Techniques for working with geopolymers

January 24th, 2011

Pouring:

Geopolymer resin is very fluid and can be poured onto and into things, and allowed to set. Bubbles in the resin are a huge problem. Also, besides pouring flat sheets and some geopolymache, I not made much this way. One could make bricks by this technique but using more aggregate, either plastic clay or non-plastic minerals, and a different method would be more cost effective.

Thixotropic casting:

A (Na/K)-PSS resin loaded heavily with sand (preferably of mixed grades) will hold its own weight when still, but will flow when vibrated. Perhaps (Na/K)-PS also has this property, I have not tried. This is called thixotrophy, and this working state is highly non-plastic. Perhaps it is technically the opposite of plasticity.

Without consistent measuring of fine particles and aggregates, I occasionally would obtain a thixotropic mixture accidentally. It cracks easily, but if vibrated into a mould the resulting piece is good. My failures here have been in trying to bend the mixture into shape, or to slump it. The non-plastic body will crack terribly if bent, and if it cures under tension (as in slumping) I have seen it crack all over. Ideally, one would vibrate it into a mould where it is fully supported all around. Compressional pressure, I expect, would be helpful, though I have not yet tried that.

Egyptian faience was made by this method, using a sodium silicate based cement.

Plastic clay:

I have mixed Na-PSS and Na-PS resins into ball clay to try to make a workable plastic clay body. This takes a lot of effort to make a small batch. Because Na-PSS will set in less than a day, small batches are all one would want to do. In an industrial setting, where a pug mill was whipping the stuff out, it may be workable, but not on my scale. Also, it is unpleasant getting alkali clay all over your hands for long periods of time. I used latex gloves as I was getting burned in those days from other alkali handling tasks also. I did make a pinch pot ash tray that was strong enough. But especially my Na-PS attempts were insufficient. The ball clay subtracts from the final strength of the object, unless it later be fired to the point of vitrifying the un-reacted clay.

From reading I learned that iron bearing clay must be fired above 250C to make the iron inert, otherwise it will prevent geopolymerization.

If I were to come back to this, I would experiment with Na-PS based drying bats (plaster of paris wont work, see section on slip casting) for Na-PSS clay bodies. Then I could mix everything together while very wet and rapidly dry it. Up till now I have tried to mix it in a thick state, which is probably a cause of failure. Alternatively, a Na-PS clay body may be workable until cured at elevated temperature. I do not know what the working time of an Na-PS resin is because I cannot consistently get it to cure well.

Slip Casting:

The make up of slip casting clay is mostly kaolin clay and mineral fillers. When I learned this I was excited to try casting with Na-PSS. However, I found that Na-PSS kills (reduces to zero) the absorbent powers of plaster of paris (calcium sulphate) which make slip casting possible. I assume this is a reaction with the calcium at the interface between the resin and the mould.

Na-PS pieces, which are absorbent like plaster of paris, were able to collect a layer of dewatered Na-PSS resin when in contact with it. Thus, that hurdle of slip casting may be over come. I cannot say anything about the service life of this mould, because its pore space may be filling up with precondensate that later polymerizes. The Na-PSS layer was gooey, and probably would not be strong enough to come out of a mould unless already set, and in that case it would probably be bonded to the mould. And it would be tricky to get a dried piece of resin to cure properly, perhaps impossible.

In his book Zeolite Molecular Sieves, D.W. Breck briefly mentions two methods of obtaining slip cast zeolite A. In the first, a piece is slip cast conventionaly and then fired at 750C for 6-12 hours, making it solid MK750. In the second, MK750 is included along with raw kaolin in the slip casting mix. Then, in both methods, the piece is digested in a caustic solution at an elevated temperature. The author does not say whether the results were confirmed as zeolite A or hydro-sodalite. These demonstrated geopolymerization in 1964. I expect that these vessels would have the interesting property of being able to distil alcohol without heat. Water would be able to migrate through the walls and evaporate on the outside, but ethanol would be confined to the inside of the container.

Also, perhaps slip casting alkali-leached kaolin and then hydro-thermally digesting it would work. Have not tried this yet.
ss2 is unstable, and will precipitate a caustic glass if it sits for several days. Heating in the microwave and stirring/swirling will redissolve it.

To 100g ss1 I add 50g Mk750, let sit for 1-2 hours, fill with 40-100g silica flour and preferably sand of mixed grades. Cures at room temperature but 50-80 degrees celcius for 1.5 hours is preferable. To 100g ss2 I add 107g Mk750 and 25g water, sit, fill, use. This cures better at room temperature than the ss1 based GP. Ideally, these would rest for 14-28 days covered in plastic to keep them from drying out (like portland cement). Then they can be dried, and as they dry can be exposed to higher and higher temperatures. Too high a temperature too quickly and they will fail from internal steam pressure.

