150KeV peak not U235?

[Mike S]

I recently started running gamma spec again after getting a new shield setup. Among the items I've run recently was a bag of fossil shark teeth that show a spectrum consistent with natural uranium and progeny (as is often found with fossil teeth). While the combination of x-ray peaks and a few gamma peaks in the 75-95 KeV range makes this portion of the spectrum hard to nail down, I'm pretty confident in the peaks at higher energies, except the one at 150 KeV.

Bag Fossil Shark Teeth 1_5x2_25 Bicron 69 hours No IRR No BKG 0-750KeV ThereminoMCA_2022_05_29_11_54_48 (2).jpg

Bag Fossil Shark Teeth 1_5x2_25 Bicron 69 hours No IRR No BKG 0-2500KeV Ylog ThereminoMCA_2022_05_29_11_50_45 (2).jpg

Note that the background is only about 3.7 cps in this shield, so background subtraction didn't have much effect other than removing the K40 and 511 KeV peaks. I figured it was better to provide the plots "raw", but I have other versions if that helps.

Looking at other amature spectra, it seems this peak is common with natural uranium spectra and most people (including myself) will label this as U235 due to its peaks at 144 and 163 KeV, which certainly makes sense and looks the part. However, in trying to confirm this is really U235, I'm finding evidence that may not be correct, but I'm not finding any good alternative explanation documented (though I have a theory).

Please keep in mind that I'm still fairly new to gamma spectroscopy and I'm just trying to understand what I'm seeing; I'm not saying anyone is wrong. With that, here's what I've found:

Here's an example showing how this peak becomes more visible as enrichment increases (but doesn't show with natural uranium).
[broken link removed - Steven]

Low Energy NaI response Natural and Enriched U.JPG

Here's another example with a CdZnTe detector and various enrichments showing those 144/163 KeV peaks don't show up until the enrichment is high enough for the U235 gammas to really come through (note the log scale).
[Found on page 2 of this PDF: http://cztlab.engin.umich.edu/wp-conten ... n_2016.pdf ]

U var enrichments CdZnTe detector.JPG

Of course, we know there are gammas in this range in natural uranium and you can clearly see them in this plot of uranium ore taken with a Ge(Li) detector. But those peaks are barely above the background and are several times lower height than the 186 KeV peak (again, note the log scale); I don't think these would show up on a NaI detector. The same organization has an older catalog of spectra using 3"x3" NaI detector but the poor resolution at this low energy makes it unusable in this discussion since these features down show up.
[Image from page 837 of this PDF: https://gammaray.inl.gov/Shared%20Documents/gecat.pdf ]

U Ore low energy section.JPG

So what is it?
My best theory is it is a coincidence peak from the jumble of lower energy x-rays that have far more counts (and therefore higher chance of coincidence), but I realize that doesn't line up perfectly and the count rate is pretty low overall. Maybe some other feature in in play like a Compton edge, but I don't know how to make that determination.

Next steps?
I figure if the 150 KeV peak is from uranium 235, then if I do a radium spectrum it should NOT be there. Unfortunately I don't have a Ra spectrum done with my current setup, but it looks like that peak may be there on some of the lower quality spectra I do have. I found someone else's Ra spectrum that does clearly show that peak. Their hypothesis was that it was from incomplete processing that left uranium in the Ra. I feel that the concentration of U235 would still be extremely low even if the Ra was grossly impure, so I don't think this explains the observation.

The other thing I could do is put some distance between the sample and the detector to try to reduce coincidence (at the cost of lower count rate).

I'm open to all ideas, corrections, and suggestions.

[Mike S]

Update:
I did a run with radium watch hands and that peak at 150 KeV still shows up. Here's a comparison with the fossil shark teeth (natural U) and the Ra watch hands normalized to the Pb xray peak.

U vs Ra ThereminoMCA_2022_05_30_13_46_50.png

As near as I can tell, the backscatter peak for the 351 KeV Pb214 peak would be right about the same spot, centered at ~147 KeV, so that's likely a big contributor. My best guess at this point is the peak is primarily backscatter. That would explain why the examples I've seen are on amature setups which presumably have smaller shield interiors and detectors. I would imagine that U235 gammas would contribute some, but I'm still not convinced they would be enough to be the primary source of the peak in natural uranium with progeny. I do see a similar, possibly less prominent, peak on previous refined U samples I've run (U-glass, Fiestaware), but that makes sense since the U235 gammas aren't getting swamped by all the gammas from the daughters. Of course, there's literally books filled with what I don't know about gamma spectroscopy and nuclear physics, so I'd love to hear a more informed explanation.


I'm running another test now with more space between the sample and the shield surface and a little more distance from the detector. I'm only 155k counts in so far but it is definitely developing a ~150KeV peak, though you can see the lower "noise" floor from putting some distance between the sample and the graduated shield surface. I'll also plan to re-run some U-glass and Fiestaware samples using the same setup so I can do a better comparison.

[Sesselmann]

Mike,

Both of your theories are plausible, sum x-ray peak will land around 150 kV and some Compton back scatter from the higher energy peaks may also land around there.

In PRA you could try running the spectrum with a very tight pulse shape tolerance ~ 0.5 to 0.8 this should eliminate some of summing events.

