some other geopolymer experiments and results

[Deleted]

Hi everyone,
I've been reading and experimenting geopolymerisation for some months now, with many failures and some success. My goal was to produce an insulating, heat-resistant castable in order to cast some cylindrical heat-risers. I'm posting some of the results, hoping they might be useful to some of you.

Pictures may be found here : webcloud.zaclys.com/index.php/s/7hMokc2rhq3iYqh

First trials :
My basic mix was the one that @karl gave in another post : 10 parts (in dry weight) of a fine alumino-silicate, 1 part of sodium hydroxyde powder, 1 part of natural hydraulic lime. Then I added vermiculite and/or fine sawdust to give it insulating properties.
I tried at ambient temperature :
    * natural pouzzolan with a granulometry between 0 and 1 mm : no hardening
    * red chamotte, granulometry 0-1 mm : hardened but still brittle
    * very fine disaggregated metamorphic rock : hardened but still brittle

Second row of trials :
I synthetised sodium silicate from 4.5 parts (in dry weight) of Sodium Hydroxyde, 5.5 parts of diatomaceous earth, 20 parts of water. I used it successfully as a deflocculant, but not in my insulating application.
I then tried at ambient temperature :
    * 1.5 L of vermiculite + 150 mL of sodium silicate, weak hardening
    * 1.5 L of vermiculite + 300 mL of sodium silicate, weak hardening

Third row of trials :
10 parts of 0-1mm red chamotte, 1 part of sodium hydroxyde powder, 1.5 part of water. I let them rest for 3h at ambient temperature. I then added 1 part of natural hydraulic lime and, after that, vermiculite hydrated with 15-20% water in volume. It allowed me to cast the combustion core of a 18 cm (7") rocket. The rocket has been in daily use for nearly three months. The geopolymer is quite resistant to abrasion and with the time the texture became like glass. The moulding has been done with wood that has been fired in the rocket, so no unmoulding was required.

In detail : 30 kg of 0-1 mm chamotte, 3 kg of sodium hydroxyde, 4.5 L of water, 3 kg of natural hydraulic lime, 30 L of vermiculite pre-hydrated with approx. 5 L of water. This allowed to cast a volume of 30 L.

Fourth row of trials :
12 parts of 0-1mm red chamotte, 1.5 part of sodium hydroxyde powder, 1.5 part of water. I let them rest for 3h at ambient temperature. I then added 1 part of natural hydraulic lime and, after that, vermiculite hydrated with 15-20% water in volume and hydrated sawdust.

In detail : 20 kg of 0-1 mm chamotte, 3 kg of sodium hydroxyde, 3 L of water, 2 kg of natural hydraulic lime, 20 L of pre-hydrated vermiculite, 20 L of pre-hydrated sawdust. Cooked at approx. 100°C for 48h, and then fired in place in the rocket.

Conclusion : very insulating, barely sticky when casted but holds together quite well after the first cooking. Not very resistant to abrasion, but it's not a problem for a heat riser. Been in daily use for two months, no sign of degradation at all.

Fifth row of trials :
10 parts of 0-0.2mm kaolin chamotte, 1 part of sodium hydroxyde powder, 1.5 part of water. I let them rest for 3h at ambient temperature. I then added 1 part of natural hydraulic lime and, after that, hydrated sawdust (1L of sawdust per kg of chamotte).
Casted in an oiled metallic mould, cooked at 100°C overnight and then fired at 250°C for a few hours in the previoulsy made, rocket-fired oven.

Results : unmoulding is impossible. The mix is quite sturdy but sticks too much to the mould. I tried to improve my mould by giving it clearance angles but it lead to another failure.

Some remarks on geopolymer :
* I've found that an alumino-silicate is fine enough for geopolymerisation when it makes a paste once hydrated. If it stays crunchy like sand the mix won't hold.
* I found that lime is not necessary. It accelerate the setting, but if you have enough time you can just cook the mix longer.
* It's better to hydrate the vermiculite/perlite/sawdust before adding it to the mix, so that it won't absorb directly all the water needed for the geopolymerisation.
* the insulating geopolymers i've tried are weak and friable until fired at high temperature (900 °C). When the ceramic bond is created the geopolymer becomes hard and glassy. At medium temperature (<250°C), the geopolymer gets harder but it's still hard to move without breaking it.
* Unmoulding is always very difficult. To me, that is the biggest problem with geopolymer in our applications. It's possible to make cynlindrical moulds with very fine sheet metal as long as the internal diameter of the heat riser is large enough (>180mm).
* It's a good idea to keep a wet cloth and some diluted vinegar nearby to remove spots. I've burned myself too many times and now i prepare it systematically. The acidity of the vinegar will stop the burn of the alkaline geopolymer. Use heavy PVC gloves to manipulate the sodium hydroxyde.
* 0-1mm chamotte costs 0.3 €/kg, 0-0.2 kaolin chamotte costs 0.8 €/kg, sodium hydroxyde powder costs 2 €/kg, hydraulic lime costs 0.4 €/kg, vermiculite/perlite costs 0.16 €/L.

