Finished analyzing yesterday a small (relatively) new BLV SIGNION HIE nw (neutral-white) duro 70-100-150 W MH I bought 2 years ago from Germany to use as a general lighter for my living room space. It's been on day on/off since then, so it has accumulated approximately 8k hours and the discharge tube is already blackened from overpowering it at ~100Ws (~0.9-1A) (I think the one I've got is meant for 70W use loading).

Posting the two spectra here for visual comparison:




Packaging box writes: "May contain traces of 58Kr and Th", so I thought it would be a new spectrum from the ones I've already recorded. Couldn't detect Kr/Th visually, so I carefully aligned the two spectrograms by shifting and centering around the Hg 436nm and Kr 432nm line and it appears that this indeed does detect some Kr, compared with the spectrogram of a Kr glow lamp.

Its CCT is reported in the BLV engineering manual as 4,200K, but I perceive it as much higher than that, certainly higher than the CCT of the reported Na/Sc American type MH (4,500), whose spectrogram I compared with. It certainly contains Sc/Hg and Na, since I am getting exact line coincidences for these lines in the spectrogram. For its lighting CCT impression, follow item #[50] on page:

https://ingalidakis.com/spectroscope/collection.html

It's quite hard o tell the difference between the desktop 4,200K HPM and the lampstand 4,200K HIE.

However, there does appear to be some difference between the two: The American Na/Sc and the Kr/Th: it's visible in the blue past the Hg line to to violet, where the Kr/Th MH doesn't have strong emissions there (but I am not sure whether this difference is due to insufficient exposure o the Na/Sc or not).

For the spectrogram comparison, see page: https://ingalidakis.com/spectroscope/amici.html#10halidep

The result appears to be confirmed using a model with contents all 5: Hg, Na, Sc, Kr, Th. For details on the ration of the contents, see [5a*] on page:

https://ingalidakis.com/spectroscope/elements.html#profil

which appears to emulate the HIE's spectrum quite well.

Note that if it turns out the HIE contains Kr, then so does the standard American Na/Sc MH.

If anyone has any more info about the contents of this Kr/Th MH, I'd be glad to read it and will go to the credits of my amici page.

Cheers,
Yiannis 

I don't think the Kr or Th content would cause anything observable. To get about 10Bq at the end of lamp life is way more than enough to allow darkness ignition after about a second or so, so for about 10 year life (5 storage, 5 service) and given the 10 year halflife we are talking about 40 atoms of Kr85 in the complete fill to do the job. Of course, in reality there is more of it, but still it is too little to be visible in the spectrum.

I don't think the Kr or Th content would cause anything observable.

Any proof for the above claim? As it stands, this is just a refutation claim of my showing (at least) one Kr line @432nm. You didn't bother to look at the page, so I am posting the comparisons as a standalone:

To get about 10Bq at the end of lamp life is way more than enough to allow darkness ignition after about a second or so, so for about 10 year life (5 storage, 5 service) and given the 10 year halflife we are talking about 40 atoms of Kr85 in the complete fill to do the job. Of course, in reality there is more of it, but still it is too little to be visible in the spectrum.

I have absolutely no idea what you are talking about. Perhaps if you split the above paragraph into understandable sentences using standard units and words.

Cheers,
UVIR 

He’s talking about the becquerel, the SI unit for radioactivity. What he’s saying is that there shouldn’t be sufficient Kr-85 to see it’s spectra, since it is added into the fill gas merely to provide just enough radioactivity to ensure the lamp strikes in total darkness.

He’s talking about the becquerel, the SI unit for radioactivity. What he’s saying is that there shouldn’t be sufficient Kr-85 to see it’s spectra, since it is added into the fill gas merely to provide just enough radioactivity to ensure the lamp strikes in total darkness.

Thanks for replying. I wasn't aware of this practice (using radio-isotopes in HIDs). I presume "strikes in total darkness" has something to do with allowing for a lower starting voltage when powered on, because of some default ionization from the 85Kr radioactivity? Why is this practice needed, anyway? This, like the American Na/Sc MH either use a CWA which supplies a high voltage spike once a cycle to start it or a thyristor ignitor which supplies a pulse of approximately 2KV per cycle. What's the benefit of using a radio-isotope in the gas fill, apart from providing some default pre-ionization to lower the striking voltage?

The spectrum of this indicates an almost total coincidence with the Na/Sc American tech, so I presume the latter also includes some radio-isotopes in the gas fill(- the SYLVANIA MetalArc type)?

