Normally you need high-power lasers (kilowatts) to mark metal. One alternative is to lay down a layer of resist (for example, black electrical tape, or permanent marker) and use a small laser (tens of watts) to burn off the resist, then etch the metal electrolytically, with acid, or with reactive ions.
It occurred to me that when you’re burning a layer of resist off the metal, you’re also producing reactive ions, and maybe those could etch the metal directly if the “resist” were chosen for that purpose. This is different from normal reactive ion etching, where you first lay down a layer of real resist, and then later put the resist-patterned workpiece into a vacuum chamber that you fill uniformly with a plasma; in this technique, the “resist” is the plasma, after you blast it with the laser, so you can do the whole thing at atmospheric pressure. This saves you the trouble of running corrosive byproduct gases through your vacuum pump.
For example, nickel (tetra)carbonyl boils at 43°, and maybe you could form it by heating an oxalate or formate salt on a nickel surface; disodium oxalate, for example, decomposes at 290°, releasing carbon monoxide. Similarly, iron (penta)carbonyl boils at 103°. (To make iron carbonyl, sulfur is used as a catalyst and the process is normally carried out at 5-30 MPa.) Carbonyl complexes are also known of molybdenum, chromium, manganese, cobalt, titanium, ruthenium, rhodium, osmium, iridium, platinum, tungsten, and vanadium, not to mention many more reactive metals, so perhaps they could be etched in the same way, but it might not be possible for all of them; it might be easiest for nickel, followed by iron and rhenium. Doing the whole process under a liquid solvent capable of dissolving the carbonyl complex, such as glacial acetic acid, acetone, or carbon tetrachloride, would likely help with increasing the pressure (and thus the equilibrium carbonyl-complex concentration), controlling the temperature, and reducing hazardous carbonyl emissions.
Carbonyls also decompose under heating to deposit the metal, and so if you were to laser-heat points on a surface in a metal carbonyl atmosphere, you could selectively deposit the metal. (Historically this chemical-vapor deposition process was used for nonselective nickel plating, but abandoned for reasons of toxicity.) Many other chemical vapor deposition processes could be selectively applied in the same way by localized laser heating. This is called laser chemical vapor deposition, and of course electron-beam-induced deposition is a higher-resolution vacuum variant of the same technique.
Alternatives to the ridiculously toxic carbonyls might include seriously toxic organometallic complexes like tetraethyllead (boils at 85°), diethylzinc (boils at 117°, pyrophoric), or triethylaluminum (boils at 128°, hypergolic with liquid oxygen.) In the particular case of etching aluminum, chloride salts such as ammonium chloride (reversibly decomposes at 337.6°) could perhaps be used to produce aluminum trichloride (sublimes at 180°). If cupric chloride (melts at 498°, decomposes at 993°) were the chloride salt, you might be able to deposit some metallic copper at the same time as etching the aluminum.
Either for etching or for deposition, you could probably use other CVD feedstocks like tungsten hexafluoride; silane (pyrophoric gas at room temperature), trichlorosilane, or TEOS; ammonia (e.g., for depositing silicon nitride); germane; ferrocene (forms from hot cyclopentadiene reacting with iron pipes, boils at 249°); uranium hexafluoride; uranocene; nickelocene (formerly used to prepare nickel films); cobaltocene (sublimes at 171°); cyclopentadienylcobalt dicarbonyl (boils at 140°); pentachlorides of molybdenum, tantalum, and titanium; methane; etc.
If you can start by cooling the substrate metal to cryogenic temperatures, you might be able to use “resists” that are unstable or evaporate at higher temperatures. Fluorine, for example, boils at -188.11° at one atmosphere, and somewhat more practical cryogenic temperatures at higher pressures, and so if you cool your substrate enough, you can bathe it in liquid fluorine which you then encourage to etch it with local laser heating. A liquid “resist” would have the earlier-mentioned advantage of producing higher pressure through confinement, thus making it much easier to produce large complexes containing low-boiling substances like tungsten hexafluoride, and also permit repeated etching of the same spot.
At higher temperatures, such as room temperature, you ought to be able to use other liquids that are fairly inert until heated with lasers. Gasoline or mineral oil, for example, could serve as a source of hydrogen for etching carbon, silicon, or glass, while perfluorohexane or hexafluorobenzene could serve as a source of fluorine for etching, say, tungsten or glass. And of course there are all kinds of metal/acid combinations that etch not at all or very slowly at STP, but which, at high temperature and pressure, or when a surface film is disrupted, etch very rapidly.
There are a surprising number of very stable materials that have low-boiling hydrides, silicon and carbon of course being the most conspicuous, but boron also does. Virtually any acid “resist” could contribute protons to facilitate such etching, but phosphoric acid seems like one of the most promising alternatives due to its high boiling point (over 800°) and willingness to part with its hydrogens. At lower temperatures ammonium compounds might be a promising alternative.
Arcs, as in electric discharge machining, are another possible way of inducing localized high temperatures and pressures at chosen points on the surface of a workpiece, and so could also serve either to promote the deposition or the destruction of solid material at the surface of a workpiece.
Traditional reactive ion etching feedstocks like sulfur hexafluoride and carbon tetrafluoride would also be viable candidates for producing the reactive ions on the surface of the workpiece, but heavier perfluorocarbons like perfluoropentane, or sometimes even carbon tetrachloride, would probably work better, both because by virtue of being a liquid at room temperature, they have several hundred times more fluorine atoms close to the surface, and because of the increased pressure as mentioned above. Gases like carbon tetrafluoride dissolved in some kind of solvent might also work better than just gases.
Glow-discharge plasmas are another possible way to supply reactive ions. By moving one or many sharp-pointed electrodes around close to a workpiece, you could selectively apply the reactive ions to certain parts of it; this is routinely done with air for “activating” surfaces with nonthermal plasmas, including even biological tissues, but of course other gases would have different effects. The reactive ions will tend to attack the sharp point as well, so it either needs to be inert to the ions, or consumable like a mechanical-pencil lead.