I finally started trying out the recipe I’d come up with in Dercuano for a kind of instant 3-D printing cement based on the precipitation of water-soluble phosphate by pretty much any polyvalent cation. (See “Likely-feasible non-flux-deposition powder-bed 3-D printing processes”.)
So I bought 2 kg of calcium chloride desiccant (AR$778, US$5.72, US$2.86/kg) and 2 kg of diammonium phosphate fertilizer (AR$850, but AR$470 of that was delivery; the marginal cost of the fertilizer is AR$190/kg, US$1.40/kg).
The first observation is that this fertilizer is not pure diammonium phosphate. The individual prills have substantial variation in color, and they do not dissolve fully in water, even at boiling. A slight ammonia smell evolves on boiling the water, and is absent from the bags of fertilizer. Filtering the liquid through a coffee filter produces a transparent brown syrupy liquid, leaving most or all of the solids behind (I’m doing this in cut-up aluminum cans, which are not as good as glassware for seeing small amounts of cloudiness).
This phosphate liquid fails to dry even upon being sealed in a room-temperature drying chamber sharing air with the calcium chloride for several days. (The chamber is Saran Wrap over the top of a cut-off can, so it may be leaky, but I don’t see any deliquescence on the calcium chloride, so it’s at least not very leaky.)
The phosphate liquid instantly produces a thick white suspension of a fine precipitate upon being poured into a solution of the calcium chloride. Presumably this is some kind of calcium phosphate, along with whatever fluoride may have been present as a contaminant.
After filtering through another coffee filter, it has the mouthfeel of pure clay, making my teeth slide against one another with quite a bit of difficulty, but no grittiness, demonstrating that there are no crystals above the micron scale. The taste is also slightly bitter and salty, so I probably didn’t wash the filtrate enough. To the touch of the hand, the suspension resembles a thin kaolin slip. It dries on the skin to a powder resembling rock-climbing chalk.
The calcium chloride seems relatively pure: it is plain white, dissolves completely in water, and its only smell is a faint whiff of quicklime. Left in open air for a few days with a drop or two of water, it gradually begins to deliquesce, producing a liquid that feels “oily” because it’s not evaporating. Nevertheless, it is not a food-grade chemical either, and it’s labeled “for industrial use only”, so I shouldn’t have tasted it.
I also added some of the phosphate solution to an aqueous solution of some magnesium chloride I had lying around, which also produced an immediate precipitate of, presumably, trimagnesium phosphate, dimagnesium phosphate, magnesium ammonium phosphate, or a mixture. This precipitate was slightly brown in color and settled out fairly quickly, while the calcium precipitate did not visibly settle out at all over the minutes before I filtered it. Presumably both of these differences owe to the crystals being larger or rounder than the calcium precipitate. The magnesium precipitate tastes the same as the calcium precipitate.
Upon drying, the calcium precipitate has a consistency somewhat like dried mud; I can pick up pieces of it with my fingers and break them apart with my finger relatively easily. However, rubbing it between two fingers breaks it into a white powder too fine to have any gritty feel, rather like cornstarch. So it seems that the ammonium chloride (or whatever) that is binding together the crystals of apatite (or whatever) isn’t able to hold them in clumps of more than a micron or so in size; it might in fact just be van der Waals forces between the apatite crystals.
After a couple of days of room-temperature air-drying, in the calcium precipitate one crystal large enough to glint in sunlight can be seen from the proper angle, but the rest of the powder still appears as a matte-white, purely Lambertian surface. Some 10% contraction on drying is evident.
I have not managed to acquire clay yet, but it occurs to me that this calcium phosphate powder (if that is what it is) is probably an adequate alternative and may be a superior one. The grain size is about the same as that of clay, the expansion upon absorbing water is probably smaller than clay’s and perhaps insignificant, the crystal habit can be made to be needlelike or (like clay) platy, the price is only a little higher, and the aspect ratio of the grains should be only a little worse. Where it might be superior is that the apatite cement I propose to selectively deposit can clearly bond well to these grains, while its ability to bond to grains of clay remains a significant unknown. Also, the lower expansivity might enable it to produce a higher-density final composite material.
