Needing more than a spark test?

[homebrewed]

I have some additional information to share. I found a source for new scintillator crystals here. The price seems "reasonable" for non-eBay stuff, and you don't have to ask for a quote to get the price! The link shows listings for a different scintillator, but that's because of stuff I learned while browsing the Theremino web site. They have a document showing the use of a SiPM coupled to a LYSO(Ce) scintillator The nice thing about LYSO is that it isn't hygroscopic at all.

And, speaking of Theremino, they have a fairly detailed explanation of their approach to squeezing as much resolution out of their system as possible. If you believe them, you want to get the detector pulse into the digital domain as soon as you can. Their comments discouraging the use of analog pulse-capturing H/W are pretty compelling. They use an interpolation scheme to find the signal peak, and also algorithms to detect "bad" pulses coming out of the detector. They use a 2-pole passive RC lowpass filter to slow down the pulses coming from the detector, and some other signal conditioning circuitry as well. This also improves the SNR, which, according to them, is a Good Thing.

Unfortunately, their English documents leave a lot to be desired as far as clarity goes. Mostly because either they used some sort of translation program, or the writer (clearly) is not a native English speaker/writer. For instance, I think they are using "pictures" in some cases where we would use "pixels". But I could be wrong....

 

[graham-xrf]

I guess the prices are sort of reasonable for what it is. If put with a PMT tube, before any other kit is added, we already have two higher-priced items.

If the SiPM detector is cheaper than a PMT, that can help. What is the approximate cost of a suitable SiPM, and is the front window a limited capture area? We have a ring of Americium sources firing away. If the sensor is too tiny, most reflections might miss.

 

[homebrewed]

An evaluation board for a MICRO-SMTPA-30035 3mmx3mm SiPm is $68.75 in unit quantities from DigiKey. You also would need a ~25V power supply, but of course you'd need a PSU for a PMT as well. The Theremino document on using a SiPM describes the use of a 3x3. A 6x6 mm SiPM is about $40 more. Buying the detector chip and putting it on your own board really won't save you much, so I wouldn't take that approach unless I was going to integrate it with other components.

The crystals that are most compatible with the SiPM's are sold in multiples of either 10 or 20. Looking at (say) the BGO 3x3x15 crystals, that comes to $2.30 each, if one were to find 19 other interested folks. The LYSO crystals in the same form factor would be a hair under $10 apiece, not counting shipping/handling. Unfortunately, none of their CsI(Tl) crystals are a good size for these photodetectors. I think one could buy the 10x10x30 crystals and cut them in half, but there's more money up-front for that and the added risk of damaging the crystal.....plus, CsI is slightly hygroscopic.

The Theremino article does have some words on making a crystal+detector assembly. That could be used as a starting point. Just keep in mind that wrapping it in light-shielding material (like aluminum foil) will absorb lower-energy X-Rays. Putting the whole thing in a light-tight box is the best way to go, as far as not limiting what the scintillator can "see".

 

[RJSakowski]

I think it would be possible to focus the light from the scintillator on a small are detector. You need a large area on the crystal in order to capture the x ray photon efficiently but you also need to direct the scintillation photons to the detector. I'm thinking some sort of light pipe will improve efficiency. If the index of refraction of the light pipe is large enough and the wall angles are right you will get total internal reflection.

 

[homebrewed]

The scintillator crystal I bought on eBay came in today. The vendor wasn't sure if it was CsI or BGO, but I was able to figure that out pretty easily. The crystal measured 1.22 x .585 x 2.99 cm and weighed 16 grams. That means the density of the crystal is 7.498g/cm^3, very close to BGO's published density of 7.13g/cm^3. I didn't remove the crystal from the plastic bag so the actual density would be even closer to the published value. CsI's density is 4.51g/cm^3 so there's no chance the crystal I got is CsI.

Fortunately the ends are polished, so it won't be too difficult to couple it to a SiPM. I'm not crazy about the idea of gluing the crystal to the detector, because I want to leave the option of using a different type of scintillator on the table as long as possible. I'm thinking the oil used for oil-immersion microscope lenses might work. It is chosen for its transparency (of course), and a good match to the index of refraction of the front glass lens. I'll need to make some sort of mount to hold the crystal and detector together, though, while ensuring it doesn't block x-rays.

 

[graham-xrf]

Fortunately the ends are polished, so it won't be too difficult to couple it to a SiPM. I'm not crazy about the idea of gluing the crystal to the detector, because I want to leave the option of using a different type of scintillator on the table as long as possible. I'm thinking the oil used for oil-immersion microscope lenses might work. It is chosen for its transparency (of course), and a good match to the index of refraction of the front glass lens. I'll need to make some sort of mount to hold the crystal and detector together, though, while ensuring it doesn't block x-rays.

Use a drop of this stuff. That bottle is about 20 years old. You don't need much. It goes a long way.



From your dimensions, you have a rectangular slab of crystal, and I think you only need some of it.
It is not hygroscopic, and can apparently "be machined". It does not cleave, so it may be a bit like glass. It may have some radioactivity of it's own. Does it then "self-scintillate"? The Silicon diode saw a high temperature in it's manufacture. It is not likely to come to harm if glued up with the type of hot-melt used for paper. It peels up and can be removed if you pull on it.

