NaI Kristall

[Peter-1]

Hello,
I'm looking for a function that calculates the effectiveness of a NaI scintillator from energy. The range is 50–300 keV. The crystal is 2" x 2". Is there a link to this?

Peter

[Sesselmann]

Do you mean the absorption efficiency?

This chart taken from a Saint Gobain publication shows the absorption efficiency of NaI crystals of various sizes.

NaI Efficiency

[Peter-1]

Hello,
what I need is the efficiency only for NaI in the range from 50 - 400 keV.

[Sesselmann]

Peter,

As per the chart I posted above your 2" NaI crystal is 100 % efficient under 200 keV and only drops off slightly at 400 keV, but the detector housing will attenuate low energy gamma below 100 keV so this is probably going to be responsible for the biggest loss off counts. You need to know the exact thickness of your detector housing and allow for this attenuation as well.

Steven

[Peter-1]

The background to my question is that I am looking for a way to examine samples for the ratio of U235 / U238. I've now found a report. However, it's still missing the efficiency function of the NaI crystal. But the first results are already quite useful.
https://physicsopenlab.org/2017/06/30/g ... -measures/

So I have a soil sample from Gottow, the old research station in 1939-1944. The measurement and calculation showed a content of 0.68% U235. According to reports from the Federal Office for Radiation Protection, the Earth is said to be slightly contaminated with natural uranium. This is very true.

Peter

[Lposter]

You will need the efficiency for the geometry - sample holder - you are using. You will need to calibrate the detector using a known standard for that geometry.

This will be difficult for you to do - in addition you will need to correct the calibration for the density difference between the calibration solution or matrix and that of your sample as this is important for the energy range you are looking at.

If you have information as to your detector - crystal size and thickness of the aluminium housing - and information as to your sample holder - material, thickness of the container walls, fill height of the sample - and details of the soil sample - density primarily - I can make you a calibration for a 2" crystal by Monte Carlo.

But - such calculations are vulnerable, at that energy range, to the precision with which you provide information as to the detector and geometry.

But they should be within 10% or so for energies less than 200 keV, even with just approximate information.

Otherwise, you will have difficult task getting calibration data.

An alternative is to purchase a soil standard material of known U content from teh IAEA, fill that in your sample holder and use that spectrum to determine some efficiency data.

Let me know if you want to pursue the Monte Carlo method. But you will have to provide the information as outlined above.

U measurements by NaI are no joke however.......

Lars

[Rob Tayloe]

When I worked at a US uranium enrichment plant we used an NaI detector as part of a system to monitor the U-235 enrichment of liquid UF6 prior to the filling of product cylinders. It has been a good while since I worked at that facility and I don't have the original paper and technical studies and didn't find the reference that I thought that I was seeking. I did, however, find a number of other related reports, papers, and such that may be of interest. Some papers require a subscription in order to access the full document; typically only an abstract and/or some summary paragraphs are available without a subscription. There are a few full documents available for downloading.
[Steven - these links work as of 19 March 2025; perhaps this posting should be cross-posted under the reference topic (if you think that there would be sufficient interest)]
Links follow -

Multi-detector system approach for unattended uranium enrichment monitoring at gas centrifuge enrichment plants
https://www.osti.gov/pages/servlets/purl/1398927
https://www.sciencedirect.com/science/a ... 0217310033

Comparison of Several Detector Technologies for Measurement of Special Nuclear Materials
https://apps.dtic.mil/sti/pdfs/ADA451969.pdf

Advancing the On-Line Enrichment Monitor (OLEM) Capability
https://resources.inmm.org/sites/defaul ... 9/a374.pdf

A Next Generation On-Line Enrichment Monitor (OLEM) Prototype
https://resources.inmm.org/sites/defaul ... 011319.pdf

Comparison of portable detectors for uranium enrichment measurements
https://link.springer.com/article/10.10 ... 008-0619-5

The Measurement of Uranium Enrichment
https://link.springer.com/chapter/10.10 ... -58277-6_8

Nondestructive Assay of Nuclear Materials for Safeguards and Security
https://link.springer.com/book/10.1007/ ... 31-58277-6

Uranium enrichment identification and assay using scintillation NaI detectors of different sizes and types
https://www.researchgate.net/publicatio ... _and_types

[Peter-1]

Hello Lars,
I can't contribute much to the factors. Pictured is my 2" NaI crystal. The Gottow soil sample is 5cm in diameter and 1cm high. This sample, in a plastic container, lies directly on the crystal. The thickness of the Al layer is unknown. The sample is said to be natural uranium. The 25mm lead shielding still has 8mm Cu. I can't write any more about it.
Greetings Peter

[Lposter]

Without detailed info as to both sample and detector ..... MC will be handicapped.

Either way.... I could not find drawings of that model on the net so I assumed it was a 2"x2" with a steel housing (it doesnt look like Al).

They are often 0.5 mm thick so I went with that. I dont know what the sample holder is made of so I just left it out. I dont know the soil density or composition so I went with a generic dry sand of 5cm x1cm cylinder. I assumed the sample was placed on the steel housing.

These generalisations will impact the accuracy of the model and its results especially at low energy.

The sample model was filled with 1000 Bq each of isotopes that emit the following energies with 100% probability, 50, 80, 120, 200, 400, 600, 1000, 1500, 2000, 3000keV.

So from the sample you have 1000 photons/s at each of the energies. This was counted for 10 000s. So you should have 10000000 photons at each of those energies leaving the sample over that time period.

Two spectra are provided I hope. One with typical NaI resolution applied. And one without - the peaks becomes lines in this one and it is easier to determine their areas.

The spectrum was between 0 and 3500 keV over 2048 channels. There should be 1.7089 keV per channel (although it doesnt really matter). There is no background - the only signal is from the isotopes in the sample.

Knowing how many photons at each energy left the sample, and how many were seen by the detector at each energy....the peak efficiency can be calculated.

Then you need to fit a curve and from that the efficiency at any energy can be obtained.

As I said.....for energies above 300 keV, this is probably within 10%. For energies below.....it gets worse as the energy gets lower.

Without very detailed info as to your detector and sample.....there can be no improvements using MC. Thats just the way it is. Normally x-rays of the detector and detailed drawings would be used to make as good a model as possible.

Spectra are in .n42 format - can be converted using Cambio or Interspec or whatever. The files were renamed to .pdf to allow upload. rename to .n42 when you get them.

Lars

[invader]

if you dont know the thickness of the Al layer, you can try to get that value indirectily, one source, Am241 would be a good candidate, make a spectrum reading , them use a sheet of known thickness of the Al, and make a second spectrum, when the thickness of the Al is the same as the thickness of the Al of the crystal, the reading of the Am241 gamma peak should be half of the first reading.

.