As far as the calculations for the amount of Kr in the tube, I have no reason to doubt the analysis, but I still get the Kr 432nm line visual (main resonance line of Kr, I think), so this may be an indicator of the presence of some significant amount of Kr - perhaps inert, not 85 in the discharge tube. Sc I has a weak line there, but it's 4/30 in terms of the strongest Sc line intensity in its spectrum, so I am not sure the line that shows at 432nm is from Sc. Th has two fairly strong lines there, so perhaps it comes from that. But then again, I'd suspect that Th would be minute traces like Kr, so I can't tell overall. Absent Kr/Th, there's nothing there, so where does this visible come from (coincidence at 432nm in comparison graphs)?

Cheerio,
UVIR 

CWA does not supply any high voltage (by itself, without any additional ignitor), it is just something around 300Vrms OCV (so 500V peak; exact voltage depends on the lamp type and application).

The problem with discharges in darkness is as follows:
When you want a discharge, you need to accelerate a free electron enough to kick out another electron from an atom, once the first electron hits it. Then you need this to continue so long, it generates enough free electron density to carry the ballast current. But itisnot always when an electron hits an atom, that it releases another free electron, sometimes it just excites some of them (and these then release a photon, once that electron falls back into an unexcited state). And sometimes that electron meets a ion lacking its electron, so get captured.
For this you need three things:
- Voltage high enough, so in average each accelerated electron kicks in average more than one electron over its free life. Higher the voltage, faster the free electrons populate so their density builds up (aka the gas ionises into plasma; but it is the free electrons, what this is all about).
- Time for the free electrons have to populate (the duration of the voltage causing the electrons to populate).
- the first (few) free electrons, to have something that could get accelerated. More of these initial electrons, shorter time it takes to build the required conductivity.

And the last is, what all the light dependence is about:
- You may pull that electron from the electrode by just a strong field. But you need VERY strong field, so VERY high voltage for that.
- Or you get some photons hitting the cathode and release an electron. So the photo emission from an ambient light is then the source of these electrons. Some HID designs feature a small auxiliary glow corona discharge tube to generate such starting photons. The auxiliary tube is not tzat current loaded, so may use some other way easier to ionize fill.
- Or you use a small amount of radioactive substance, prefferably a beta emitter (fast electrons - already electrons at speed sufficient to kick others out from their atoms; plus very easy to shield from the environment). Here practically a single electron from the decay during the ignition attempt is enough to start the electron multiplication avalanche, in real life having few dozens help to make the starting more reliable (some avalanches just tend to die due to unfavorable chance, it is a random process).
The thing with any radioisotope is, it disappears over time, regardless if the product is used or not. So when choosing the amount (described by its activity, so how often a decay event occurs in average; 1Bq is one event in 1 second), you have to take into account how much of it will be still present at the moment of the last ignition the lamp is supposed to still work. So when we want a lamp to ignite within one second, we need at least one beta electron to besucessful in that one second. Giving the chances (the lifetime of the free electrons, that coinciding with the peak voltage sufficient to breed them,...), you need some dozen of decays to happen over that second, so the radioactive primer is supposed to have activity those few dozens of Bq at the end of the life of a lamp, that has been in storage before as well. So knowing the substance halflife, you may calculate how much activity you need to start with (assuming you need 10Bq for a reliable start, with a substance that has 10 years of halflife, you are expecting 5 years of storage before use and then 5 years of service, that all gives yo the need to produce the lamp with 20Bq of that radioactive substance).

Now if that radioactive substance were 100% (so only the desired isotopes), it yields few billions (my example about 6.3 billions) atoms, which is not that much, compare to all the other fill components, so even when they will contribute to the lamp radiation (Kr85 behaves just like any other Kr until it decays), the contribution is so small I doubt it would be visible.

But what is true, the radioisotopes not always come as 100% clean (as the activity density would be too much to handle in a reasonable way during lamp production), but are dilluted among other materials, mainly other isotopes of the same element (to make an easy to handle homogenous matter) and some decay products. That means the total amount which has to be used may become so much it then becomes visible in the spectrum. But what is vissible is mainly the auxiliary dilluting material, not the radioisotope itself (due to their nucli being of slightly different mass, affecting the balances intheir electron shells, their specteal lines are in some cases shifted a bit).

CWA does not supply any high voltage (by itself, without any additional ignitor), it is just something around 300Vrms OCV (so 500V peak; exact voltage depends on the lamp type and application).
...
Some HID designs feature a small auxiliary glow corona discharge tube to generate such starting photons. The auxiliary tube is not tzat current loaded, so may use some other way easier to ionize fill.

Hi, many thanks for the detailed answer. The ones I was familiar with, older type Na/Sc MHs - mostly of 1990-2010 American technology, do indeed have an auxiliary starting electrode used in the same manner as in HPM lamps. Most I've seen did not require an external thyristor ignitor in a CWA ballast. I have a SYLVANIA 1000 BD MetalArc that uses a CWA, has an auxiliary starting electrode and lights up fine without an ignitor. A smaller 315W WattMiser Metalarc (Sodium) retrofit from Sylvania has an auxiliary starting electrode, but has problems starting in LPF non-CWA MH circuits without an ignitor. I guess I've missed on all the improvements since the auxiliary electrode thing was dropped with newer tech MHs.

