You can get neutral waterglass to foam up into an open-cell-foam glass just by heating it to its softening point (I didn’t measure with a thermometer, maybe 500°?) over the course of a few minutes. The structure of the resulting foam is fairly coarse, and the density is not that low.
A couple of possibilities occur to me to improve the structure.
Water is probably the best blowing agent for this kind of thing, just based on its low molecular weight. Ammonia might be a possibility, too.
First, waterglass solution doesn’t solidify at room temperature until it’s something like 35% waterglass by mass, and so when you heat it up slowly, any water in excess of that 65% (or even a bit more, since solubility increases at higher temperatures) will bubble out in the liquid state, before the solid foam can form. Adding small amounts of polyvalent cations (such as calcium or magnesium, or maybe even aluminum or boron) is well known to decrease its solubility to by orders of magnitude, so perhaps they could permit the formation of a solid hydrogel with much higher water content, perhaps 90%, 99%, or 99.9%, with a corresponding decrease in the density of the final product. There must be some density limit below which you get glass mist rather than glass foam, but given how viscous glasses can be just above their softening point, the limit might be pretty low.
Second, if the expansion is carried out very quickly rather than over a period of time, bubbles will have less time to coalesce into larger bubbles and convert the foam into the open-cell form. So perhaps the solid hydrogel precursor could be foamed more effectively in the way that rice and breakfast cereals were originally puffed (and still are by roadside vendors in China and Korea), by first heating it under pressure to its softening point, then suddenly releasing the pressure. For most cereals this is done by heating the cereal to the proper temperature inside a cereal cannon, then whacking the valve on the front of the cannon with a hammer, allowing the cereal to blow the cannon door open; but for popcorn it is instead done by encasing a small ball of the starch hydrogel in a hermetic hull which contains the pressure until it ruptures. Both possibilities might work to produce lighter-weight glass foam.
In the cannon case, there’s the risk that the glass hydrogel will completely melt inside the pressure chamber and stick together into a single mass, and also to its walls. If the hydrogel is initially produced in granules, their surfaces can be treated to prevent this, for example by rolling them in quartz flour (which can be made to stick either by doing it before they have finished solidifying, by making the quartz flour hot enough to partially melt the surface, or by a hot air blast) or by replacing most of the alkali ions near their surface with polyvalent cations like those mentioned above.
Many minerals might serve as viable alternatives to quartz flour here, including chalk, quicklime, alabaster, magnesia, zeolites, clays, talc, feldspar, aluminum hydroxide, mica, or rutile — almost any mineral used in formulating ceramic clay bodies or glazes will have a much higher melting point than the waterglass hydrogel, except perhaps oxides of lead. Boric acid and borax don’t have a higher melting point, but still might be an alternative, by forming a borosilicate network in the surface of the granule.
Most of these treatments to increase the melting point of the granules’ surfaces would also increase the surface’s strength, which would permit the use of the cannon-free popcorn process.
An alternative means of puffing rice is “hot salt frying”, in which salt is heated to a high enough temperature to puff rice (but not melt the salt), and then the rice (often parboiled) is mixed into it. The hot salt transfers heat to the rice much more rapidly than hot air would at ordinary velocities, so the rice puffs instead of dehydrating through vapor diffusion as it normally would. The coarse-grained puffed rice is then easily sieved out of the fine-grained salt. Modern continuous-process cereal puffing works the same way, but using high-speed air or steam rather than solids. This might be a viable alternative way to rapidly foam waterglass hydrogels, but media you wouldn’t want in your food could be used instead of the salt — anything from the foregoing litany of quartz, chalk, quicklime, etc., and also carbon. These would permit the use of higher puffing temperatures. Using air, of course, would permit puffing glass at higher temperatures still, but its low thermal conductivity and thermal density means you need high-pressure air jets.
You could load a refractory mold full of many such beads of the hydrogel, but without any such surface treatment, then run hot air or steam through the mold cavity at high speed in order to expand all the grains and cause them to fuse together, like expanded polystyrene. This could be a pretty quick process if the grains are small; the hot air could convert them into a sort of fluidized bed.
