Only sharing this because I can - it is risky, not healthy and required proper PPE !
What is marking fluid?
Essentially ‘just’ chemicals that bond with the material.
To make it durable though it has to be able to chemically bond with the material, or where that is not possible (like glass) to bond through heat.
In theory one could use powdered PETG to melt it onto a surface but where’s the fun in that…
Molybdenum trioxide is the main ingredient you require.
Comes as a yellowish to whitish powder or in crystal form.
It dissolves reasonable well in Hydrogen Peroxide of 30% or above concentration.
For lower concentrations you require a lot of stirring and time but don’t bother with anything below 15%.
Once all is dissolved you can either wait or slightly heat it to break down the peroxide - this will make the yellow stuff crystallise out as ultra fine particles.
The tricky part is finding a suitable binder.
Tons of binders out there - not many able to stand up against the tri-oxide…
What works reasonably well is birthing/fisting lube in powder form.
Turns into like a slippery slime but if used in only minor amounts makes a great surfactant and it holds the fine particles in suspension for a very long time.
Gum Arabic works to some degree but decomposes rather quick, only lasts for a day or two as the trioxide is a strong oxidiser.
In terms of more exotic binders as those commonly used for industrial paint - stay away!
The trioxide will break down most them sooner rather than later and the result isn’t good when it happens.
For ready to use batch sizes you can even try white wood glue thinned down with water as a binder.
For faster drying one might try alcohol mixed in but I think the trioxide would break it down after a few days.
Why use the trioxide and not just the black Molybdenum powder to make a marking fluid?
Power - with a diode laser the black stuff really struggles to bond.
On most surfaces the bond will happen through the heat the powder takes on but on glass or metal the results are not really lasting.
The trioxide will be badly affected already by a 10W laser and you don’t even have to go to snail speed…
The laser and the created heat start a chemical reaction during which the trioxide breaks down while releasing energy - this extra energy is what provides a chemical AND thermal bond to almost any surface.
Meaning the stuff actually migrates into metal surfaces while the rest on top turns rock hard.
Let’s say you not only found Molybdenum Trioxide at a more or less reasonable price but are also dumb or experienced enough to try mixing you own solution…
Then I guess you first reaction is somewhere in the region of “Are you nuts??? the stuff costs a fortune!”
It sure does but same is true for any quality laser marking fluid you can buy in the shop.
Thing is that you need far less of the trioxide than what you might think 
Ok, how to get going with the this stuff?
Put on your PPE and prefer to do it outside in an area where the wind won’t surprise you - you DO NOT WANT to breather in this yellow powder and you don’t want to get it on your skin, let alone into your eyes…
Take some tiny glass jar with a properly sealing lid, like these tiny jam jars or such.
Fill it to about one quarter with Hydrogen peroxide - 20% or above strength preferred.
You don’t want more than about 15 to 20ml for a first test batch…
Add a small amount of the powder and stir until it fully dissolved.
No need for a fancy stir bar if you don’t have one, the rear end of a plastic spoon will do fine…
Don’t get too frustrated as it takes a bit more time than dissolving sugar or salt…
Once it is all gone add a little bit more powder - if you prefer take notes about how much you added as it helps if you ever need a larger batch size.
When it starts to take a really long time to dissolve stop adding powder.
If need be take a break and mix again once the fingers have enough feeling again.
DO NOT screw the lid on and start shaking…
The hard part is over, now for the even harder part - waiting…
While the lube can be added now as it should not react with the hydrogen peroxide it still binds and gas bubbles that will form…
I made the mistake once to add some thinned down latex paint right away and ended with a big foamy mess that I had to let drip down into a larger jar.
The peroxide will break down naturally, especially if the jar is placed in sunlight.
Once enough peroxide is broken down to water the trioxide will participate out of the solution, forming micro crystals which will slowly sink to the bottom.
Once a day or so give it a little stir so the crystals won’t bind fully together.
