Find all your DIY electronics in the MakerShed. 3D Printing, Kits, Arduino, Raspberry Pi, Books & more!

“Geopolymers” are artifical stone materials that are somewhat like cements, and have interesting applications in many of the same areas. There are, however, some important differences, both practical and theoretical, between a geopolymer, and, say, a mortar or concrete based on Portland cement. Practically speaking, geopolymers show impressive performance, in some tests, above and beyond those of run-of-the-mill cement and concrete mixes. On the theory side, geopolymers are critically different from cements because they don’t depend on the hydration of lime (CaO) to set up. Lime is made by strongly heating limestone to drive off carbon dioxide, a process which, given the huge amounts of lime required by the huge amounts of cement our world consumes, is a major contributor to atmospheric CO2 emissions. Hence, a lot of the excitement about geopolymers, apart from their potential high-performance applications, has to do with reducing our collective carbon footprint.

Anyway, big-picture takeaway: A chunk of cast geopolymer is a very different thing than a chunk of cast cement. If you’re a hands-on type, like me, your first question on hearing about this, or any kind of fancy material, is likely to be how do I get my hands on some? Data, theory, and popular science all have their uses, but ultimately that’s always the best way for me to understand something: Hold it in my hands, maybe poke it with a stick. Best of all, of course, is if can make it myself.

So I started digging for the practical protocols the lab folks were using to make little blocks of geopolymeric stone to run compression tests on. This kind of thing is always frustrating because a lot of the hands-on info is locked away behind academic publishing paywalls, and I have no intention of shelling out $39.95 to download one six page paper from ten years ago, thank you very much Mr. Elsevier. Anyway, perseverance paid, and I eventually found a 2008 Journal of Materials Science paper by Australian researchers J. Davis, et. al., posted (probably illegally) to scribd.

Table 2 is particularly instructive. It includes four formulae for geopolymer compositions, one of which (“SGP”) is both highest-performing, in terms of compressive strength, and simplest to make. Pound for pound, it’s also probably the most expensive, which would be a problem if I wanted to build a viaduct out of it. But all I want to do is cast a couple small pieces. So we’re good.

What follows are my notes on adapting the “SGP” protocol from this paper for garage prep. It is a straightforward formula, and most of the leg-work, as is so often the case in this kind of thing, is in sourcing materials in a way that does not require going through one of the big, expensive chemical supply houses that really doesn’t want to deal with citizen scientists. It’s important to note that I haven’t actually done this, yet, so I can’t vouch for its efficacy or safety. But I wanted to publish the reference and my thoughts about a DIY version before concrete month gets away from me. If you’re interested, read on, and please do comment if you spot anything fishy; if not, stay tuned for a follow-up about how it works out.

Starting Materials

Sodium hydroxide – AKA lye. Available at many hardware stores as drain cleaner, e.g. “Red Devil.” This is strong base and you need to understand how to handle it safely.

Metakaolin – This is a form of kaolin, a common clay, that has been chemically changed by heating at about 750° C for several hours. It is commonly prepared this way in the literature, but most people don’t have a temperature-controlled furnace in the garage for performing this operation. Fortunately, so-called “highly reactive metakaolin” is available commercially, online (in the US, anyway) for use in cement countertops. A 25-lb bag will be well more than we need. If this process works, I may divide up my leftovers for sale in small, cheap portions for those who want to play along at home. The MSDS suggests that it is not particularly hazardous, but as with all fine powders, a dust mask is probably a good idea.

Sodium silicate solution – It is probably possible to prepare a homemade solution of sodium silicate that will work in the geopolymer process by using a procedure like this one from NurdRage, in which finely-crushed silica gel desiccant is dissolved in strong lye solution. In the literature, however, geopolymer samples are seemingly always prepared from a ready-made commercial sodium silicate solution in water. Unfortunately, the “grade O” commercial sodium silicate solution specified is only available from specialty suppliers; however, I think I can fudge it by adjusting the composition of the common “grade N” solution used, e.g. to repair mufflers, by adding lye flakes. So we’ll need some “grade N” sodium silicate, to start. “Grade N” is also called “Grade 40″ and “water glass,” and Googling turns up several online sources.

