Many HID lamps have a halogen component such as metal halide and many UHP lamps. Do HPS and mercury vapor also have halides in them? Why or why not?

If the halogen cycle redeposits tungsten on the electrodes, why does the arc tube darken over time? Is it exclusively the thermionic emission coating?

Thanks

Good question! I may have an answer! metal halide lamps usually have an argon gas fill, with various metal "salts" HPS has sodium and mercury mixed together in an amalgam. Mercury vapor has mercury and argon, a lot of small computer projectors use UHP (ultra high performance) lamps some use metal halide lamps (not sure what those are called for projectors) there's XBO which is a xenon short arc lamp, just filled with xenon, and nothing else, while HBO lamps are metal halide short arc lamps. There is no halogen cycle in an HID lamp, what you see evaporating, is the metal salts or mercury, or sodium inside of the burner! (Anyone feel free to correct me that's all I know)

I think the halogen cycle is effective only for tungsten, while in MH burner other components, capable to darken the arctube, are present.
In MH the halogens are present to at first form salts, that evaporate easier then the pure active metals and as second (mainly in quartz burner types) to attract sodium, so it does not attack the quartz.
In quartz halide lamps the big darkening factor are sodium atoms, what for sure are not recovered by the halides. The only defense is to place something positively charged arount the arctube, so sodium ions are repelled towards the discharge center. Quartz shroud serve such function during the lamp life (it is charged by photo-emission excited by the UV from the arctube)

@RCM: "UHP" stands for "ultra high pressure", it use pure mercury as active material and utilize the continuum radiation present at high arc load discharge.

@Medved: UHP is commonly referred to "Ultra High Performance" rather then "Ultra High Pressure". MH lamps for projection called HMI usually but also MEI.
@Xytrell: UHP lamps are the only type of discharge lamps that includes a halogen (Bromine) to prevent arc tube blackening.
In the HPS, MH and MV lamps, the loading of the arctube isn't sufficient for the halogen cycle for the electrodes to become effective.
Halogen cycle on discharge lamps are effective only for projection lamps (HMI, UHP), when a very high power loading per unit length is required.

Really? I've seen 70W halogen capsules that are larger than 70W metal halide arc tubes. I've also seen the quartz from many (all?) arc-tubes glow red hot when turned off, and I don't see that when I turn a halogen lamp off.

@Xytrell: Arctubes of HID lamps operates at 1000degC or 1,832degF and even more. Halogen lamps operates at only around 600degC or 1,112degF.
Also ceramic arctubes cools down slower then quartz arctubes, results in a red hot glow for a longer time in ceramic arctube.

Exactly. So what do you mean by the loading not being enough?

The temperature of the electrodes and the arctube in projection lamps during operation is much more higher than in regular MV, HPS and MH lamps. The loading on the electrodes is also higher. This is sufficient to be able to react with a halogen and therefore to make a halogen cycle. In HPS, blackening from electrodes emitter isn't an issue as only the ends are blackens, but it is a problem because of the loss of sodium or leakage, which can blacken the arctube.
European MH and MV lamps still uses an anti-blackening emitter that turns the arctube white or gray.
In projectors the arctube blackening more of a problem, as the color of the image should be stable and the size of the source should be small.
Also, even in projection MH lamps (HMI), i don't think that halogen cycle is effective, because the halogens from the halides may prevent the free halogen from doing the halogen cycle.
UHP contains only mercury, so it is easy to include bromine for a halogen cycle.
Medved, Max or James may be know about this better.

In fact many general lighting metal halide lamps have an operating halogen cycle to keep the wall clean.  It common especially in smaller lamp types which require a high wall loading, and is nearly always present in the short arc versions like HMI, BriteArc, MSR etc.

Normally the salts in a metal halide lamp are added in the form of metal iodides.  These have a low chemical activity and don't really contribute to any halogen cycle.  However if we add the halides in the form of bromides instead (or occasionally even as chlorides such as in the old Thorn CSI/CSD types), their chemical activity is greater, and it's possible to achieve a working tungsten transport cycle to keep the walls clean. 

The increased chemical activity however does not only remove condensed tungsten from the relatively cool arc tube walls and deposit it on the hot electrode tip, it can also begin to remove tungsten from the relatively cool tungsten shank of the electrode, and transport that to the hot tip.  The result is called beavering - the electrode shank is eroded away in the same fashion that a beaver eats through a tree trunk.  Eventually the electrode will fall off and this kills the lamp.  There is only a very narrow window of having the correct thermal and chemical balance in the arc tube such that the halogen cycle will keep the wall clean without beavering the electrodes, and the rate of the reactions is sensitive to small thermal changes, e.g. use in different kinds of fixtures.  So in general lighting lamps its not common to set up a halogen cycle.  There could be difficulties, for instance, in a colder fixture, that the electrode erosion would proceed too fast and this could lead to premature lamp failure.  One lamp which does have a working halogen cycle however is Sylvania BriteSpot, where it was possible to control the thermal distribution quite consistently thanks to the arc tube being accurately pre-focussed in its own reflector.

Mercury and HPS lamps do not have any halogen cycle.  It is possible to dose a mercury lamp with mercuric bromde and get a perfect halogen cycle operating (this is in fact done in UVC mercury lamps like Philips HOK / Sylvania Germicidal) but the presence of halogens, which are extremely electronegative, means that the lamp is difficult to start.  It’s also not possible to have emitter coating in the electrodes because that would react with the halides, so the lamps are doubly difficult to ignite.  This is why an ignitor is needed for metal halide lamps.  Mercury lamps have been popular largely because of their simple control gear requirements, so setting up a halogen cycle in them would raise system cost undesirably.

HPS lamps cannot be made with a halogen cycle in their standard format.  The electrical feedthroughs are made of Niobium metal.  The halides react extremely vigorously with niobium and a halide-filled lamp would fail rapidly.  The electron emissive coatings on the electrode are also not compatible with metal halides.  This was a problem which had to be overcome to realise the ceramic metal halide lamp – basically to come up with an alternative ceramic-to-metal seal in which no niobium parts are exposed to the metal halides.

I hope this helps answer some of your questions on this interesting topic!