With a minimal ammount of filler the resin is very fluid. (The ammount of filler I give above is general and I have not tested that 40g is the ideal minimal ammount.) One can keep adding silica flour and the resin becomes more putty like. It stops being wet and will cure on things but not stuck to them. If instead of adding large ammounts of very fine particles, one starts adding mixed grade sand the mixture becomes thixotropic (holds its own weight but flows when vibrated). I will discuss this further in the section on working techniques.

As the piece cures it is very fragile. While fluid it can be moved, and once solid it can take weight, but the slightest disturbance anytime in between these two points and the piece will crack in a dozen different places. Another symptom of this fragility is that the thixotropic mixture mentioned above will break all over if it cures while in tension, ie if it is draped over something or is hanging.

Sodium geopolymers will “bloom” after a couple weeks while drying. I understand that this is unreacted sodium that migrates to the surface, probably as sodium carbonate. The bloom can be washed off easily, and may come back once but is not very persistant in my experience. Perhaps a perfect specimin with no unreacted alkali will not do this. Potassium geopolymers possibly do not bloom because they are not molecular sieves (the sodium has no channles to migrate through). I have only losely gotten this impression from reading and have not compared sodium and potassium in this respect).

In his groundbreaking book on hydraulic mortors, Le Châtelier discusses substances that are slaked by humid air but not by water. Over-burned lime will air slake over time and destroy what looked like a good concrete. Le Châtelier sites sodium sulfate as an excellent example of a substance which seems to set fine with water but then will blossom in humid air. I have seen this impressively in sodium chloride and in sodium carbonate mud mixtures. In geopolymers I believe the free alkali reactes with carbon dioxide from the air, and then expands and grows as humidity allows it to migrate. The change in volume causes internal pressure which weakens or destroys the piece. This is just my thinking.

Another obstacle to having a good final product is the ammount of air bubbles trapped in the thick resin. Many of my pieces seem to be over 50% air by volume, which creates unattractive surfaces and structural weakness. No ammount of resting will de-gas the resin, nor will careful mixing. The huge and large bubbles can be vibrated out by vibrating the stirring rod. Apparently, the resin must be placed in a vacuum and the lack of atmospheric pressure lets the bubble rise.

Calcium Geopolymers:

A mixture of quicklime and commercial sodium silicate sets almost instantly. The reason for this is the precipitation of calcium (poly?) silicate. Free sodium hydroxide is also formed. A mature sodium geopolymer resin cosists of sodium silicate, sodium aluminate and sodium poly-aluminosilicates. Therefore, quicklime would also cause a setting of this resin through the formation of calcium silicates, calcium aluminates and calcium poly-aluminosilicates.

According the Breck (”Zeolite Molecular Sieves”), in the calcium-alumina-silica system: “although hydrous aluminosilicate gels have been employed as reactants, significant crystalization of a zeolite phase does not occur below 225C. Most successful experiments were conducted in the temperature range of 300C-400C” at autogenous pressure in an autoclave. (Breck also reports that no Magnesium zeolite phases have been synthesized under any conditions, though he does not discuss the sodium-magnesium system). In my own experience, lime and Mk750 alone was not significantly stronger than lime with an inert aggregate.

The addition of sodium however allows for ion exchange and precipitation routes which do result in zeolite phases. This has been my experience, without quantitative data. In my understanding, to a stoiciometric Na-PS or Na-PSS precondensate resin, one would add calcium hydroxide and enough Mk750 to absorb the free lye released by the ion exchange of calcium for sodium. The weight ratio of Ca(OH)2:NaOH:(MK750 or activated kaolin) would be 74:80:(444 or 516).

Rather than trying the above ratio, I made a Na-PS paste and mixed it 10% by volume into a lime paste. The mixture set very quickly, within 15 minutes, and though I had mixed it as I would mix an hydraulic portland cement mixture, the Ca-geopolymer quickly sequestered the free water and was thirsty for more. This is the behavior of an hydraulic cement, which neither lime nor Na-PS are on their own. The resultant block was much harder (more difficult to scratch) than lime mortar, and set when Na-PS alone would not. It was decently strong, requiring solid effort to break off marble sized pimples.

This water starved specimin was the only Ca Geopolymer I made, being more interested in the Na-PSS. I have not experimented with optimizing this system and I do not know if my thinking behind the proportions or mechanisms is accurate. This should have been clear evidence for me to be more careful in using wollastonite and silica flour interchangeably, though I did not heed it.