The liquid on top will turn clear, indicating all of the trioxide is now out of solution.
Do a little test with a drop of the liquid and some yeast - if it foams then there is still peroxide in it.
If the liquid is fully clear like water once all is settled for a while you can suck the liquid out with a syringe - leave a tiny bit on top of the yellow stuff so you won’t suck that in.
Now you can add water (no need if the peroxide was fully broken down and you did not remove the liquid layer) and your preferred binder.
The key is to create either a thick slurry that you can brush on to give a THIN and even coat or to have it runny enough to use an airbrush gun - for the later go outside to ensure you won’t breathe in any mist…
How to apply it?
Grab some scrap metal, like the lid of a cookie tin can - the inside is usually untreated and unpainted but a fine sanding to be sure won’t hurt.
Clean with soapy water followed by an IPA wipe.
Mark out a few spots you want to engrave and prefer to use the same logo, text or whatever.
Apply a really thin coat of the marking mix you made - a slight haze on the metal.
Only apply where you actually engrave, don’t paint the whole thing…
Set your 10W laser to 90% power and engrave your first logo or just square with a speed that works good for you to engrave on wood.
Check the result after rinsing off the marking fluid - preferable in a small plastic bowl or such, not in your sink…
No effect means going slower, a thick black marking protruding out of the surface means you have to engrave faster.
If you do get a good effect that can’t be removed with rubbing hard using the the rough side of a kitchen sponge but that looks pale use a second coat - to do not make the first coat thicker, do it like you would when applying spray paint.
What you want to find is the right balance between the amount of marking fluid on the surface and the engraving speed.
The first defines how black it gets and how thick the marking will be.
The second provides the penetration and required heat for the reaction.
If you get a really dark black but it wipes right off you might have a too thick coat or engraved too fast.
Once you honed in on the values that give you best result write them down!
The remaining 10% on your power scale are are the margin to work with for different materials.
E.g.: For stainless you might want to use 100% and the speed rather on the slower side while for polished Aluminium you might get better results going very fast on the first pass to just create a dark grey or black top to then do the second pass with the usual settings.
If you want to engrave glass this way prefer to go THROUGH the glass…
Trying to engrave from the top usually ends with micro-fractures in the glass and a very rough look and feel as the marking stuff creates too much heat.
Disposal…
Do yourself a big favour and used a water based binder.
With that you can use water and a brush to clean the engraved part and can also keep this water for further washes.
If you have a warm enough place you can just let the water evaporate from the bowl…
What’s left, if the engravings went well, should not have too much black in it.
That yellowish to grey residue can be used again
Just collect it and once you have a bit replace about half of the fresh oxide powder with the recycled remains - be aware that it will have some binder in it and that it might foam a bit while the peroxide disappears.
I prefer to just mix the recycled stuff with a bit of water until the consistency of the binder is just right.
Great for larger engravings where a true black or 100% perfect result is not that important.
And yes, in case you wonder> You can use the heated bed to speed up the drying of the marking fluid but do not cook it…
Great, but I have the 40W laser or a CO2 laser…
Good for you…
Same as above applies for the application of the coat(s) to get the desired thickness and durability.
Only that you can engrave with much higher speeds and/or higher power.
What’s 9W gonna help you?
Probably not much as down there the output power usually isn’t that linear, so go with what works great on wood to get going.
With the extra power you have at your disposal you can do some nice things as well.
Like going slow and with full power on stainless to create a surprisingly deep engraving - or with more marking fluid applied a marking that ‘stands out’.
With a very thin coat that you basically burn away you can even create coloured engravings on stainless steel and with a bit of fiddling even on titanium.
Here the marking fluid only provides enough thickness to allow for a blackening effect that allows the laser to heat the material far more than what it otherwise could.
Takes some time to hone in the settings and works best using these test patterns people use to define the speed and power ranges for greyscale engravings.