Needing more than a spark test?

OK - it appears that the "X" in "XRF" on the name does refer to X-Rays,a and many detectors do use a small X-Ray tube to make a pulse of X-Rays, very confined, and not going through the user.

ALSO - it seems, the thing works with stuff other than X-Rays, like Alpha particle emitters, and electron beams.
We discount electron beams, because you have to do the trick in a vacuum.

Here is where the smoke detectors come in. Americium 241 is an Alpha-particle emitter with a half-life of 432.2 years. The helium nuclei being emitted are at a level far below the background level, which is presumably why the levels measured when folk try going at smoke detectors with radiation detectors get much the same count no matter how far they strip it down.
--> LIKE THIS

The stuff is feeble, failing to penetrate a piece of paper. It shows on radiograph film from up against after 2 hours exposure. You can handle it. Just don't eat it!

We live on a naturally radioactive planet. The potassium in bananas is radioactive, and the sand from Chesil Beach in Dorset, as --> THIS Radioactive Beach in Brazil. Maybe nearly every shingle beach anywhere, but we are talking lots more than from a smoke detector.

Can you amass something harmful? Yes - but you have to try hard. Mixed with Am241 from about 100 smoke alarms + radium from old clocks + thorium from gas lamp mantles, purified with lithium from batteries, and a block of lead with a hole drilled part-way through.. sure, it becomes a breeder reactor which will make a mess.
ref --> David Hahn (no longer with us)
--> David Hahn Wiki
--> The Radioactive Boy Scout

Eight or ten or a hundred smoke alarms are not going to do anything nasty. There are a good deal more in most public buildings. You can expect there will be a portion of the population for whom anything beginning with "nucl" will rub on their beliefs kinda hard, but I try always to stay informed by the science.

The PhysicsOpenLab LINK DIY XRF provided by @homebrewed was really useful. (My thanks for that)

This is the sort of thing one can apply (say) a Raspberry Pi or Arduino to, and turn into a relatively affordable project, and I am sure there is enough expertise in HM to draw on. In a nice practical form, it could require some parts made in the shop, and at the end of it, you can know with some certainty which heat treat regime to use for the thing in the making. Know how hard will be the knife edge, of the gun barrel. Know from the get-go whether the steel you are using for whatever will resist corrosion. Know whether the bronze for the bearing is something other than a boat ornament.

Now that I look at it harder, it seems a safe an viable project that need not strain the finances and looks to be something that an HM member can do - and then post exactly how with plans and instructions in 3-part harmony!

It seems the "Theremino" part is a generalized USB interface to a PIC microcontroller.

MasterDIL-V5_3D_Top.jpg

MasterDIL-V5_SCH-1024x727.jpg

Perhaps this board could be side-stepped for not many dollars by using a $50 Raspberry Pi, and that is like an Arduino on steroids.! A full-blown computer (Pi-4) with all the operating system, USB, video, and GPIO connect pins one could want. You can even watch YT while you analyze steel.

I admit, I am tempted, but there is too much to get right on my house, my workshop man-cave in construction, and my two machines, to go at this one just yet, though I might just keep some of the parts I have around a bit longer.

- - - - - - - - - - -
P.S. Please note that there is no "XRF" allusion to anything related to my user name graham-xrf, where the "x" denotes "ex" like in "ex-girlfriend" and similar.
 
Last edited:
OK, here's my take. The americium 241 is an x ray source. The x-rays from the source impinge upon the sample and electrons from the inner shells of the contained elements are knocked up to a higher shell. When they fall back to the lower shell from whence they came, they emit a photon (photoelectric effect) . The energies invoked are such that the photons are in the x-ray part of the spectrum. This is the fluorescence part.
Those photons hit the scintillation crystal and lower energy visible light photons are generated. The number of secondary photons is proportional the the energy of the original x-ray photon. Some percentage of those photons impinge upon the phoromultiplier which generates an electric current proportional to the impinging photons and thus proportional to the energy of the fluorescent x-ray. An electronic circuit measures the current for that original photoelectric event and records an event in a particular channel according to its energy. There are multiple channels, each of a different energy, and those are scanned after observing a sufficient number of events to generate the x-ray spectrum.

So, required components are: x-ray source, scintillation crystal, photomultiplier, and multichannel analyzer plus computer and software to display the spectrum.

When I was in college fifty years ago, we had a newly acquired multichannel analyzerin our nuclear physics lab. It had only about fifty channels and the readout was by columns of neon lamps. Digital calculators were almost a decade away and computers were mainframe affairs. I recall that we did some rudimentary analysis of gamma radiation with it.

