule in your controller while firing.  Consult your controller manual to learn how to do these things.

Stop the soak when complete and advance to the anneal. Continue the slumping soak if not complete after the 30 mins.  In some cases, you may need to also increase the temperature by 5-10°C.

Annealing

The annealing point will be about 40°C below the point that the glass visibly starts the slump. If you want a more accurate determination of the annealing point, this post gives information on how to conduct a test to give you both the slump temperature and the annealing point.  It also helps to determine the lower part of the tack fusing range (the lamination state), since it is not far above the slumping point that you will observe.

The annealing soak for a single layer, 3mm glass need not be long – 15 to 30 minutes.  The annealing cool can be as fast as 120°C down to 370°C.  For thicker glass and slumped bottle glass you will need a longer soak – 30 to 60 minutes – and a slower cool.  The annealing cool in this case could be about 60°C/hour to 370°C.  You can turn the kiln off at 370°C, if you wish, or keep the temperature controlled to about 50°C.  The rate for the final cooling can be approximately double the first cooling rate.  For a single layer of stained glass this could be 240°C, and for thicker glass about 120°C

Use of Untested Glass - Kiln Forming Myths 22

You must use art glass rather than recycled glass.

This seems to refer to the use of untested glass in kiln forming.  If you are going to use untested glass for kiln forming, it does not much matter which you use.  Because, in every case you will need to test for forming and annealing temperatures to be able to make use of the glass with unknown properties. 

Of course, people use glass that is not tested fusing compatible in many circumstances.  Float glass is frequently used in many kiln forming applications.  And bottle glass is of very little different in composition.  So-called art glass can be used in a variety of ways also.  There are many other variations of glass including handmade, casting, lamp working, and borosilicate, among others.  Each has their own set of characteristics, which overlap with each other.  The forming and annealing temperatures must be determined to enable you to use them. Some of this information is often available from the manufacturer’s web site or other sources.  Many times you have to do the testing for yourself.  One guide to help determine the critical temperatures is here. 

One characteristic that all untested glasses share is a tendency to devitrify by the second or third firing, so attempting to get the most work done in the fewest firings is a good idea.  This tendency to devitrify is frequently shown when manipulating bottle glass.

Float Glass

A question about sharp raised points on the corners of a square bubble plate made of window glass is the occasion to discuss some characteristics of float glass. 

It is necessary with float to find out which is the tin side and which is the air side. The tin layer of the glass produces a bloom that resembles devitrification when compressed. Put the tin side down for a slump.  If you slump with the tin side up, you will create a tin bloom by compressing the tin. If the tin is on the bottom, you will be stretching the tin and so avoid the tin bloom.

Sharp, pointed and raised corners are the result of devitrification.  Devitrification is the crystallisation of glass. Mild devitrification appears to be dirty streaks across the surface. Extreme devitrification produces a crumbling glass surface. Raised, sharp corners are the result of intermediate devitrification. The tin side does not protect against devitrification.  It does provide a separating action when against the shelf, although kiln wash is still needed.  Float glass devitrifies easily. I have only ever been able to get two firings without devitrification.

Cleaning is of great importance in avoiding devitrification. Clean well with only a little detergent, rinse and then polish dry with paper towels. Any residues left on the surface will promote devitrification.

A general way of reducing sharp corners is to nip or round the corners with diamond pads. I nip the corners - it is quicker and does not leave any microscopic pits for devitrification formation.

Paint, stains and enamels will interact with the tin to produce variants of the colours.  Stains most often become darker than when put on the air side. Powder, frit and mica will not usually react to the tin.

Remember, float glass is not manufactured to be a kiln forming glass.  You will always be at risk of devitrification.

Float Glass in the Kiln

An important characteristic of float glass is that a very small amount of the tin is embedded into the glass on the side it touched. The tin side is easier to make into a mirror and is softer and easier to scratch than the air side. The characteristic of float glass having a molecular level of tin left on the “tin side” but not the “air side” is important to distinguish. There are short wave UV light sources to help determine this. The tin side gives a whiter glow than the air side. If any forming of the glass is planed after fusing, the tin side needs to be on the side being stretched, as when in compression the tin side will show a “tin bloom” similar to devitrification.

If the tin side is down on both sheets, and it is slumped into a mould there will be no tin bloom because the tin layer is stretched. If the tin side is up on both sheets and it is slumped into a mould there will be tin bloom because the tin layer is compressed. If you have placed the tin sides together, or on both the top and bottom, one of the tin surfaces will be in compression and so will show tin bloom. This is often mistaken for devitrification, and no amount of any devitrification solution will help.

A borax solution can help with the devitrification on float glass in some circumstances. It is not a perfect solution. This is because tin bloom and devitrification are often not distinguished correctly. But a high level of cleanliness and polishing the glass until squeaky clean is the best start.

The heat characteristics of Float glass depend in large part on which company manufactures the glass being used, so the temperature characteristics are given in ranges.

The softening point is around 760C

The annealing point is around 560—540C

The strain point is around 515-495C. The strain point being the temperature below which no further annealing occurs, although the glass can still be thermally shocked below this range.

Due to the robustness of float glass, it can be fired with a quicker initial temperature rise than glasses formulated for kiln forming. The down side is that it devitrifies very easily and very badly. Rarely can you perform more than two firings before the devitrification begins to become troublesome.

All window glass now seems to be referred to as float glass. However, the float glass process was invented in the 1950’s. Prior to that time, window glass was drawn. Float glass can use more iron in its composition, because it does not have to be drawn up out of a molten vat of glass as the drawn glass did and still does. Float glass is formulated to be stiffer at forming temperatures, whereas the drawn glass has to be flexible due to the mechanical stresses it is put under during the drawing. Except for low iron glass, the float glass has a distinct blue green colour when viewed through the edge. Drawn glass has a variation in thickness and is much paler when viewed through the edge. These visual differences can help distinguish the two kinds of glass, but are not foolproof.

More information on the general characteristics of float glass can be found here.