A 2011 paper from Gargouri et al., “Synthesis and Physicochemical Characterization of Pure Diammonium Phosphate from Industrial Fertilizer”, explains that in Tunisia the “diammonium phosphate” industrial fertilizer is only 75% diammonium phosphate, the remainder including “Co, Cu, Fe, Mn, Mo, Ni, Zn, F, As, Al, Hg, Pb and Cd”. They report getting their cheap industrial DAP almost as pure as the laboratory DAP they bought from Fisher, simply by recrystallizing it with 70% water and 30% alcohol, decoloring with charcoal. They report these results:
| ppm | Fe | Al | Mg | Ag | As | Co | Pb | Hg | Si | Sn | Ti | Cr | Zn | Cd | Cu | Ni | Mn | V |
| plant DAP | 6769 | 4273 | 4907 | 6 | 26 | 5419 | 22 | 3 | 150 | 382 | 93 | 525 | 1203 | 34 | 59 | 25 | 65 | 1341 |
| plant DAP recrystallized (water-alcohol) | 24 | 37 | 14 | - | 3 | 3 | 7 | - | 70 | - | 2 | 27 | 41 | 3 | 4 | 17 | 1 | 47 |
| commercial DAP (Fisher) | 15 | 22 | 9 | - | 3 | 2 | 7 | - | 38 | - | - | 25 | 11 | 3 | 2 | 17 | - | 9 |
This amounts to a reduction from 2.5% of these impurities down to 0.3%. 2.5% is a lot less than 25%, and I’m not sure what happened to the other 22.5%; it might be impurities they also removed but didn’t measure, such as O (for example in OH or SiO₂), Ca, and F. Their analysis of the P and N content before and after their purification (46% and 17.7% before, 49% and 18% after) does not support the possibility that 25% of the original material was made of non-ammonium, non-phosphate components. However, some of the “25% of impurities by weight” they cite might have been compounds like ammonium fluoride and magnesium phosphate. Or maybe it was just a typographical error where they were missing a decimal point.
I should see if filtering with charcoal reduces the brown color. Also, especially if I can get vacuum filtration set up, recrystallization as per the standard procedure would eliminate impurities that are still soluble in the solution after cooling. Greg Sittler suggested a water-driven venturi as a vacuum pump.
Another approach is to make the solution basic, which will precipitate hydroxides of (among other things) iron, nickel, copper, and cadmium, but not ammonium or phosphate:
g. All hydroxides are insoluble except those of the alkali metals. ... Ammonium hydroxide does not exist.
The usual way to do this is with lye, but I don’t have access to lye; however, household ammonia solution should also work. Also, sodium carbonate or sodium bicarbonate, which I do have, would precipitate the transition metals (“e. All carbonates, sulfites, and phosphates are insoluble except those of sodium, potassium, and ammonium”), but by the same token you would think those would be precipitated already in a phosphate-rich environment. (Iron(III) phosphate, ferric orthophosphate, is slightly soluble in water, but probably not enough to give the brown color.) So maybe I should try it and see what happens but not expect success.
Previously I’d written that you’d want to get the ammonium chloride out of the finished piece by leaching it out with water. But ammonium chloride evidently dissociates and “sublimes” at 337.6°; initially I thought the mix of corrosive gases it produced would be something I wouldn’t want around, but apparently the gas on cooling re-neutralizes to ammonium chloride rather than going around corroding solid objects it encounters, so that might actually be a reasonable way to remove the side product.
I guess the immediate next step is to dissolve some calcium chloride in water and soak a little sand, a little of the supposed calcium phosphate powder, and a little of a mixture of both with it, then let it dry. Actually ideally I would do this with both calcium chloride and what I suppose to be DAP in order to see what the resulting substances are like, since I suspect that calcium chloride in between the grains of filler will work better than the other way around because it will favor needlelike nanocrystals. But that might turn out to be wrong.
A further thing to try might be to use different pH levels. I have household ammonia to alkalinize the mix pretty thoroughly, but like the ancient alchemists, no strong acids.
As always with scholarship, there is danger from the thoughtless prejudice of the ignorant, which so often has turned into violence, as in the cases of Giordano Bruno, Aaron Swartz, Alan Turing, the Maya codices, and Qin Shi Huang’s burying of the scholars.
Ammonium chloride is on the national list of “controlled chemical substances” which unauthorized people are not allowed to have or make; but, then again, so are everyday products like aqueous ammonia solution, lye, acetic acid, ethanol, isopropanol, methyl ethyl ketone (the solvent in dry-erase markers), quicklime, slaked lime (whitewash), acetone, ethyl acetate (nail polish remover), red phosphorus (as found on matchboxes), nitromethane, sodium carbonate, sodium bicarbonate, and kerosene. Phosphorus, hydrochloric and sulfuric acids, toluene, dichloromethane, and acetone are even in “list 1” along with actual drugs I won’t mention here. Ammonium chloride is in “list 3” along with ethanol, isopropanol, sodium sulfate, and kerosene.
The definition of “product” is something of 30% purity or better of (the total of) substances from lists 1 or 2 P/V (which I suspect means “per volume”), or 20% purity or better of hydrochloric acid or aqueous ammonia; except that if it’s impossible to separate the substances by physical means, higher concentrations may be approved on a case-by-case basis. Perhaps this is the reason I can buy vinegar at the grocery store, lye at the hardware store, and nail polish remover at the pharmacy, even though they are all in list 2: they are dilute.
Notably absent are sulfates (of anything but sodium), sulfur trioxide, sodium percarbonate, phosphoric acid, nitric acid, and nitrates (though sodium nitrite is included in list 3).
So, I think as long as I stay away from acetone and hydrochloric and sulfuric acids, I shouldn’t run into any pitchfork-wielding peasants.