Fast forward to 1997, I worked for a medical device manufacturer and we made products for calibration of diagnostic x-ray machines amongst others. Along that line, we made ionization chambers which measured x-ray energy by passing the x-rays through a chamber comprised of two parallel plates with a voltage applied to them. An x-ray photon passing through the chamber would eject electrons fro the contained gas, some of which would drift to the positively charge plate, generating a small electric current. This current was then amplified to generate a signal proportional to the energy of the x-ray beam.

I don't know if the signal from the ion chamber would be sufficient to feed a multichannel analyzer but it could, in theory replace the scintillation crystal/photomultiplier. If you think about it, a modern digital camera functions in a similar fashion arranging the "bins" spatially instead of by energy levels.
 
The other approach for doing elemental analysis is EDX, energy dispersive x-ray spectroscopy. In this method an electron beam is accelerated to 5-30KV and focused on the sample. This basically replaces the x-ray excitation with e-beam. From there, a pulse height detector is used to generate the x ray spectrum, very much like the XRF system. The equipment is (naturally) expensive, so why would you go that route instead of XRF? Simple--spatial resolution. Your typical EDX tool is installed on a scanning electron microscope, which has very high resolution. Not needed to analyze a hunk of unknown alloy.

Back down to earth, I started wondering if it might be possible to update the old spark test by aiming a home-brew spectrometer at the trail of sparks and see what spectral lines show up. Here is a web site that sells inexpensive spectrometer kits.

Iron and carbon would burn and generate light but it's not clear to me if nickel, chrome, vanadium, molybdenum etc. would.

Again, even if the "new age spark test" scheme is able to detect the different metals in your alloy, it would be a challenge to get quantitative results. Still, it might be good enough to aim you in the right direction -- and if augmented with other data (magnetic/nonmagnetic etc) you MIGHT be able to narrow down the field of possibles even more.

An interesting discussion!
 
The other approach for doing elemental analysis is EDX, energy dispersive x-ray spectroscopy. In this method an electron beam is accelerated to 5-30KV and focused on the sample. This basiHe
2+
cally replaces the x-ray excitation with e-beam. From there, a pulse height detector is used to generate the x ray spectrum, very much like the XRF system. The equipment is (naturally) expensive, so why would you go that route instead of XRF? Simple--spatial resolution. Your typical EDX tool is installed on a scanning electron microscope, which has very high resolution. Not needed to analyze a hunk of unknown alloy.

Back down to earth, I started wondering if it might be possible to update the old spark test by aiming a home-brew spectrometer at the trail of sparks and see what spectral lines show up. Here is a web site that sells inexpensive spectrometer kits.

Iron and carbon would burn and generate light but it's not clear to me if nickel, chrome, vanadium, molybdenum etc. would.

Again, even if the "new age spark test" scheme is able to detect the different metals in your alloy, it would be a challenge to get quantitative results. Still, it might be good enough to aim you in the right direction -- and if augmented with other data (magnetic/nonmagnetic etc) you MIGHT be able to narrow down the field of possibles even more.

An interesting discussion!
A spark is a short-duration plasma
Suggesting one does it with a plasma instead of a spark- almost the equivalent of the Bunsen burner flame test, but much hotter, and with a wavelength detector (usually diffraction grating) to let you see more "colours".
OK - let us pick out the possibles.

To resolve first the X-Ray thing. Yes indeed, you can excite the metal atoms with X-Rays, but you can also use other primary excitation sources.

1) β- radiation is, as I understand it, an electron beam. It will do. β+ is positrons. We won't do that!

2) Protons. The stuff left over when we ionize hydrogen to get electrons. Hmm.. That is the stuff they use in the LHC. The hydrogen they use for many weeks supply is the size of a in-car fire extinguisher.

3) α Alpha particles. These are helium nuclei. 2 protons + 2 neutrons. Among other things, they come from smoke detectors. Written as He2+ or He2+ with energy 5MeV, and they move fast - like about 15,000 km/sec.

I can do, and have done X-Rays. Expensive, awkward to use, risky, needing high energy evacuated kit.

I once used a hospital X-Ray machine to put a beam through a 1mm thick little square of copper which I used as a neutral density filter to cut down the brightness of the beam, so I could develop an auto-focus video device for the photomultiplier display to allow continuous X-Ray diagnosis. The only way I could get my face close enough to the eyepiece was to sit on the machine, and let the beam go down onto the copper quite close to where I was ..er. sitting. I had the heavy white apron and all, but the beam was to the side of me, and I had it switched on for 20 seconds and longer at a time. X-Rays is not what we (OK - I) would want to use.

