ECM for machining nonmetals?

I’ve looked a bit into electrolytic machining (usually “ECM”) of glass, but I think probably it’s not a good idea; sandblasting (“abrasive jet machining”) is a dramatically more reasonable idea for glass. But I was thinking about marble, and I think electrolytic machining would probably work really well for marble.

Electrolytic machining of marble and similar basic minerals

Suppose we use a neutral, very dilute NaCl electrolyte, saturating the pore spaces of the marble, with a high enough overvoltage that most of the electrolytic product is hydrogen and oxygen (1.23 V) rather than sodium and chlorine. At the anode, we produce (conventionally) H⁺ ions and oxygen by stripping an electron from water: 2H₂O → 4e⁻ + 4H⁺ + O₂. If the anode is close to the marble or in contact with it, these H⁺ ions (really hydronium, H₃O⁺) will immediately react with the CaCO₃ to reform water and carbon dioxide: CaCO₃ + 2H⁺ → H₂O + Ca⁺⁺ + CO₂. Thus the marble around the anode will be eroded, neutralizing the acid and producing CaCl₂. If the anode is either silver-plated or made of carbon, I think it will not itself suffer erosion.

We can pump a fresh stream of electrolyte constantly through a hole in the center of the anode; if the anode is sheathed in an insulator, such as teflon, the hole in the insulator where the electrolyte squirts out can ultimately determine where the workpiece erosion happens, rather than depending on the geometry of the possibly-eroding anode itself.

I don’t think much chlorine gas will be produced, if any, because the electrons being sucked out of the anode are coming from the abundant hydrogen ions, which replace them from the abundant marble, rather than from the scarce chlorine ions. If that does turn out to be a problem, alternative electrolytes exist.

Aside from lime and its carbonates, the same anodic-attack approach should work with other nonconductive minerals subject to easy acid attack, such as dolomite (Mg/Ca), siderite (Fe), smithsonite (Zn), magnesite (Mg), malachite (Cu), azurite (Cu), and perhaps portland cement (calcium silicate hydrate).

A waste product of strong alkali will form at the cathode, though perhaps this could be ameliorated with a suitable buffer, perhaps based on acetate, citrate, or borate, to compensate for the buffering the carbonate provides to the acid produced at the anode; lacking this, ultimately the liberated calcium ions will find their way to the cathode and precipitate slaked lime.

Choice of electrolyte

It’s important for the anions in the electrolyte to maintain the calcium ions in solution, unlike phosphate (apatite), pyrophosphate, oxalate (weddellite/whewellite), fluoride (fluorite), hydroxide (slaked lime), sulfate (gypsum), tartrate (beerstone, barely soluble) or titanate (perovskite). Chloride is a good choice, but other choices include bromide, iodide, cyanide, thiocyanate, nitrate, acetate (34.7 g/100mℓ), chromate (2.25 g/100mℓ), or formate (16 g/100mℓ at 0°).

Calcium borate is a weird boundary case. In dicalcium hexaborate, the least soluble borate of calcium, water can dissolve 202mg/100mℓ of boria, which works out to (* 202 (/ (+ (* 2 40.078) (* 2 15.999) (* 6 10.81) (* 9 15.999)) (+ (* 6 10.81) (* 9 15.999)))) = 310 mg/100mℓ of the salt, though US Borax gives 470 mg/100mℓ.) Generally, borates are complicated and not very soluble, much like silicates, phosphates, and silicoaluminates, because of the possibility of oligomer or polymer formation.

Choice of anode material

It’s simultaneously desirable to use anions that won’t form soluble salts with the anode material itself, both so you don’t end up with nasty anode salts all over your nice cut marble, and so you don’t have to keep feeding in more anode as it’s consumed (and suffering imprecision from anode wear uncertainty). A gold-plated anode would permit the use of just about any electrolyte (except maybe cyanides, which have other disadvantages), and even silver should resist chloride and the other halogens (except fluoride). Ordinary copper would permit thiocyanate, and lead might permit the use of iodide and bromide, though the resulting lead salts would be soluble enough to pose real risks of contamination. Because copper is lower in the reactivity series than hydrogen, you’d think it could avoid forming copper chloride in this use, but in fact copper plating using chloride or acetate baths is totally a thing. I have definitely anodically destroyed copper in salty vinegar.