Steven

[Sparky]

Mike S,

For what it is worth, I had a similar plot of a U bearing fossil and Radium watch hand in my files. I was mostly looking for difference at the early U daughters Th234 peaks with the fossil at the time. The peak around 150 kev (144 for U235) is still there with the Radium, but much less pronounced. I think your right that there has to be some backscatter there, but I believe a U235 component also.

Mike L

[Mike S]

Steve, thanks for the thoughts.

Mike S,

For what it is worth, I had a similar plot of a U bearing fossil and Radium watch hand in my files. I was mostly looking for difference at the early U daughters Th234 peaks with the fossil at the time. The peak around 150 kev (144 for U235) is still there with the Radium, but much less pronounced. I think your right that there has to be some backscatter there, but I believe a U235 component also.

Mike L

Mike L. Thanks for the comparison. I got to wondering what the difference would be if I used something other than the Pb x-ray peak for normalization. I pulled the data into Excel and did a crude normalization with the 186 and 352 peaks for what it was worth. If anything, it seems to me that the Ra sample looks more like backscatter and the U sample looks like backscatter plus perhaps some gammas. The Ra sample had a lot less backscatter due to the sample and orientation (the fossil teeth were much larger in volume and less optimally placed), so I think that explains some of the reduction in the peak.

U vs Ra Normalized 186keV.JPG

U vs Ra Normalized 352keV.JPG

If I normalize on the dip about 165 keV, then the peaks look roughly the same height, but on the U sample it is a little taller and broader.

U vs Ra Normalized 165keV dip.JPG

I feel like my original thought about the sum peak was a stretch. At such a low count rate (~20 cps) it just seems unlikely there would be a noticeable number of pile-ups. I do have my filtering set pretty tight and casual monitoring of the pulse waveforms that are getting through shows there's rarely even another pulse within the displayed time window.

I think my theory is moving to "The peak around 150 keV is caused by the backscatter peak from the 352keV Pb214 gammas; if U235 is present, its gammas around this energy (143 keV and, to a lesser extent, 163 keV) add to and broaden the apparent peak."

Thanks again for the thoughts. I'll keep experimenting and researching.

Mike S.

[Sparky]

Mike S,

There is a lot going on here. It is hard to sort out. I don't think normalizing at 186 between NU (with progeny) and Ra226 is useful because Ra226 and U235 both have high probability gamma emissions near this energy. NU (with progeny) might have proportional greater counts at this energy than Ra226 alone. I don't have the math to compare the two, maybe Ra226 dominates enough that there is no difference. If anyone knows how to determine the relative intensities and would care to teach, that would be great.

Normalizing at 360 kev is somewhat better. Note the elevated region for the tooth and that the peak is slightly to the left compared to the watch hand. I think this might be the effect of U235 at 144 keV for the tooth.

There is often differences in source thickness and geometries that can confound comparisons like this also.

Note that my Ra226 plot was done without shielding. This creates much less backscatter in this region than the shielded tooth spectrum because the Cu floor of my shield is a great backscatter target. I think that is why that whole region is elevated with the tooth spectrum. The peak near 144 does appear to be better defined for the tooth.

[Sesselmann]

Here is a quick comparison spectrum with PRA of an old clock face.

Steven

RA.png

[Mike S]

Thanks for the responses; it was a busy week and it took me a while to get back to this topic.

Steve,
It looks like their may be something developing about 150 keV on your plot but it would be competing with a lot of background (I think). I wonder if a bump would have shown up there if it ran for longer? I noticed that some of my previous Ra spectra done with a 2x2 detector didn't show the bump as prominently, but it wasn't immediately apparent why; I assumed it was due to lower resolution at low energy and higher background with previous shields. My little watch hands are giving this detector ~20 pps with a background of <4 pps. Do you suspect that what I'm seeing may be an artifact of Therimino?

All,

I did do a second run that spaced the sample away from shielding material; it went long because I just let it run while I was tied up during the week. I was impressed with the improvement in low energy peaks. It wasn't large, but it was clearly an improvement and my first clear results showing the impact of scatter from shielding material. Previously I had been so focused on eliminating Pb x-rays that I didn't invest much effort into scatter reduction. To be fair (to myself), my prior shield setups didn't really have much space for separation. I found it particularly interesting that the low energy peaks were also taller even though the absolute difference in counts was only 2%.

Coming back to the topic of the thread, the peak at ~150 keV was still clearly there. It was lower in absolute value, but still had roughly the same shape and size relative to the background. This comparison is in absolute counts per second with no scaling, so you can see where the backscatter reduction removed and added counts. Slight (and imperfect) energy adjustments were needed between runs due to temperature changes during the second run.

Ra 15hrs vs. Ra 69hrs with space.jpg

For kicks, I pulled the values into Excel and normalized on the 352 keV peak.

Ra 15hrs vs. Ra 69hrs with space Normalized to 352 keV.jpg

I'm now doing a run with some uranium glass marbles. I plan to let it go a while longer, but here's a preliminary look at how it is shaping up. Note that (as in the first image) this is in absolute counts per second without scaling between the two samples.:

Ra 69hrs spaced vs U marbles 25hrs spaced.jpg

Thanks again for the input, I feel like I'm learning something and improving my skill.

--Mike