Electron beam guns is also something I have got to grips with. The up to 40kW kind where the beam is focused onto the top of the (titanium) being vaporized, using scan coil magnets to move it around. Huge vacuum pumps, and 35kV power supplies (variable), running up to near 1Amp. All is OK unless you lose control of the beam. If it ever hit the side of the chamber and broke through, that would be a huge dose of X-Ray.

I came to dislike electron beams. I still have two of the old Leybold Hereaus electron guns intact with cooling jackets and a high-isolation filament transformer. They are nice big chunks of beautifully shaped stainless steel that can be re-purposed. Alternatively, if any HM member wants them, even for a door stop or boat anchor, they are there for the taking, but you have to fund the shipping - and remember, I am in the UK! We can throw in the safety isolation filament transformer and a length of molybdenum filament wire. This stuff is a bit heavy.

We don't need X-rays, nor electron beams.
Americium is an Alpha (α) source, and a weak gamma (γ) source that is unknown in nature, the only artificially produced element, made by neutron bombardment of plutonium or uranium, so it is, in fact, a component of reactor waste. The tiny amount needed for a steel analyzer can come from a few smoke detectors.

Can we use the light as @homebrewed suggests?
Perhaps one can, though I think this is still the stuff of a small vacuum chamber. Perhaps of stainless pipe about 50mm diameter, and having quartz windows. A cheap plastic diffraction grating, and one of those "line array", or matrix photodiodes. The carbon arc can be started with a spark, stabilize the light, and the sample being tested is in fact one of the electrodes, or is placed on a piece of carbon. We only need a brief flare - perhaps 400mS.
The photodiodes can be sampled much faster, perhaps at 20 - 100kHz, and the outputs put through a calibration scaling before display. A Raspberry Pi fitted with an e-Bay A-D converter add-on board (about $10 to$40) might do.

So far, for me, this version is just a mad, free thinking, possibly stupid speculation. Like designing on the back of a napkin while the food is arriving! I have no idea whether this high speed, high tech, version of a Bunsen flame test will work. Feel free to trash the musings if you like.

The radioactive way - a proven design exists!
That is why I like it. None of the bits are outrageously expensive. Much of the design outline is already in this thread. It has some nice science in it.

How we get the spectrum plot out of it is something I did not quite understand. I (think) the energy that excites the metal atom is released as the electrons return to their un-hyped-up quantum state, and the wavelength of what is released is related to the wavelength by Planck's Constant, which is extremely er.. constant. So measuring the relative brightness of the scintillation leads to a place on the plot.

We surely do not need a build as clunky as this!

xrfEquipment3.png

The smoke detector set looks like this. Again, I am thinking it need not be like this.

xrfEquipment2.png

americiumSpectrum.png

OK - not fully checked out yet, but this last option looks so very feasible that it is surely worth a scrutinize by HM Forum crowd critique.
 
Last edited:
Hmmmmm ... high voltage power supply for the photomultiplier .... Maybe this one?

-or-

use one of the Rooshin night vision scopes! They include photomultipliers (though an imaging type).

(2) One thing I think I missed in your description of a DIY XRF rig was a monochromator - something to separate the frequencies of the detected radiation, so you get a spectrum, as shown in your illustration. Your most recent post touches on that, but I don't think brightness and wavelength/frequency are interchangeable. RJ's explanation helped a little, but I still don't "get it." My studies in chemistry did not include XRF or EDX. The closest I came was in grad school, where I did ESCA spectroscopy. That entailed analyzing the energy spectrum of electrons emitted when a sample was bombarded by Xrays. Electrons aree a WHOLE lot easier to manipulate than photons!

(3) Good thing you mentioned "P.S. Please note that there is no 'XRF' allusion to anything related to my user name graham-xrf, where the 'x' denotes "ex" like in "ex-girlfriend" and similar." I was really starting to wonder!
 
Last edited:
(2) One thing I think I missed in your description of a DIY XRF rig was a monochromator - something to separate the frequencies of the detected radiation, so you get a spectrum, as shown in your illustration. Your most recent post touches on that, but I don't think brightness and wavelength/frequency are interchangeable. RJ's explanation helped a little, but I still don't "get it." My studies in chemistry did not include XRF or EDX. The closest I came was in grad school, where I did ESCA spectroscopy. That entailed analyzing the energy spectrum of electrons emitted when a sample was bombarded by Xrays. Electrons aree a WHOLE lot easier to manipulate than photons!