regards,
yasin

[Deleted]

To form strong bonds geopolymer binder needs to be as fine as cement.
Alumino-silicates have a density of about 3g/cm2, thus 1g gives a cube of about 0.7cm
with a surface of about 3cm2(0.7*0.7*6).
Standard portland cement has about 3000cm2/g.

To give an idea:
Each time the cube is divided by 2 the total surface doubles.
To get 3000cm2/g the cube needs to be divided 10 times by 2.

Irregularly shaped particles as created by grinding have a significantly larger surface per volume,
thus less divisions would be required to get a large enough surface.

I have recomented to use 200 Mesh grog which is equal to 0-0.074 mm
0-0.1 mm equals 155 Mesh and 0-0.2 mm about 70 Mesh.

With relatively large particles a significantly smaller part of each particle can be dissolved
and the chemical composition rearranged.

Not to forget heat can travel larger distances in larger particles before it hits a boundary,
thus the conductivity will be higher.

[Deleted]

Kaolinite and quarz have densities of about 2.6g/cm3.
2.6 parts kaolinite and one part water(~38%) will result in a dry porosity of 50% without shrinkage.
In my experiments I have mixed 80% grog 0-0.1 mm, 10 lye and 10% lime with up to 66% of water,
which gives for 100 g solids a volume of about 104 ml with 66 ml pores, which equals a porosity of about 63%.
The mixture cured at ambient temperatures and became still quite hard at a density of about 1Kg per liter.
A similar mixture with Molochite 200 Mesh became even a bit harder.

To prevent cracks geopolymers with such a high water content need to be dryed carefully.
Some short mineral fibers could be added make it more crack resistant.

Only if the density of your fired insulating mixture is significantly lower than 1Kg per liter
it may have a lower thermal conductivity as this simple mixture.

[Deleted]

Thanks karl, very interesting as always. I tried to use the most common materials I could find. Sawdust will be fired off so the density will be lower after firing.

My biggest problem is about unmoulding, do you have ideas on how to cast a cylindrical heat riser without breaking everything on unmoulding ?

A side question : when curing my geopolymer at ambient temperature, I noticed a white stuff migrating on the surface. It doesn't happen at higher temperature. Do you have ideas about what it could be ?

regards,

[keithturtle]

Jul 23, 2016 20:25:40 GMT -8 @yasintoda said:

My biggest problem is about unmoulding, do you have ideas on how to cast a cylindrical heat riser without breaking everything on unmoulding ?

I have successfully used an oil-based form release compound in steel and PVC moulds for these round risers; keep in mind this was for castable refractory and not geopolymer so I cannot speak to that issue. Adequate dry time under heat is a must before attempting release

Material similar to this www.bestmaterials.com/detail.aspx?ID=19544

Turtle

[lawry]

Great report yasintoda! The white stuff that settles on the surface is efflorescence. Karl explained it to me in the other geopolymer thread.

"Karls Geopolymer Experiments"

r.tapatalk.com/shareLink?url=http%3A%2F%2Fdonkey32%2Eproboards%2Ecom%2Fpost%2F21484%2Fthread%2F1269&share_tid=1269&share_fid=106896&share_type=t&share_pid=21484

[Deleted]

Pure geopolymer binder made from very fine grog with lye and lime can become quite hard,
even if it is still full of water.
Other binders eg. perlite or natural zeolites especially if made with waterglass without lime
my require to loose a significant part of the water to get really hard.

As you have saturated the vermiculite with water before mixing it with the other material
it cannot take part in the pozzolanic reaction with lye or waterglass and thus remains
a bunch of loosely layered material, significantly weakening your mixture.

Vermiculite has an inner surface of several square meters per gramm,
which makes it to a quite reactive pozzolan.

You should saturate it with lye or waterglass, not water, and then let it rest for some time
before adding it to the mixture.
The pozzolanic reactivity could be increased by driving off all the water,
which is physically adsorbed to the inner surface, by heat treatment.