(Of course, graphite or carborundum electrodes will withstand arbitrary acid or base attack at ordinary temperatures, and platinum electrodes withstand nearly any reactive environment.)

Here’s a solubility chart formulated for the purpose:

† indicates compounds that I don’t think will form electrolytically from unoxidized metal and relevant anions.

(Ugh, I don’t have zinc in the chart. But it’s almost the same as magnesium. Also, I don’t have tartaric, lactic, and phosphoric acids.)

I tried reducing the above solubility chart to an easier-to-use form a few times, but I never succeeded.

Alternative solvents

Another approach is to make the electrolyte from ions dissolved in a polar solvent other than water; for example, anhydrous ammonia, formamide, tetrahydrofuran, acetone, isopropanol, methyl ethyl ketone, pyridine, DMSO, dichloromethane, or deep eutectic systems; these will yield different solubilities for various ionic substances.

For example, only 0.2 grams of muriate of potassa dissolves in 100 mℓ of DMSO, and 0.013 g of potash, but it can dissolve 20 g of the iodide or 30 g of muriate of Mars, while the muriate of lime is entirely insoluble. In DMSO, hydrated cupric acetate and muriate are insoluble, but the acetate and muriate of zinc are quite soluble, as are the muriates of tin, so a copper anode with a zinc-muriate electrode dissolved in DMSO might be able to electrolytically etch salts of tin or iron with impunity.

Some solvents may not produce electrolysis products of their own that are useful for the electrolytic etching process, the way water does; for example, the anodic reaction converting carbonate ions to oxygen and carbon breaks apart and then reforms water molecules in the process. DMSO in particular seems likely to produce electrolysis products much more noxious for human life.

Etching with a cathode instead of an anode

If the electrolytic cell’s cathode rather than its anode were the active tool, it should work for acidic or amphoteric materials attacked by strong bases, most notably sapphire (slowly, at tens of megapascals and 400° or higher), gibbsite, and amphoteric oxides like those of zinc (philosopher’s wool, a refractory (1974°) thermochromic transparent piezoelectric direct-bandgap n-type semiconductor with a 3.37-eV bandgap), titanium (rutile, a UV-blocking strongly birefringent photo-superhydrophilic photocatalytic transparent refractory (1843°) n-type semiconductor with refractive index 2.61 and a 3.05 eV bandgap which becomes an excellent dielectric when stoichiometric), tungsten (an electrochromic photocatalytic semiconductor), vanadium (a refractory (1967°) transparent semiconductor with an 0.7-eV bandgap that becomes metallic and IR-reflective in 100 fs above 68°, a temperature that can be adjusted with tungsten doping), and tin (cassiterite, a refractory (1630°) n-type semiconductor with a refractive index of 2.0 and a specific gravity of 7).

The amphoteric oxides can be etched just as well by the anode-acid process described at the start, but etching them with a cathode means you can use any metal for the tool electrode, since it won’t be vulnerable to anodic dissolution.

Metal sulfides might be another candidate. Leaching with very dilute alkali has been successfully used to separate antimony from stibnite (antimony sulfide) without affecting other metals, with etching speeds around 10 microns per minute. Alkaline leaching has also been used to extract lead, tungsten, zinc, vanadium, and chromium from various ores. Mostly, though, these processes are very slow.

These cathodic etching processes, instead of producing waste alkali at the cathode, would produce waste acid at the anode, and the same comments about buffering apply to avoiding undesired acid accumulations.

More speculative directions to explore

A very interesting question for this kind of electrolytic work: these semiconductors, zinc oxide, rutile, tungsten oxide, vanadia, and cassiterite, are all immune to anodic dissolution; but they are amphoteric enough to be unstable for this kind of work. But perhaps other nonmetallic semiconductors other than carbon and carborundum, such as GaN or InP, may be alternative electrode materials.

Pulses of high voltage on small-diameter electrodes should be able to produce plasma discharges to overcome activation barriers, a sort of corona-discharge EDM/ECM hybrid, though this would surely also erode the tool electrodes.