I placed a thin irregular slip of air-dried waterglass on a bed of expanded vermiculite, then dumped a small bowlful of preheated construction sand on it. The bowl and sand were not hot enough to glow visibly, but hot enough to immediately char the towel I was using to hold the bowl and the towel on the floor that I spilled some of the hot sand on. The expanded waterglass foam retrieved from beneath the sand was about 47 mm long at its longest, 27 mm wide at its widest, 20 mm wide at a narrower point, and 10-12 mm thick over most of its area. It weighed some 600 mg. If we approximate the volume as 45 mm × 22 mm × 10 mm, we get 9.9 cubic centimeters, so 0.06 g/cc, suggesting that it is about 98% air. Upon immersing it in a glass of water weighing 421.4 g while supporting it from above with chopsticks, the measurement increased to 430.4 g, suggesting it was displacing 9.0 mℓ of water, and upon weighing afterwards, its weight had increased to 1.5 g, suggesting that it had absorbed 900 mg of water, also giving a volume of 9.9 cubic centimeters. Although an extra decimal place of accuracy would give more confidence, particularly given how the scale readings were drifting upwards while the foam was held underwater, it seems safe to say that it’s probably 60 ± 30 mg/cc, so maybe 97%-99% air. (But contrast that 0.06 g/cc to the 0.74 g/cc reported for pumice in Material observations.)
The resulting piece of white foam can be easily handled without breaking it, but it’s fragile enough that dropping it on the table sometimes breaks off a corner. Contact and rubbing make sounds similar to the sounds of extruded polystyrene, suggesting that the velocity of sound through the material is similar (and thus a similar density to stiffness ratio) and that acoustic coupling to air is pretty good, as you’d expect it to be if its density is only 60× higher than air’s.
For context, a silica aerogel produced in the 01990s was the record-holder for least-dense solid material for a while at 3 mg/cc, more recent aerogels have reached 1 mg/cc, and extruded polystyrene panels for building are typically 20–80 mg/cc.
I was concerned that the sand or vermiculite might stick to the waterglass as it foamed up, but this didn’t seem to happen — there were three pieces of vermiculite stuck to it, but I think they were there in the unexpanded waterglass, which had oozed onto a polyethylene sheet out of some vermiculite I was trying to glue together earlier.
Later, I reheated the foam to drive out the water and verify that it still weighed 600 mg. After removing it from the hot bowl on the stove, I could hear crackling inside the foam as the thermal shock induced fractures inside the foam.
Fragments of this foam melted and collapsed when heated with a butane torch to red or orange heat (say, 700–900°). In an effort to raise this softening point, I tried boiling another foam fragment made in a similar way in aqueous magnesium chloride for a while, then boiling it in tap water for a while to get rid of the magnesium chloride. The idea was that the magnesium, without changing the structure of the foam, would replace some or most of the sodium to produce some kind of mostly magnesium silicate, which would be water-insoluble and have a much higher melting point. After a first such tap-water boiling, there was some white deposit around the water (probably MgCl₂) but the foam no longer tasted noticeably like magnesium chloride. Just in case, I boiled it in fresh tap water for a while longer, which left less of a white deposit around the water.
Wet, the fragment sank in bottom of water and weighed 400mg. Dried, it weighed “0”: under the 100mg low end of my crappy scale. After heating it with the blowtorch to an orange-yellow heat (800°–1000°?) for several minutes, which produced some sodium yellow in the flame at first, it appeared slightly smaller and less white, more translucent. Upon putting it in water, rather than floating at first as before, it immediately sank, and its wet weight was still "0", so, under 100mg. This suggests that its pore space had diminished by at least a factor of 4 from this treatment, and that the pores had become larger.
This survival for several minutes contrasts strongly with the behavior of the freshly prepared foam, which shrivels up to a tiny bead in seconds upon being heated in the same way.
I’m repeating the MgCl₂ procedure with another piece from the 9.9mℓ chunk of foam just to make sure I’m not fooling myself. I kind of fucked it up because I let it boil dry, and then didn’t let it cool back down before adding water, so water was boiling fiercely as it re-wet the foam, which probably did some real damage to its structural integrity. Also, the first time I did it in the cut-off bottom of an aluminum can; this time I’m doing it in a steel bowl with a badly oxidized plating of something like nickel, so there may be different compounds running around. And the water is cloudy white (perhaps due to particles of insoluble silicate broken off when I re-added the water) rather than transparent as before.