Step 1: Prepare sodium silicate solution

The easiest way to measure ingredients for this process is by weight. You’ll need a scale with a capacity of at least 1000g. Put a 250 mL beaker on the scale, record its weight, and, taking appropriate precautions for handling strong base, add 4.4 grams of lye flakes. Now add an additional 62 grams “grade N” sodium silicate solution. Remove from scale and stir to dissolve lye. Once this solution is well mixed, cover it and let it stand, at room temperature, for 24 hours before use. Note that this solution is now slightly more dilute than commercial “grade O” sodium silicate, but the Davis SGP recipe actually calls for diluting the commercial solution just a bit, anyway, and the math works out pretty closely.

Step 2: Add metakaolin

Remove the cover, put the beaker back on the scale, and bring the the total weight of the solution up to 100g by adding dry metakaolin powder. You may want to add it in batches, with stirring in between. The metakaolin will not dissolve; you should end up with a paste or slurry. The polymerization reaction will begin as soon as you start adding the metakaolin, but you should have at least an hour’s pot life. If it’s working, the mixture should begin to give off heat.

Step 3: Casting

Transfer the mixture to a small metal mold. I will probably use a steel muffin tin—they’re cheap, about the right size, and if I want to I’ll be able to cast and cure multiple samples in a single pan. Seal it well with aluminum foil. The idea is to keep water from getting out during the curing.

Step 4: Curing

The “SGP” mixture should polymerize at room temperature, but geopolymer samples are generally cured with mild heat, and if you want to experiment with other aluminosilicates besides and/or in addition to metakaolin (e.g. fly ash), the curing step seems to be necessary. Generally, the protocols call for heating to 60° C for 24 hours. That’s 140° F, and I’m comfortable doing that in my kitchen oven, as long as I’m not going to be leaving the house during that time.

Final thoughts

Demolding the cast samples may be a problem that eventually calls for some kind of mold release. But it’s not reported in the literature protocols I’ve seen, so I’m going to start without it. I’ll probably want to experiment with adding aggregates, and sand is convenient and conventional. The Davis paper describes mixtures with 40 wt% and 60 wt% sand additions to the “SGP” formula.

More:
Super cements aka “geopolymers”

Sean Michael Ragan

I am descended from 5,000 generations of tool-using primates. Also, I went to college and stuff. I write for MAKE, serve as Technical Editor for MAKE magazine, and develop original DIY content for Make: Projects.


Related

Comments

  1. elt2jv says:

    Excuse me, but it may be ill advised to put a lye-containing mixture into a metal muffin pan. Even less advisable is to cover it with aluminum foil. I might suggest a thermally stable glass or ceramic.

    If you have no choice but to use a metal container, then consider coating it with a non reactive product like wax.

    1. Sean Ragan says:

      That’s a good comment, thank you. The Davis paper specifies a “metal mold,” but I need to think about that interaction, and I haven’t yet.

    2. James Vooght says:

      I also agree about the low pH and aluminum reaction. Try a silicone muffin pan. Should be non-reactive and you can easily peel it off after the product sets. Try a silicone cutting sheet as a cover. it will certainly withstand the temps you quoted. If the reaction is exothermic don’t use your home oven. Setting it in the sun outside on a warm day may do it.

      1. Sean Ragan says:

        Thanks, silicone is a good idea. I’m less concerned about the exotherm at 140F in my oven. I’ll be surprised if this reaction, even with pure metakaolin, is energetic enough to cause any problems. But I’ll try your suggestion and see how the product compares. Course, I’ll have to think about a testing protocol…

      2. I’m afraid this mixture is *high* pH, not low (high is basic, low is acidic. NaOH is a strong base, silica is a weak acid and thus isn’t going to change that a lot.)

      3. Sean Ragan says:

        FrozenFire, thanks for keeping us honest and pointing this out. It is critical not to forget that basic solutions have numerically *high* pH, while solutions of “low pH” are in fact acidic.

        For whatever reasons, I think it’s interesting to note, this seems to be a common slip of the tongue/pen, even among folks I think I can safely describe as “experienced chemists,” by which I mean the folks in the department when I was in grad school for Ochem at UT. It was common enough that it became a kind of running joke.