Oil is not a good release agent for geopolymer,
because it is hydrolyzed by the alkali.
Hard wax can withstand chemical attacks better.
Silicone oil has a good alkali resistance.
The best surface is plastic.

[Deleted]

You (the others too) are likely completely wrong with your assumptions about thermal conductivity
in granular materials.

After reading the paper linked below you may reconsider your mix design.

The Effect of Particle Size on Thermal Conduction in Granular Mixtures.
www.mdpi.com/1996-1944/8/7/3975

[Deleted]

Haha karl, I like your frankness !
thanks turtle and lawry for your answers. I'd like not to use PVC because I cook the geopolymer in the same oven I cook my food in. I'll give a try to bee wax as a release agent.

I thought using oil as a release agent was a good idea because of the hydrolyzation which leads to the formation of soap which should not stick to the mould. But maybe the soap is destroyed by the alkali ?

In the paper you linked, insulating particle smaller than the conductive particles are used. In this case, the smaller the insulating particle, the better the insulating effect (volumetric fractions being equal).

Here, the results of the paper don't apply because the insulating particules i'm adding in my last mix (sawdust that will become void after firing) are much bigger than the conductive particles of the geopolymer. I don't see how macro voids like this could favor the thermal conductivity.

regards,

[Deleted]

Macro voids allow convection streams and heat can travel fast by radiation.
The convection streams in restricted volumes can travel quite fast,
faster than you may belive,
thus raising the conductivity of the gas inside significantly.
From a mathematical point of view, from a pragmatic view as well,
voids can be seen as insulating particles.
Microporous materials can have a conductivity even lower than air,
as no convection is possible and radiation will be reflected many times.

Commercial mixtures use any kind of sugar or derivates thereof
and sugar like polymers eg. proteins as additives in geopolymers,
which may burn out and leave micropores.
BTW starch and cellulose are sugars too.
Some sugars eg. table sugar may have a retarding or accelerating effect depending on the amount.

If you do a patent search you may be surprised which substances are mixed into geopolymers to create pores or for other purposes.

By the wording of some patents even the use of sawdust infrings those.

Anyway, very fine or even liquid burnout additives require much less material
to achieve the same effect als larger ones, while weakening the structure less.

[Deleted]

Soap is water soluble contrary to oil and will simply be dissolved.

[Deleted]

Thanks for these infos karl !

I didn't imagine the insulating particles needed to be so small. In my mind, fine sawdust was fine enough to stop convection.

I often use an insulating mix made with 50% sawdust and 50% creamy clay to make insulating bricks or mortar. Next time i'll let the mix ferment as long as i can so that the sawdust is transformed into smaller particles. The insulating effect should be better.

[Deleted]

One mole of any substance contains 10^²³ molecules.
One mole of any gas occupies about 22 liter volume at 0°C.
If you devide the mole volume by a billion it contains still 10^¹⁴,
in words hundred thausend billion molecules.
Thus a gas volume in the size of a sawdust particle contains sill an extremely huge amount of molecules,
which allows to establisch enoug density difference to cause convection streams.

[firewatcher]

Jul 23, 2016 1:51:46 GMT -8 @yasintoda said:

Hi everyone,
I've been reading and experimenting geopolymerisation for some months now, with many failures and some success. My goal was to produce an insulating, heat-resistant castable in order to cast some cylindrical heat-risers. I'm posting some of the results, hoping they might be useful to some of you.

Pictures may be found here : webcloud.zaclys.com/index.php/s/7hMokc2rhq3iYqh

First trials :
My basic mix was the one that @karl gave in another post : 10 parts (in dry weight) of a fine alumino-silicate, 1 part of sodium hydroxyde powder, 1 part of natural hydraulic lime. Then I added vermiculite and/or fine sawdust to give it insulating properties.
I tried at ambient temperature :
    * natural pouzzolan with a granulometry between 0 and 1 mm : no hardening
    * red chamotte, granulometry 0-1 mm : hardened but still brittle
    * very fine disaggregated metamorphic rock : hardened but still brittle

Second row of trials :
I synthetised sodium silicate from 4.5 parts (in dry weight) of Sodium Hydroxyde, 5.5 parts of diatomaceous earth, 20 parts of water. I used it successfully as a deflocculant, but not in my insulating application.
I then tried at ambient temperature :
    * 1.5 L of vermiculite + 150 mL of sodium silicate, weak hardening
    * 1.5 L of vermiculite + 300 mL of sodium silicate, weak hardening

Third row of trials :
10 parts of 0-1mm red chamotte, 1 part of sodium hydroxyde powder, 1.5 part of water. I let them rest for 3h at ambient temperature. I then added 1 part of natural hydraulic lime and, after that, vermiculite hydrated with 15-20% water in volume. It allowed me to cast the combustion core of a 18 cm (7") rocket. The rocket has been in daily use for nearly three months. The geopolymer is quite resistant to abrasion and with the time the texture became like glass. The moulding has been done with wood that has been fired in the rocket, so no unmoulding was required.

