Photo peak duplication when the source is on the side of the scintillator

[Richard]

Hello again! It was more than two years since I posted on the old forum http://www.gammaspectacular.com/index.p ... t&path=265

I have had some time to play with my GammaSpectacular GS-2812 during the summer. I have worked with a deconvolution routine that takes into account the Compton continuum, backscatter peak and single and double escape peaks. During this work, I discovered that not only the backscatter peak but also the rest of the spectrum is very dependent on the position of the source.

If I put the source on the side of the detector, 10 cm from the front - slightly behind where the scintillator crystal ends and PMT starts - I get very strange spectra compared to the spectra taken with the source in front of the scintillator crystal (see figure above). The photopeaks are duplicated, creating ghost peaks at around 14% lower energy than the original. Also the Compton spectrum is distorted (see the Cs137 spectrum where the Compton continuum has been squeezed in energy). It is like there were two paths with different PMT amplification creating an overlayed ghost spectrum on top of the original spectrum. In the Lu176 spectra one can see that the ghost peak of the weaker 202keV photopeak is stronger than the ghost peak of the 307keV peak, indicating that the effect creating the ghost peaks is stronger at lower energy. The ghost peaks are even stronger than the original peaks for energies below 500keV. In the Co60 spectrum the ghost peak of the 1.33 MeV peak is probably covered by the 1.17MeV peak. The duplication effect is sometimes visible also when you have the source in front of the detector. In the Cs137 front spectrum one can see a small artefact due to the ghost peak in the area between the photo peak and the Compton edge. All spectra include also background radiation but it is weak compared to the signal, except in the Lu176 "side" spectrum where the low energy peak is dominated by a combination of X-rays and background radiation. The energy axis is the same on all graphs but the y-axis is in arbitrary units due to the different count rates.

Does anyone on the forum know or have any idea about the explanation of these ghost peaks? Can you reproduce it?

Since the effect is so obvious I guess that it has been thoroughly investigated and explained even if I couldn´t find it mentioned in my textbooks.

[luuk]

Hi Richard,
I would say there are two options:
1: your crystal is broken, best way to find if it is use a Am-241source hold it against your detector and move it oround your detector while looking at the screen.
You could see a sudden jump in the peak signal from high to low or reversed, if that is the case your crystal is broken`.
2: your crystal is partly wet again use a Am-241 source do the same with the source and try to find the point were the peak is at it lowest point try to calculate the resolution.
Then move it to the point were the peak is at it highest and caclulate the resolution, is there is a big difference(lowest peak is really wide) you have a partly wet crystal.
Why use a Am-241 source, because the radiation will be absorbed in the first 1-2cm so it is more easy to find the bad worst and best point in your detector.
Luuk

[Richard]

Hi luuk,

Thanks for your reply which made me a bit concerned. I have tested with a Am241 source. Yes there is a rather abrupt change of peak energy at around 7 cm from the front. If the source is located 6 cm or closer to the front I get a peak energy of 63.4 keV (the discrepancy to 61 is calibration error) with SD=4.5. If it is located 8 cm or more I get a peak energy of 55.6keV with SD=5.2. Inbetween the peak energy varies rather continuosly with the peak boadened. I can not notice any rotational asymmetry. The crystal is 2.8 inch long which is 7.11cm, almost exactly where the transition occurs. However I don´t know it the crystal is placed exactly at the top.

What speaks against that the crystal is broken or wet is that I kept it dry and handled the probe very carefully. It has also rather good resolution with FWHM around 6.5% at 661keV with PRA and a peak to valley of Co60 between 7-10 depending on the distortion tolerance setting. I don´t really understand why a crack would give double peaks but if it is typical for a damaged crystal and does not occur otherwise, it is still the most likely explanation.

It would be interesting if someone else with a similar probe could test if the effect is reproducible.

[luuk]

Hi Richard,

The sudden jump when you scan the crystal with Am-241 indicates that it is a real serious crack
Looks like you crystal has a horizontal (parallel to the pmt cathode) crack.
If you irradiate e.g. Cs-137 from the side the now two separated parts will produce both light, the light made in the part close to the pmt will transport the light easy to the Cathode.
But the light produced in the other part will bounce on the crack and will bounce a few time but and that cost some light, eventually it will reach the pmt but the lost light will give a second peak lower than the original peak.
If you irradiate from the front most light will be produced(depending on the energy) in the first part or second crystal and not so much in the part close to the pmt so you will see “one” peak, So it looks like there is nothing wrong.
Some crystal can break suddenly jus by a light mechanical or temperature chock or , it happens that the crystal has internal tension it just happens nothing to do about it.
But your detector is build up as far as I know with a separate C style NaI(Tl) couple to a pmt so it should be possible to replace the crystal, but Steve knows all about that.