The problem with discharges in darkness is as follows:
When you want a discharge, ...

But what is true, the radioisotopes not always come as 100% clean (as the activity density would be too much to handle in a reasonable way during lamp production), but are dilluted among other materials, mainly other isotopes of the same element (to make an easy to handle homogenous matter) and some decay products. That means the total amount which has to be used may become so much it then becomes visible in the spectrum. But what is vissible is mainly the auxiliary dilluting material, not the radioisotope itself (due to their nucli being of slightly different mass, affecting the balances intheir electron shells, their specteal lines are in some cases shifted a bit).

Thanks. So it would appear, indeed. That this line in the spectrum (432nm) is from Kr/Th I and not from 85Kr. As for the shift you are talking about, I doubt it would be easily visible in a low resolution spectroscope like this small one (Δλ~17Α). If memory serves, isotope shifts I've seen on the Hg 254nm line are of the order of 0.2-0.5A - of similar order to hyper-fine structure shifts. You'd need a huge R spectroscope to detect that. What shows in the photo is thermal broadening of the Kr/Th I 432nm line, instead.

Cheers,

Yiannis 

The auxiliary corona glow discharge tube was/is used only on pulse start lamps, it needs high dV/dt pulse toemit the priming UV flash. In fact these auxiliary discharge vials is often the way to avoid the need for the radioactive aid. Handling radioactive substances is very expensive, so it is a way to save production cost...

The probe start lamps work more on the strong field (the aux probe isvery close to the main electrode), relying on the mixture of ambient light/cosmic radiation/ decay of the thorium sometimes used as tungsten dopant in the electrode material, the resulting electron then being multoplied by a huge factor by that field being present for a long time (normal mains sinewave top, not just the ignitor spike).

In fact these auxiliary discharge vials is often the way to avoid the need for the radioactive aid. Handling radioactive substances is very expensive, so it is a way to save production cost...

Except that it looks like MH tech is advancing backwards: The 58Kr/Th HIE and most probe-less MHs are more recent, while auxiliary probe tech is at least 30 years old. If one wanted to save on production costs, one would go with auxiliary probe starting, not with radioactivity, so it doesn't make much sense timewise.

Cheers,
Yiannis 

Not really. Probe starting was abandoned not because of cost, but because all the electrodes and their lead in seals could not fit the small size arctubes. Plus all the lead seals form a heat bridge, carrying the heat away from the area which is already marginal.
In other words the more compact "two main electrodes only" arctubes allowed way more efficient lamps and that was all the reason behing the starting probes becoming a thing of the past.
It had nothing to do with the lamp cost, it was pure performance driven. But it had to wait till a suitable ignitor technology become available (semiconductors become reliable and cost competitive enough to be used in lighting; the HPS development took lead in this).
The drivers were European markets, where anything requiring more than 220V OCV means cost and efficiency penalty of transformer, instead of just a series choke ballast.
There the reasoning was:
A probe start need more expensive and lossy ballast, or an igniter to rise the voltage. The igniter idea won.
Now when we need the igniter anyways, why the starting probe, which poses only problems otherwise (too big and bulky seal for two lead wires means heat losses, so lower efficacy; the probe needs a resistor and a shorting bimetal switch, which are a reliability burden), when there is then already developed HPS ignitor, just boost its peak voltage a bit. So comes the idea to design the MH so, they could share the same gear with HPS (later was found out, the MH failure modes need a thermal cutout in the ballast or they may overheat it and set on fire, but nothing prevents to use such protected ballast for both MH and HPS).
In order to improve the starting reliability, they used the radioactivity aid trick already used with glow discharges (practically all glow discharge based devices, obviously except the radiation detectors like GM counter or so, or some low noise RF power switches or surge arresters).
Only in very recent decades, when the safety/security regulations regarding the use and handling of radioactive substances tightened so made them more expensive, lamp makers resorted to other tricks, but generally the reliability was hurt in the process.

By the way the probe start used radioactive aid as well: The thorium doping of tungsten in the electroees - all thorium isotopes are somewhat radioactive...

Not really. Probe starting was abandoned not because of cost, but because all the electrodes and their lead in seals could not fit the small size arctubes. Plus all the lead seals form a heat bridge, carrying the heat away from the area which is already marginal....

Cool. Thanks for that info.

By the way the probe start used radioactive aid as well: The thorium doping of tungsten in the electroees - all thorium isotopes are somewhat radioactive...