In detail : 30 kg of 0-1 mm chamotte, 3 kg of sodium hydroxyde, 4.5 L of water, 3 kg of natural hydraulic lime, 30 L of vermiculite pre-hydrated with approx. 5 L of water. This allowed to cast a volume of 30 L.

Fourth row of trials :
12 parts of 0-1mm red chamotte, 1.5 part of sodium hydroxyde powder, 1.5 part of water. I let them rest for 3h at ambient temperature. I then added 1 part of natural hydraulic lime and, after that, vermiculite hydrated with 15-20% water in volume and hydrated sawdust.

In detail : 20 kg of 0-1 mm chamotte, 3 kg of sodium hydroxyde, 3 L of water, 2 kg of natural hydraulic lime, 20 L of pre-hydrated vermiculite, 20 L of pre-hydrated sawdust. Cooked at approx. 100°C for 48h, and then fired in place in the rocket.

Conclusion : very insulating, barely sticky when casted but holds together quite well after the first cooking. Not very resistant to abrasion, but it's not a problem for a heat riser. Been in daily use for two months, no sign of degradation at all.

Fifth row of trials :
10 parts of 0-0.2mm kaolin chamotte, 1 part of sodium hydroxyde powder, 1.5 part of water. I let them rest for 3h at ambient temperature. I then added 1 part of natural hydraulic lime and, after that, hydrated sawdust (1L of sawdust per kg of chamotte).
Casted in an oiled metallic mould, cooked at 100°C overnight and then fired at 250°C for a few hours in the previoulsy made, rocket-fired oven.

Results : unmoulding is impossible. The mix is quite sturdy but sticks too much to the mould. I tried to improve my mould by giving it clearance angles but it lead to another failure.

Some remarks on geopolymer :
* I've found that an alumino-silicate is fine enough for geopolymerisation when it makes a paste once hydrated. If it stays crunchy like sand the mix won't hold.
* I found that lime is not necessary. It accelerate the setting, but if you have enough time you can just cook the mix longer.
* It's better to hydrate the vermiculite/perlite/sawdust before adding it to the mix, so that it won't absorb directly all the water needed for the geopolymerisation.
* the insulating geopolymers i've tried are weak and friable until fired at high temperature (900 °C). When the ceramic bond is created the geopolymer becomes hard and glassy. At medium temperature (<250°C), the geopolymer gets harder but it's still hard to move without breaking it.
* Unmoulding is always very difficult. To me, that is the biggest problem with geopolymer in our applications. It's possible to make cynlindrical moulds with very fine sheet metal as long as the internal diameter of the heat riser is large enough (>180mm).
* It's a good idea to keep a wet cloth and some diluted vinegar nearby to remove spots. I've burned myself too many times and now i prepare it systematically. The acidity of the vinegar will stop the burn of the alkaline geopolymer. Use heavy PVC gloves to manipulate the sodium hydroxyde.
* 0-1mm chamotte costs 0.3 €/kg, 0-0.2 kaolin chamotte costs 0.8 €/kg, sodium hydroxyde powder costs 2 €/kg, hydraulic lime costs 0.4 €/kg, vermiculite/perlite costs 0.16 €/L.

regards,
yasin

Hi yasintoda,

You mentioned with your recipes that the amounts of material that you used was enough to cast a 7 inch stove...how thick were the walls? Sorry if I missed the informatiin in your post.
I'm trying to keep my options open on materials due to cost consideration. Thus far, the best I've been able to come up with is a mix which will cost about $80/cubic foot, buy this figure counts on a supplier delivering on a promise to get me some powder perlite inexpensively through a "specialized" delivery process which is probably unlikely to happen as it's been 3 weeks since last i called them and I haven't heard anything from them since then.

[Deleted]

Hi firewatcher,

have a look at the picture here webcloud.zaclys.com/index.php/s/7hMokc2rhq3iYqh#//IMG_4205.JPG

I don't remember exactly but I think the side walls were 3" thick and the front and back ones 4" thick.

Good luck with the materials,