(3) Good thing you mentioned "P.S. Please note that there is no 'XRF' allusion to anything related to my user name graham-xrf, where the 'x' denotes "ex" like in "ex-girlfriend" and similar." I was really starting to wonder!
OK - we step carefully here - why we don't need a monochromator. All spectrum analysis of the various frequencies light kind do need it. Diffraction gratings, Fresnel things, anything to bend the light varying amounts depending on the color.

We are finding out the unique wavelength - or frequency if you like, of the light emitted from an element. We are hitting it with radiation energy, albeit not from a Bunsen flame, nor a spark. We use a radioactive energy.

Now here is where we get right into the murky waters of "is it a particle, or is it a wave"? There are now lots of good explanations on the equivalence of matter and energy, and the "Quantum Standard Model". The quantum model is an elaborate scientific fairy story which does not care that the concepts are arbitrary and non-intuitive. What is known is that it allows exact calculation and prediction to better than 10 digits of extraordinary, repeatable precision.

When that "energy", whether thought of as a helium atom with momentum, or simply as a bunch of energy, slams into the metal, the electrons in the atom will only accept strictly discrete amounts - i.e."quanta", and after a short time, from nanoseconds to hours, depending on material, the electrons drop back to their previous state, releasing back exactly that amount of energy.

Now we get to the nice bit. That energy comes out as radiation - photons - it is light. It zips away at the speed of light. And best of all, it has to be a unique color. At this point, you can say "ahh - we can figure what element that came from is we use a monochromator grating, and make it sit at it's place in a spectrum. This could be done if one was madly energising it, like the phosphors in a flourescent tube.

As I understand it, this method is different. The photon that is released hits a scintillator material where the whole excitation thing happens all over again, but this time in a constant, known, material. So the now second-hand light from the scintillator crystal excites a photomultiplier tube, or photodiode, or maybe solid-state photomultiplier. It results in an electric pulse which can reveal meaningful information. The nature of the pulse, the timing, the slopes, and the magnitude of it, are a signature of the metal that started it all.

I am not an expert. I am still reading up on exactly how this trick works. A whole bunch of metals in the mix can apparently produce a whole bunch of little slopes in the response, unique to each, and the magnitude of the departures can indicate relative proportions. The math, and science in this is not trivial, and is completely contrary to the reasonable simple, low cost kit that is involved.

The XRFSpec_ENG.pfd file posted earlier in this thread has the pulse data signatures for nearly every element.

OK on the "xrf" in my username. It was pure coincidence!
 
Last edited:
Maybe a silly question, but if you dissolved the metal in acid, and shine a wideband filament type light through it, and then spread the light with a grating or prism, would you see a "dark line" corresponding to the absorption of light wavelength specific to the stuff in the solution?

Yeah - I know, if it was that easy, it would be widely done already!
 
Alpha particles have to be accelerated to a very high velocity in order to generate x-ray fluorescence in a target. This would require a linear accelerator. In addition, the beam would have used in a high vacuum. The Americium 241 source emits a relatively low energy alpha particles, easily stopped by a few sheets of paper. It is the 59.7Kev gamma ray that is responsible for the fluorescence in target materials. The fluorescence photons re not light. They are much higher in frequency , above UV and in the x-ray region where their "frequency" is measured in terms of energy( e =hf).

To detect these x-rays, they are passed through the scintillation crystal where again the photoelectric process takes place and visible light photons are generated. The sum of the energies of the photons generated is equal to the energy of the x-ray photon.
 
Maybe a silly question, but if you dissolved the metal in acid, and shine a wideband filament type light through it, and then spread the light with a grating or prism, would you see a "dark line" corresponding to the absorption of light wavelength specific to the stuff in the solution?

Yeah - I know, if it was that easy, it would be widely done already!
Actually, you do not see a single line. If you run a spectrum of the light passeing through the solution it is a rather broad absorbance band. Offhand, I can't give you the physics of it but I have run enough of them to know it for a fact. Additionally, the colored solution resulting from absorbance is not dependent on an interaction with the nucleus. Copper chloride is a different color than copper nitrate. Add ammonia, and the solution turns a very intense blue.

It is possible to see absorbance lines (Fraunhofer lines) if the matter is in a gaseous state. This is how the atomic absorption spectrometer works.
 
Back
Top