Doubts about spectroscopic isotope identification

[tim.hbn]

Hi

I have some doubts about the whole practice of spectroscopic isotope identification.

As an example, if, hypothetically, I was to give any of you a random sample of soil or something else and you examined it using a gamma spectrometer and you saw a peak at approximately 662keV, would you assume that you had found 137Cs?

From the following link, I have seen that it could also be 137mBa, 50mMn, 88mNb, 181Re, 181Au or a few other very close neighbours.

http://nucleardata.nuclear.lu.se/toi/Ga ... 60&Max=664

How can you be sure of what you have found?

Thank you very much.

Kind regards

Tim

[Sesselmann]

Tim,

You are right, there are plenty of gamma rays and many have energies in the same region of interest, especially when it comes to NaI(Tl) spectrometry where the peaks are relatively wide. HPGE spectra have a lot higher resolution and leave less for guessing.

It is important to look at the intensity as well as the energy. In the table you linked to you can see that the intensity of some of the gamma rays are simply so small that you would never be able to see them, most likely some of these are derived theoretically rather than experimentally, we simply know they are there because they have to be.

Scientists, despite their claims are never sure of anything, all they can do is give you a probability, so when the P factor is very high we go with it.

When we find 662 keV gamma in Japanese soil after a nuclear reactor in Fukushima melts down and sends out plumes of radioactive smoke, we put it down to Cs137.

"If you hear the clopper of hoofs, think Horses not Zebras"

Steven

[GigaBecquerel]

Tim,

When you see a 137Cs spectrum the 662 keV peak actually comes from 137mBa.
137 Cs undergoes beta decay to 137mBa, which then decays to 137Ba with the emission of a 662 keV gamma ray.

It most cases nuclides emit more than one gamm ray, so looking at relative intensieties of these makes more sense.
662 keV is also visible from 241Am, but if you don't also see a 59 keV peak with orders of magnitude higher intensity it's safe to bet that you don't have 241Am in your sample.
Also, look at the half life, daughtes and parents of your suspected nuclides, if it's synthetic and lasts 2.2 µs you can be fairly certain that it's not in a sample you picked up in a forest three weeks ago.

[tim.hbn]

Hi Steven and GigaBecquerel

Thank you both very much indeed for your very informative replies.

It therefore looks as if it is a matter of detective worrk.

Kind regards

Tim

[Sesselmann]

Tim,

You are right, there is a bit of detective work in spectrometry, and just like a good detective a good spectrometrist can look at a spectrum and see a lots of information, such as;

a) type and size of detector used
b) isotopes present
c) likely length of recording
d) if lead shielding was used
e) if crystal has been exposed to moisture etc..

Not to mention badly configured software, so whenever I have a client with spectrum issues I always ask for a screenshot of the spectrum, it usually tells me all I need to know.

Steven

[tim.hbn]

Hi Steven

Thank you very much for your help with this.

Kind regards

Tim

[jneilson]

Yes, there's sometimes quite a bit of detective work in IDing nuclides. There's quite a lot of potentially interfering peaks that can potential cause misidentification, especially with low and intermediate resolution gamma spec. Things get easier with professional HPGe detectors as the sharp peaks and gain stability mean there's less uncertainty in the energy of the peak, but there's still close peaks that can cause issues; 186keV Ra-226 or U-235 is a common difficult one. There's a few different things to look at and think about when making identifying decisions:

1. Does the suspected nuclide have any other peaks you should be able to see? (if you've got more than one matching peak, you can be much more confident in your identification)
2. Does the suspect nuclide have any daughters with peaks you should detect?
3. Are any of the parent nuclides of your suspect nuclide present?
4. Could the suspect peak be a sum peak, escape peak, or backscatter peak from anything else in your spectrum?
5. Do the peak areas of each peak match up? Of course, if you've identified one nuclide by confirming with different peaks, how can you be sure part of the peak doesn't come from an interfering nuclide with similar energy? Checking the ratios of peak areas against the gamma intensities from the nuclear data can reveal if there's a discrepancy. If one peak has an intensity of 99% and another 50%, you expect the second to be roughly half the area (you've also got to account for the difference in energy-dependent detection efficiency). If a peak is taller than it should be compared to the other peaks, that's an indication there might be something else there.
6. What's the half-life of the suspected nuclide? If it's very short, it's likely to have decayed away before you could measure it, so it's probably not that - but do keep in mind the decay of parents can cause short lived nuclides to be constantly produced (leading to secular equilibrium)
7. Where did the sample come from and is it likely to have the suspected nuclide in - the medical saying of thinking horses not zebras that Steven posted can be good guidance.

I've got plenty of interesting detective cases of resolving these kinds of ambiguities in my professional work, I'll put a couple below:

1. The Cs-137 that wasn't:
A metallic sample was neutron irradiated so the microstructural neutron damage and lattice dislocations caused by irradiation could be studied. Before it was sent to the materials lab for the microstructral testing, a quick gamma spec was done using a detector with a small CZT crystal, checking on the neutron activation and confirming and there wasn't any contamination from the facility where it was irradiated. The initial gamma spec showed the neutron activation products we expected (probably something like Co-57, I don't remember exactly what it was), but there was also a small peak around 660keV. This concerning seemed like Cs-137 contamination; so it was rescanned with a HPGe for a longer count time to confirm. On the HPGe spectrum, the energy came up more precisely as 657keV, and we could also see additional slightly weaker peaks at 884 and 937 keV, which had not been seen on the previous scan due to the lower detection efficiency for higher energy peaks on the small crystal. This new energy precision and the presence of the other peaks enabled us to identify Ag-110m rather than Cs-137 - and that could be explained via neutron activation of a small impurity in the sample.

2. "Am-241" in a welding rod:
Someone had some old thoriated tungsten welding rods on a nuclear site, and needed to get rid of them, but because they were in a controlled area, they needed to be appropriately monitored for contamination. They were scanned by the health physics team using a portable automated radionuclide identifier (RIID), which reported the expected Th-232 but also Am-241. Am-241 contamination in that area would have been very unusual so they were sent to the lab for proper gamma spec, and were put in front of a HPGe. The initial analysis by a junior analyst reported the concentrations of the thorium decay chain, but also a peak in the Am-241 region of interest at 59.5 keV was reported. I looked at the spectrum and sure enough, that peak was there. However, doing the due diligence I saw that the 59.5 line could also be explained by tungsten xrays, which would make sense as the majority material is tungsten. Tungsten also has a characteristic xray around 68keV, and looking back at the spectrum there was something there too. Analysts often ignore small features in the x-ray region as these x-ray peaks are typically not useful for quantifying activity, so that had been missed by the initial analysis. Next was demonstrating whether the 59keV peak could be fully explained by tungsten xrays or whether we might still have Am-241 present as well - for that, I considered the relative intensities of the 59 and 68keV peaks, and the ratio matched what you'd expect for just tungsten, so we could rule out Am-241 and just report the expected Th activity.