I'm curious though:  Water is H20, Hydrogen and Oxygen molecules.  So if the lamp was put under vacuum (air pumped out) wouldn't just hydrogen be left? I'm not a physicist of course. 

If you left water inside the lamp, as you apply the vacuum the water would start to boil, at 29.22 inches of mercury it will boil at 69 deg F and turn into steam, this will just get sucked out by the vacuum, even if you separated the oxygen from the hydrogen though chemical reaction and heat it'll just get sucked out by the vacuum pump until nothing is left, but how Medved is showing the separation (which chemically will happen as soon as you heat the tungsten)  the Oxygen from the water and tungsten combination when heated will allow the oxygen in the water to form tungsten oxide and this leaves the hydrogen free to be itself the balanced chemical equation would look like this WO3(s) + 6H+ + 6e- <-> W(s) + 3H2O  and this reaction is fully reversible so as you cool it you'll see the tungsten turn from oxide back to tungsten on the cooler spots so the outer glass and the hydrogen will recombine with the oxygen and the process will keep repeating until you blow the lamp.. But a single drop of water it might take several tens to hundreds of hours but don't expect to get the full life out of the bulb..

Homebrew if you want to make Hydrogen all you need is DC power two wires and water, just stick the wires in the water and turn on the power the negative wire bubbles will be hydrogen and positive will be oxygen, and if you want to make it in quantities you might want to use platinum foil on the end of those wires, copper doesn't last long. Just stick a jar full of water over the wire and it will fill up with your gases, and if you mix the two and use a match you'll be left with water and very loud bang and maybe a busted jar (yes it'll explode, and yes your parents will get mad if they hear it, use a plastic pop bottle and a yard stick to hold the lit match, and do it outside, and keep in mind the explosion might propel the bottle like a rocket watch out so your not in the flight path, and your gun totting neighbor might come out thinking some one is shooting up the place, you might want to give them a heads up)     

I knew it is explosive. .. As soon as i read the whole cycle thing producing hydrogen the word that came to mind was Hindenburg lol.. Kinda messed up but thats how i remember hydrogens capabilities... Fun thing is there is literally pallets full of thousands upon thousands of glass jars at that abandoned factory lol... Might pull some hooliganry ... With proper safety procautions of course

I think Water can get in in a couple ways...

 - Most GLS lamps are not Vacuum, they are filled with Argon at above atmospheric pressure. I think we can consider what would happen when the Water is introduced in the Argon flushing stage i.e. as if the Argon is contaminated with Water

 - Some of the lamp's structures might have an atmospheric Air bubble inside. The biggest would be the internal part of the stem tube, when the center support wires are added and the tube is melted back and pressed with a plate (to form the stem head) - the volume of this tube might seal off, and later crack (some little invisible hole) into the inner lamp atmosphere at a later stage



You know the Water bottle rocket, thats launched with a Wine cork and bike pump

I made a version launched electrically. Same Wine cork, through it is going electrical flex, connected to 2 electrodes inside (a Plastic tube rolled from PET from another bottle, with 2 sections wrapped with Aluminum foil). The other end of th flex plugs directly into 230V. The thing boils (and electrolyses ? but then the Hydrogen and Oxygen would be produced alternatingly by each electrode, so be mixed right away) and the pressure rises

After a few minutes it shot the rocket quite well, though it hit the corner of the roof of the house....



With DC i think one electrode would be inherently protected by the electrolysis process (though it might become plated over with material from the other electrode)

For the other one (or for both if the system is supposed to stay unharmed when not in use for long periods), wouldn't Stainless Steel do and be a bit cheaper than Platinum ?

Ash Stainless would work but it will still suffer over the long term from corrosion you can literally watch is go from shiny to brown sludge forming in minutes and longer term you get Hydrogen embrittlement and you have to deal with the toxic Hex-Chrome to Tri-Chrome by products, sure small scale it's not an issue but doing it in bulk Platinum is the way to go the electrodes last basically forever and when you start playing with platinum foils like 0.025mm there really isn't much weight to them so the cost of like 30 cm2 is only like $15 USD, we aren't talking aluminum foil thickness, they are very thin, or for a few bucks more ($35) you can get them fused in glass with rubber stoppers so they fit nicely in most lab glassware, but if the platinum is still to expensive just for playing around with, you can always use Carbon rods they also work very well, and produce virtually no toxic by-products but if they are copper coated you have to remove the copper or you end up with toxic copper by-products and the water will go blue/green...

Water is immensely damaging even to gas filled incandescent lamps, due to the water cycle described by Medved. A single droplet of water in an argon cylinder capable of filling a half a million lamps would cause every one of them to fail early.  Water content in a gasfilled lamp has to be reduced well below 1 part per billion to achieve rated life, and for metal halide lamps even greater purity is needed.  Quite apart from 'droplets' of water, it is essential to bake out water molecules that are adsorbed onto the surface of the glass and metal parts.  And just pumping vacuum to make the water evaporate away is not enough, because water molecules are chemically bonded to the glass surface.  It's necessary to heat the glass almost to its softening temperature during pumping, and even that is not enough - hence the use of chemical getters to eliminate water molecules that continue to escape after the lamp has been made.

For anyone interested in the detail of this topic, I can highly recommend GE's vacuum book from the 1960s - The Scientific Foundations of Vacuum Technique by Saul Dushman.  He provides a good account of Irving Lamgmuir's research on this topic, which made the gasfilled incandescent lamp possible.  It is one of a few classic books still used today by every serious lamp production engineer.

Interesting, I will definitely keep that book in mind, did they manufacture incandescent lamps with getters? Also I did not know getters could get rid of water...

Incandescent lamps were made with getters after 1894.  The first getter was developed by Arturo Malignani at an Italian lamp factory, and GE immediately sent its engineers there to study and copy the process.  It was subsequently adopted by all lampmakers worldwide.  Malginani's getter consisted of red phosphorus suspended in alcohol, combined with the process of "Blueing" in which the lamps are overloaded at higher voltage than normal.  This creates a glow discharge around the filament at the same time the phosphorus is vaporising, and the electric field accelerates the atoms of impurities sufficiently that they are removed from the lamp atmosphere. 

Red P was the main incandescent lamp getter until 1969 when GE developed Phosphorus Pentanitrile (P3N5).  This has the advantage that it works for much longer.  Red P is only an effective getter during the blue glow, and it stops working when lamps leave the factory.  P3N5 keeps on working during lamp life, and is more effective at gettering moisture as well as oxygen.

Higher power incandescent lamps or high value types like automotive versions also use zirconium/aluminium getter, a metallic paint applied to the leadwires or stems.  This substance sorbs hydrogen, which is present in tiny quantities in incandescent lamps and due to its high thermal conductivity, reduces lamp efficacy.  Higher output lamps can be made when hydrogen getters are used.  Its also standard practice to use this getter in soft white coated lamps, because the white powder brings a lot of moisture into the bulb - getting rid of the hydrogen helps to combat the water cycle in such lamps.

Interesting I've never actually seen one in a lamp.. I'm guessing it's not like a patch around the inner wall but like on the stem? You see when I think of getters I think vacuum tubes and Sodium lamps, because there's that very obvious patch of silvery material on the inner portion of the outer envelope...also unrelated but You know a lot about this sorta thing... Do you read books or are you a historian, or maybe a lighting engineer that knows history? I don't mean to throw this thread off track but I've seen what you post around and it's more in depth than most others I have seen.

So that is what the little black spots are on the stem that look like they are painted on, I've always wondered about that, thanks James for explaining it..