The second magnesium-infused chunk of foam did indeed withstand the blowtorch flame; but I think it lost significant weight during the magnesium-infusion process, and it looks significantly smaller. At any rate, once dry, it weighed in at “0 g” again. Wet, it weighs 1.1 g.
To test it further, I put it back in the bowl of vermiculite and built an aluminum-foil arch over the top of it, then heated it to yellow heat (parts to white heat) for several more minutes with the butane torch. It appeared unharmed, but after cooling was evidently more fragile than before, and its wet weight was 1.0 g, so it may have lost 10% or 20% of its pore space through this treatment. Also, the aluminum part of my butane torch nozzle melted and the brass tip fell out on the floor, so I had to stop.
To test compressive strength, I cut a 13 mm × 18 mm rectangular piece of the non-magnesium-treated foam and sanded two surfaces fairly flat, making it about 6 mm thick. I placed it on a larger scale and crushed it by pressing down slowly on a slab of granite placed atop it with my hand; it began to crush around 1 kg and completed crushing around 2 kg. 1 kg gravity / 13 mm / 18 mm is about 40 kPa, so the compressive strength of the foam is probably somewhere in the neighborhood of 20–80 kPa. This is an order of magnitude lower than construction-insulation styrofoams, which are typically in the 150–700 kPa range, measured by the DIN 53421 standard, which evidently specifies 10% deflection as the limit.
(See also Synthesizing amorphous magnesium silicate.)
The glass foam resulting from foaming waterglass can be abraded or sawn (or crushed) very easily, and the possibility of cutting it with a hot wire, like styrofoam, is very appealing, especially if its density can be decreased further. Although it’s quite weak, it could be very useful for supporting granular slightly-denser materials such as perlite, vermiculite, alumina foam, and cheap carbon foam, for example while an adhesive sets (see Leaf vein roof for one use for this). Because of its relatively low melting point and very low density, it might be possible to “burn it out” in such cases, leaving only a thin layer of residue.
It was easy to bore a hole through one of the scraps prepared in the crude kitchen experiment described above using an 850-μm spring-steel wire, just by poking at the foam with the end of the wire. The material did not visibly shatter or chip as the wire poked through the back side.
A different way of using this glass foam to support stronger materials is to first get it into the right shape (whether by expanding beads into a mold, gluing together a bunch of pre-expanded beads and pushing them into a mold, by cutting or abrading at low temperatures, or by hot-wire cutting) and then use the resulting form as either a mold or a stucco substrate. This is more or less the same way styrofoam is used for molding, for example, concrete, or as a base for a fiberglass layup. By painting, spraying, wrapping, laying up, or otherwise depositing a stronger material onto its surface, you can make a strong, hard shell. This may need to be done in stages, first building up a lightweight shell that can be supported by the foam, then a stronger shell supported by the lightweight shell, then perhaps a solid object filling the whole shell.
Why would you use glass foam for this rather than styrofoam? Well, it doesn’t require any organic materials, so it’s potentially much cheaper. Being more rigid means you can cut it to precise dimensions more easily, and it will bend less when you’re doing things to it that impose slight side loads, like painting or stuccoing it. It can withstand common solvents without any complaint, unlike styrofoam, and it can withstand higher temperatures than any organic polymer. Counterintuitively, it might be possible to get it to a lower density than styrofoam, because silicate glass (even waterglass) has a much higher strength-to-weight ratio than polystyrene; however, experiment so far has not realized this possibility.
Also, if the density gets low enough, you can use glass foam for lost-foam casting, particularly for casting of materials like basalt, fused quartz, lead glass, or soda-lime glass that will happily dissolve the glass-foam residues.
It might be possible to raise the melting point of the foam once it has been shaped through ion exchange, for example with salts of magnesium, iron, boron, or aluminum. Saturating the foam with an aqueous solution of such soluble salts might be enough. This could enable it to be used directly as a refractory.
The standard established approach to foaming glass is to mix carbon and an oxygen source such as MnO₂ into it, so that as the glass starts to soften they react to form CO₂ (and, say, Mn).