Yes, it would seem so, since the spectra of the two kinds (Na/Sc-Kr/Th) above, coincide in the 432nm area and that area feature doesn't come from Sc. Whatever's responsible for this in the HIE, is also present in the Sylvania Na/Sc MetalArc.

Btw, do you know if Cs is also used as a radioactive dopand for the daylight /D type (~6,000K) MHs? It is stated in the contents, but after some old investigations, I suspect it does not participate in light production, so maybe it's there for similar reasons to 85Kr/Th in Na/Sc MHs? Thanks.

Cheers,
Yiannis 

I don't know about the Cs exactly, but there are way more reasons, why certain element finsd its way into the fill mix:

There are obviously the luminous components, atoms that tend to radiate in the desired spectrum when excited.

Then the plasma buffer, the atoms tgat are responsible for becoming ions and so generate the conductive plasma, supplying all the free electrons. Not necessary the same with above, as for good efficacy, with the radiating atom you want its electron to be excited, not kicked away. And vice versa, on the ionizing agent you want its electron to be rather kicked free than just being excitedand radiate outside of the desired wavelength range.

Then arc  - substances forming colder plasma layers, forming "jackets" holding the main arc stable in its desired position. Many components tend to form a thin arc, which then tend to swirl around, yielding lamp flicker and other unwanrted effects.

Then there are the starting aid - component that lower the ignition voltage (noble gas penning mixtures,...) or the (radioactive) free electron generators

Then there are evaporation aid and chemical neutralizers. Elements preventing the other fill components from attacking the arctube structures (halogens, readily forming inert salts with all the predominantly metals which would tend to attack the quartz otherwise) and as well forming salts that are way easier to melt and evaporate than the actife elements in their bare metal form.

Then there are residues from the lamp manufacture. These are materials used to allow the bulb assembly, as well as the fill dosage and even handling and storage, which could not be removed before the lamp is complete.

Many elements serve multiple functions, many only one, many elements act partially in some of the other job (that is, why we see spectral traces of the elements whose function is other than generate light),...
And Ihave no detailed overwiew which element is used for what, beside the real basics.

Then the plasma buffer, the atoms tgat are responsible for becoming ions and so generate the conductive plasma, supplying all the free electrons. Not necessary the same with above, as for good efficacy, with the radiating atom you want its electron to be excited, not kicked away. And vice versa, on the ionizing agent you want its electron to be rather kicked free than just being excitedand radiate outside of the desired wavelength range.

Is there really any appreciable ionization taking place in these MHs? I recently got a quote for a T~5,300K and T~4,500K for the arcs of HPM and MH arc centers from D.L Klipstein and these don't seem to be enough for ionization. I am getting much higher T's for substantial ion presence for some of these elements, via E(eV)=5-7kT. For Hg this gives for example a T~23,000K minimum for appreciable ionization and that's way off 5,300K. Do you happen to know what's getting ionized in these MHs?

Is there really any appreciable ionization taking place in these MHs? I recently got a quote for a T~5,300K and T~4,500K for the arcs of HPM and MH arc centers from D.L Klipstein and these don't seem to be enough for ionization. I am getting much higher T's for substantial ion presence for some of these elements, via E(eV)=5-7kT. For Hg this gives for example a T~23,000K minimum for appreciable ionization and that's way off 5,300K. Do you happen to know what's getting ionized in these MHs?

Of course, there is strong ionization, without that there would be no free moving charge particles capable to carry any current.
Yes, by solely heating the matter to high enough temperature, the electron would fly out from the atom on their own. But that is not the only way to get ionization and definitely not the way any ionization happens in the discharges, the temperatures (unless you consider the accelerated electron speed as their "temperature") are not that high.
All the ionization is by the accelerated electron collisions, some by ionizing radiation (the radioactive discharge ignition aid).

Some literature describe the situation using the term "temperature", but in that context the electron acceleration means these electrons could become heated to 100k+Kelvins (dont take that number exactly, read it as "hot enough to cause ionization"). In that terminology the ions have the temperature just about 5kK, but the electrons are orders of magnitude hotter.
It is sometimes called "cold plasma", so a situation where the temperature of ions and electrons is not equal, so where electrons are order of magnitude hotter than the ions.
I dont like this terminology style with discharges, because it just adds extra layer of unit conversions. But it may become handy with hot nuclear physics, when tze difference is not that big (in Tokamak,...), where it inherently sums up all the contributors (random plus systematic component in electron velocities) to the ionization equations.

But with just normal discharges, the ion temperature is too low to contribute, so all ionization energies come from electron acceleration by the field and there directly the electron energy (in eV) is more convenient wayto express it (you need a potential difference of xx V over the mean free path to accelerate an electron to the xx eV energy in order to get some collision action going...).