Luuk

[Richard]

Hi Luuk,

Yes, after searching more information on the internet I have also come to the sad conclusion that you are correct even if your explanation doesn´t explain why I get higher energy (=MORE light) at the front than when the source is closer to the PMT. Maybe it can be explained with a larger crystal volume in front of the crack. Anyway, there is probably a fracture parallel to the PMT that runs through the entire crystal. I haven´t suspected that anything was wrong with the probe since it had good resolution and didn't give any major artefacts when I measured from the front. However, when I now go back and reexamine old data, this explains the mysterious spurious "bumps" that sporadically occured in the measured spectra, e.g. the weak bump between the Cs137 photo peak and Compton edge. Most probably the probe was damaged already in the freight to Sweden.

I have to live with the knowledge that my dear GS-2812 "Red Baron" is not as perfect as I thought and account for that when I analyze the spectra in the future. Even if it is possible in principle, it is too costly for me to replace the crystal and I would not dare to do it myself after seeing videos and pictures how hydroscopic and corrosive NaI is. The end result would not only be a completely ruined gamma probe but also a thallium-poisoned family :)

Anyway, thanks for solving the doublepeak-mystery so quickly!

Richard

[luuk]

Hi Richard,

I am not 100% sure because I have never seen your type detector, but if I remember right the crystal is a C style so that means that it is a completely sealed crystal with on one side a glass window, the pmt is coupled to the glass of the C style with some coupling compound that all together is mounted in an aluminium housing.
So it is possible to remove the C style without the risk of leaking, the crystal is completely sealed.
I only don’t know how Steve glued the housing and if it is possible to open it again and replace it for an other C style crystal.
It is maybe possible to replace it with a cheaper eBay detector one that is not broken, it could be worth a try.
But you better ask Steve what the possibilities are to open the detector.
Luuk

[Sesselmann]

Richard,

I agree with Luuk, the double peaks you are seeing are not typical of NaI detectors, you should not be seeing these.

Your detector is built with a 30x70 mm Russian crystal and an American ADIT PMT (Politically explosive combination which could easily lead to fractures).

The crystal and tube assembly is fitted into the housing surrounded by high density foam, hopefully it will slide out easily. I recommend finding a small hook and pulling the assembly out from the hole in the middle of the divider circuit board (rather than tugging on the cables.)

Remove the black electrical tape and the mu-metal shielding.

WARNING: Never power up the PMT in open light, you will instantly damage the PMT and risk electrocution.

The crystal is completely encapsulated (safe to handle) and glued to the pmt with Selley's "All Clear" glue. This glue does not dry hard, it remains soft and crystal clear. it should be possible to carefully separate the crystal from the PMT by using a pointed tool to remove all the visible glue around the edges, then a thin blade to remove some of the thin film between the pmt and the glass, finally a gentle twist should do it (use a glove in case the PMT should accidentally break)

I don't have any more of these crystals but I believe there are some still on the market, the resolution on these crystals were excellent, some of the best I had.

This adds another dimension to gamma spectrometry for you, best of luck with the surgery,

Steven

[Richard]

Thanks Steven for the instructions how to replace the crystal. However, the cost and time to do this, together with the uncertainty of the outcome, makes me doubtful if it is the correct way for me to take. I saw that there indeed was a crystal of same size for sale on eBay but with import costs and swedish tax, the total cost is comparable with the cost of a good used probe. I had the same contemplation when I decided to put some extra cash to invest in a ready built of guaranteed performance, instead of building one myself. I will try to live with the crack for the moment. The probe has still a good resolution and the fracture causes very small artefacts if the probe is illuminated from the front. If I am doubtful if a peak is real or not, I can make an additional measurement with the source on the side and check if the peak height increases.

The annoying mystery is when and how this crack happened. The probe has really been treated like a baby and been stored safely in the wardrobe in the foamed camera case when not used. It could of course have been a small defect in the crystal that has propagated with time but if I understand correctly, these crystals are old surplus so it is strange that it didn´t occur within a a few years of its fabrication. A possibility is that the internal stress from the assembly has caused an originally small defect to propagate. Another possibility is that the fracture has been there all the time but unnoticed since the artefacts it causes when the probe is illuminated from the front are so small.

Thanks again for all the advices and instructions. My measurements were not all in vain since they give a good example for others how a spectrum from a fractured scintillation crystal can look. :)

[brehwens]

Hi Richard!

Sorry to hear that there is a crack in the crystal...

Steven: is there no solvent that could be used over a few days to soften up the joint in order to avoid damaging a perfectly good PMT, which I believe is the single most expensive component in the detector??

Richard: most of those 30x70mm crystals are sold from the Ukraine or russia, you pay no sales tax. If you cant fid the time to try to open up the probe, let me know and maybe I can give it a try. I have some tools and some non-hardening coupling grease which allows for non-permanent mating of a new crystal to that PMT.

[Steve Dubyk]

Hello folks! This is a topic in the "sad but true" category. I have attached graphs from the defective crystals I have encountered over the years, including one featuring the double peak. What is most unfortunate is that the sealed probe here was the first one I used for gamma spec, and spent many months trying to determine what was wrong. I have never opened that unit so it is unclear what the problem is, but likely a degraded crystal.

Richard, as to repairing your probe, it really is worth the effort. My first operational probe was one I built using a tested CsI crystal provided by Luuk, and it remains one of my best. Good luck!

Steve Dubyk