I’ve been thinking more about chemical reactions that can produce inorganic solids that are hard and/or strong and/or refractory from conveniently shapable precursors. Previously I was mostly thinking of things that react in aqueous solution, but now I’ve also been thinking about granular substances you could mold into shape (in some kind of carrier, such as a thermoset resin) and then heat up to activate the reaction.
In Dercuano in file berlinite-gel I mentioned Grover et al.’s
low-temperature synthesis of alumina ceramics bonded by berlinite
(aluminum orthophosphate) by partly dissolving alumina in aqueous 50
wt% phosphoric acid at 130° (with a 5:1 weight ratio of alumina to
phosphoric acid) for 1–4 days, then heating the resulting a thick
puttylike gel of hydrated aluminum phosphate to only 150° for 1–3 days
to get a 20%-porosity solid of nearly 50 MPa compressive strength.
Similar substances (“MALP” or “MAP”) are widely sold as castable
refractories. (See Cola flavor for more on this substance.)
With respect to adding phosphate cross-linkers, the ammonium phosphates should not be overlooked; some are deliquescent under ordinary conditions, easily dehydrated to crystallinity with gentle heating, and yield phosphoric acid when the ammonia is driven off with stronger heating.
An interesting benefit of phosphoric acid as a binder (or acid phosphates like monoaluminum phosphate or monomagnesium phosphate) is that it can bind silicates like soda-lime glass, but doesn’t destroy them the way lye does. Soda-lime glass fiber is only about US$3/kg, according to Potential local sources and prices of refractory materials, which is enormously cheaper than any alternative fiber reinforcement of similar strength. Maybe by combining phosphoric acid, some kind of high-strength foam (see Glass foam) and oriented glass fiber, you could produce a structure analogous to wood. Space Shuttle TPS protective LI-900 HRSI tiles were a felt of fused-quartz fibers, but they were adhered together with, I think, borosilicate.
More discussion of such composite materials is in Fiberglass CMCs?.
A different way to get phosphates of aluminum might be to react sodium aluminate, easily prepared by digesting aluminum in lye, with a soluble phosphate, such as one of the phosphates of sodium or ammonium. I seem to have had some success with this with fire; see Material observations, 02021-08-06. WP says sodium aluminate is thus used to precipitate soluble phosphates in water treatment plants.
Sources of aluminum may be other useful bonding agents; water-soluble sodium aluminate precipitates aluminum hydroxide in water upon cooling, part of the Bayer process; even the insoluble hydroxide itself may serve as a useful bonding agent due to its tendency to form a metastable hydrogel which eventually crystallizes. Calcining the hydroxide yields sapphire.
Aluminates of calcium are also potentially interesting, especially for 3-D printing; the highly reactive tricalcium aluminate is notorious for causing “flash setting” of concrete immediately upon hydration, though a small amount of sulfate can prevent this. WP says:
Water reacts instantly with tricalcium aluminate. Hydration likely begins already during grinding of cement clinker due to residual humidity and dehydration of gypsum additives. Initial contact with water causes protonation of single bonded oxygen atoms on aluminate rings and leads to the formation of calcium hydroxide.4 The next steps in the sequence of the hydration reaction involve the generated hydroxide ions as strong nucleophiles, which fully hydrolyze the ring structure in combination with water.
But of course if you want to 3-D print a stone object by spraying water onto a powder bed, this kind of instant setting may be highly desirable, and tricalcium aluminate's other deleterious effects on concrete may not be relevant.
My previous thought of precipitating calcium phosphate through a double metathesis reaction still might work as a cement, even with melting rather than in a solvent system. It’s similarly fairly instant. I had settled on calcium chloride and agricultural diammonium phosphate, and the thought at the time was an aqueous reaction; but DAP decomposes to yield liquid phosphoric acid (and other crap) at only 155°, which I think can attack the solid calcium chloride (which melts at 772°). Maybe it can also attack solid metallic aluminum to get phosphates of aluminum. Monoammonium phosphate is cleaner, decomposing to yield only ammonia and phosphoric acid at I think a slightly higher temperature.
Reportedly you can react alumina with boria at 800° or 600°–800° or 1200° and get the exotic aluminum borate, which reportedly sublimes at 1050° or above 1300°; reacting aqueous borax with aluminum sulfate below 45° reportedly gives aluminum borate directly. For Al₁₈B₄O₃₃, Vickers hardness is reported as 6 GPa, Young’s modulus 400 GPa, tensile strength 8 GPa (65% higher than basalt fiber, 15% higher than carbon fiber, three times maraging steel), Mohs hardness 7, crystal habit acicular, and the whiskers have been used to strengthen aluminum. There are a few different aluminum borates or boron aluminates, though.