Fiberglass for insulation is super cheap (Potential local sources and prices of refractory materials reports that it’s around US$3/kg) and has substantial tensile strength, even if it’s not as high as S-glass and M-glass. Chopped glass fiber for mixing into plastics as reinforcement (probably E-glass) also sells for about US$3/kg (AR$9975 for 20kg). (I think fiberglass insulation is A-glass.) Glass fibers are commonly used to reinforce organic polymers like epoxy to make them roughly as strong as steel, but this matrix is itself fairly expensive in bulk; 6 kg of epoxy from Tienda Bepox costs AR$21000, US$129, or US$21/kg, seven times the price per unit mass of the fiber reinforcement. Polyester resin is significantly cheaper, at AR$7300/10kg, US$4.50/kg, but also much more rigid and, I think, more fragile. (Vinyl ester and phenolic casting resins are apparently harder to find, locally at least.)
For polyester thermosets, Matweb gives useless ranges of 10–123 MPa for ultimate tensile strength and 54–265 MPa for flexural yield strength (implying that at least one polyester thermoset has a flexural yield strength twice its ultimate tensile strength), and for “epoxy cure resin” 5–97 MPa and 76–1900 MPa, and for “epoxy, cast, unreinforced” 8–97 MPa and 14–131 MPa. And I have no idea where in these very wide ranges these resins are.
Matweb says fused quartz (Saint-Gobain Quartzel) is 6000 MPa UTS and generic A-glass fiber is 3310 MPa UTS, while Micarta RT500M glass-reinforced epoxy is only 269 MPa UTS and “Goodfellow E-glass/Epoxy composite” is only 490 MPa UTS. So there’s a lot of room for improvement.
Could you make a ceramic-matrix composite from insulation fiberglass, getting the usual cracking-resistance performance improvements of CMCs, even if the improved performance doesn’t approach the kind of performance you get out of things like SiC/SiC CMCs? Could you do it cheaply?
You clearly can’t heat it up to the high temperatures normally associated with CMC manufacturing; those are over 1000° and glass wool craps out around 230–260° (reportedly, I haven’t tried, and I suspect this may just be the max temp of the polymeric sizing — soda-lime glass normally doesn’t soften until around 700°), while rock wool (which is also about US$3/kg) should maybe be good to 700°–850°. So you might need some kind of hydrothermal process. Moreover, the hydrothermal process needs to be gentle enough to not eat the glass reinforcing fibers, so in particular ordinary portland cement and carbonation of slaked lime are probably off the table, and so is anything that involves making silica highly soluble in water.
A few candidate matrix systems:
To any of these binders you could add functional fillers such as clays, talc, mica, quartz sand, quartz flour, carborundum, perlite, vermiculite, glass foam, or sapphire.
Aside from the spatula layup approaches normally used to produce fiber-reinforced plastics, chopped-fiber mixing is a possibility, and spin-coating of the binder/filler/fiber mixture would tend to produce a biaxial orientation of the solids, which would tend to be very advantageous to both solids density and to applying the strength of the solids in the appropriate directions. This approach would be best at increasing the density of the solids if the binder decreases greatly in volume after the spinning, for example by dehydrating.
Another way to increase the toughness and flexibility of such a stiff composite material is to fabricate it in thin sheets and then laminate them together with a softer or weaker binder. The binder would tend to stop crack propagation or at least redirect it parallel to the surface of the material, and if it is soft rather than just weak it would tend to shear to allow the overall composite sheet to flex, like the pages of a book. Sodium chloride and highly-hydrated sodium silicate are two candidate binders that might be capable of such shear, especially over time.
Normally when you’re making glass-fiber-reinforced things you’re concerned with sizing the fibers to improve adhesion to the matrix, because otherwise the low-modulus, low-strength matrix can’t effectively transfer load to the fibers, and you get pullout failures at much lower loads than would be needed to actually break the fibers.
In CMCs the objective is the opposite: the matrix has strength comparable to the fibers, but both the fiber and matrix have the very low elongation at break characteristic of ceramics. Instead, the fibers are incorporated to arrest crack growth by bridging cracks, for which a much longer length of fiber must be recruited to stretch, for which the fiber needs to slide through the matrix. So sizing is necessary to weaken the bond between the fiber and the matrix, causing material failures to be pullout failures — still at a lower stress than would be needed to break a solid, defect-free block of the material, but with enormously improved toughness. This can even translate to higher tensile and flexural strength in the case of non-defect-free brittle materials, where microcracks can reduce their theoretical tensile strength by many orders of magnitude.
So, what kind of sizing could you apply to the glass fibers to weaken their bond to the matrix? You want it to be much softer than the other materials involved, and to completely cover the glass fibers, or at least almost so, which suggests that you’d like it to be amorphous. You want it to be water-insoluble so it will survive when the matrix is infiltrated into the fibers, but applicable hydrothermally rather than with some kind of white-hot gas or something. And you want it to be cheap, which eliminates most organics. Precipitating amorphous calcium phosphate or aluminum trihydroxide from aqueous solution might work; there are also a number of softer insoluble phosphates and carbonates of polyvalent cations, such as rhodochrosite, siderite, vivianite, wolfeite, smithsonite, hopeite, malachite, azurite, magnesite, struvite, and calcite. Oxide and hydroxide minerals might also be candidates.
Amorphous deposition isn’t essential, but if we hydrothermally deposit a crystalline sizing material, there’s a good chance there will be gaps between the crystals that leave the glass exposed.
Another approach would be to expose the fibers to something that reacts enthusiastically with the exposed glass (for which there are relatively few candidates at ordinary temperatures, mostly silanes) and forms a monolayer on it, and then something else that reacts with the monolayer, perhaps repeating the process several times.
Infiltrating loose glass fibers very briefly with a low-density hot plasma of something that’s solid at room temperature might work to deposit a thin, even layer on the surface of the glass without damaging it, if the heat in the plasma isn’t sufficient to melt the glass. The layer would tend to be of relatively constant thickness because the places on the glass where the plasma has already frozen will be hotter than the places it hasn’t frozen yet. For example, tenorite and maybe cuprite (Mohs 3.5–4) will tend to precipitate from a plasma formed by passing too much current through a copper wire which is allowed to mix with air. In an inert atmosphere, maybe you could use vapor of aluminum or zinc.
If the sizing layer is thin compared to the fiber diameter, which would be necessary in this crazy plasma process but maybe not some of the others, the cost of the sizing is less of a concern.