I speculate a procedure here, but based on experience of coupling coherent fibre-optic tapers to CCDs.

You can likely cut it slowly the same way as glass, using a mudsaw, for which read, hacksaw + abrasive grit + water + a little sink dishwasher squirt like "Fairy Liquid". Amazingly, the hacksay will be OK for cutting metal afterward.

Any old wet-type tile cutter is also likely to work. You can use a aluminium disc fro an old hard drive. It needs to turn on an arbor with some kind of variable speed motor. Difficult to do with a hand held electric drill. You need something steady, and you can gently control the pressure with the fingers.

You can also "make it round", by cutting through with a hole saw, but using abrasive.

The taper light pipe as suggested by @RJSakowski might be made of acrylic or polycarbonate. You can polish that stuff easily, and it is very transparent. You can get it as rod, or just raid it from a cheap drawer handle.
--> Something like this LINK

You are way ahead of me on this. I have yet to read and understand all the Theremino info. My Arduino experience was only one A/D datalogger. I have more experience with Raspberry Pi, but that does not extend to deep Python programming. My EDA program for schematics and PCB design has languished for too long, and needs the update.
I have some urgencies at home but will keep following the thread. I hope my PMT will arrive soon.

[Edit: Oops - when you said the ends were polished, I think you meant the 1.22 x 0.585 small ends. I was imagining cutting a piece off, and shining through the 0.585 dimension. It need not stop you. You can polish the "sides" before you cut.]

 

[homebrewed]

I have a low-speed diamond saw setup, made from a stepper motor and home-made arbor, and a lapidary-class low concentration saw blade. I've used it to nick HSS rods to break them into shorter lengths so I should be able to cut the crystal down more if I want. It appears that BGO is about as hard as glass, so there should be plenty of polishing media that will work to polish it.

I had thought it might not be necessary to polish the BGO if the immersion oil has a similar index of refraction but BGO has a pretty high index (2.15). Immersion oil is around 1.5 so a rough surface won't completely "disappear" if it's covered by the oil. Oh, well.

BTW, LYSO has an index of refraction of 1.8 so it might be OK for mounting up a crystal that has a less-than-perfect surface polish.

 

[graham-xrf]

When coupling two (flat) transparent materials of different refractive index n1, n2, using a drop of immersion oil with index n_oil, what should the oil index be?

The major use, to couple the space between a special oil-only microscope objective, and a cover glass to eliminate two air-glass surfaces. A tricky, close-up procedure, only possible with specimens a few microns thick.

In this case, the crystal-glass join is flat, surfaces closer than a quarter wavelength. It is much more like cementing achromatic doublets, or making nicol prisms joining glass to calcite crystal. We only have to fill the surface roughness at the join.

When n1 and n2 are different, I would hazard a guess, and seek a coupling medium sqrt(n1*n2), but that is difficult, and I am not sure it is necessary. This is qa bit like the treatment given to emeralds and rubies to "enhance" them by filling micro fractures in the polished surfaces. They get up to all sorts of procedures to get the stuff into the gems, but unless you are OK to cement the scintillator to the SiPM, I suggest just use a drop of oil, and mount the surfaces together, The oil will not dry, nor fall out.

 

[graham-xrf]

Just a passing thought..
is it OK to be sharing a desk with radioactive items where 8 button sources of Americium 241 + 2 radioactive thorium-loaded lamp mantles compete for space with the coffee mug?

OK - now I get it why the crystals have to be so thick in the direction axis the incoming photon passes.


It's all about the energy range. I guess a high enough energy incoming can pass right through a thinner crystal without getting stopped, and having to deliver it's glow. That said, the range of eV energies we need for XRF might suit thinner crystals. Also, Area matters.

I am soaking up YT. This fellow uses a little Australian made voltage supply for the PMT, and is plugging it madly.
Also - using the sound card as the signal capture mechanism.
He uses the expression "pulse shape analyze".
He mentions "pulses being accepted". There is some discrimination there.

Y-Axis is counts.
X-Axis is "energies".
But - quite instructive (I cannot bring myself to use the word "illuminating")!
--> Inexpensive Gamma Spectroscopy!!!

Take a look at the options in data acquisition menu to the right.

He may be looking more at sources for isotope detection. We want to look at scatter that souces have worked on.

 

[homebrewed]

If I' correctly understood the Theremino description of their approach using a sound card, they also are using pulse shape analysis to: 1. Accurately determine the peak height of a pulse; 2. Detect and reject instances of so-called pileup, where a second photon arrives before the first one decays (so the pulse height of the second one is offset by the first); and (3) reject pulses that are contaminated by system noise.

I'm not convinced that point 1 can't be achieved using a properly-designed analog track-hold circuit but the other functions would be difficult to implement using an analog-only system.

BTW, I don't think that a simple deconvolution can handle the case where pileup occurs. There's a difference between photon energy overlap and simple pulse overlap. The first situation can be addressed (to some extent) by assuming a particular system resolution and using that to de-convolve the data. In the second you have a variable time relationship for two pulses that overlap in time. Only S/W could "look" at the data and determine if any part of the pulse(s) can be used to generate a good spectrum.