    3. Elliot Hallmark says:

      I use coffee cups for those heavy paper, waxed coffee cups for this. No release compound necessary.

      1. Elliot Hallmark says:

        edit: I meant I use heavy paper, waxed coffee cups as the mold I cast in.

  2. Sean thanks so much for doing the legwork on this – I’ve been mulling over a similar material using a silica/clay gel mixed with cellulosic filler — maybe pine needles or straw — to make a castable mid-temperature insulator in rocket stoves, etc. Setting aside the addition of organic material like straw, what do you imagine the thermal performance properties of your blocks will be?

    Also, do you have any idea as to the water-resistance of this material? When you make casting cores using waterglass and CO2, the resulting material is very hard, but loses its integrity with the addition of water.

    1. Sean Ragan says:

      Dominic, thanks for your comment. Here’s an open paper…

      http://www.azom.com/article.aspx?ArticleID=3171

      …concluding that a very similar composition (using potassium salts instead of sodium salts, but otherwise identical) is a castable refractory suitable for continuous use at 1000 C.

      That water glass / CO2 process for fixing sand cores is cool, but the chemistry of a geopolymer is very different. The silicate sand core process essentially glues the sand matrix together with crystals of sodium carbonate (washing soda), which are held together by the electrostatic attraction between sodium (+) and carbonate (-) ions, and remain highly water soluble — that’s why the cores come apart when you wet them. A geopolymer involves formation of very long molecules in which the atoms are held together by covalent bonds, analogously to the polymers in, say, a plastic fork, except geopolymers are not long chains of carbon atoms–they’re long chains of silicon and aluminum and oxygen. They are thoroughly and completely waterproof, and in fact are generally more resistant to all forms of chemical attack (especially strong acids) than Portland cement-based materials.

  3. Guitarsavvy says:

    Isn’t it counter productive (CO2-wise) to use a product that requires “… heating at about 750° C for several hours.”

    If all this hassle is over CO2 and Global Warming, consider that a single volcano can out CO2 decades of human activity. Hate to see

    And (as not reported in the mainstream media) there are growing glaciers…
    tinyurl.com/inconvenient-fact

    Wishing all the best to you!

    1. Sean Ragan says:

      Good question, and here’s why it’s different: Making lime to make cement requires all the CO2 from (presumably) the fuel burned to heat the furnaces that heat the limestone to around 850 C. ADDITIONALLY, that process itself releases CO2, in great quantities, that was formally trapped in the limestone. So you’re making a bunch of CO2 to heat limestone, which produces a bunch of additional CO2 in the process of changing to lime.

      The heating step in the preparation of metakaolin does require the energy input which could be adding CO2 from power plants, etc. Critically, however, the chemical transformation of kaolin to metakaolin *does not give off CO2.*

      In any case, metakaolin based geopolymers are an ideal lab system but, for a number of reasons including the energy requirements of the heating step, are considered too expensive to be a viable cement replacement. Much of the ongoing research into geopolymers is about producing them without the need for metakaolin using aluminosilicate industrial byproducts such as flyash and/or foundry slag.

  4. I’d like to know what kind of textures you can get with this process and the weight properties.

  5. Sean, In the picture I see 4 class of results. What are the differences in the preparation.
    There is other data than called my attention, in the Step 2, you say “up to 100gr” . Shouldn’t it be 1000gr?
    Thank you for the article

    1. Sean Ragan says:

      Those of you with a scientific background who are coming to this article may be confused or irritated by the form of its presentation. To be clear, it probably should have been titled “Notes *toward* a Garage Geopolymer Prep.” Again, it’s important to note that I haven’t actually done this, yet, and this procedure is just a plan for how I think one respected literature protocol could be adapted for home experimenters. The photograph is related to the content of the article only in that it shows samples of freshly cast geopolymers prepared by someone else, possible using other method. It is here only because blog posts need pictures. Apologies for any confusion this may have caused.

      And no, “100gr” is correct in step 2. This method will prepare a very small sample, my guess is single a cube about 4cm on a side.

      1. patriciocardo says:

        It is also very usefull for me. If you remember the origin of the image, I’d appreciate you publish, thank you!