Mobile spectrometer

Scintillation crystals, PMTs, voltage dividers etc...
Post Reply
ChrisqMalibu
Posts: 29
Joined: 13 Mar 2022, 08:00
Contact:

Mobile spectrometer

Post by ChrisqMalibu » 19 Mar 2022, 18:20

Hi all,

I just bought Atom Fast 8850 & it's delivered today to one of my brothers in Bangkok, Thailand. It can't do spectrometer (but it's ok because I bought him a set of GS + 1.5" NaI probe. I need to buy one more for my other bothers.

I just saw this online called RaysID with FWHM 662keV = better than 8.5% for ~ 599.99 EUR but it doesn't support iOS yet (Atom Fast apps is available for iOS). RadiaCode 101 may not be able to submit their iOS app due to the sanction. So I might buy the 3rd Atom Fast 8850 which works well with iOS just for its sensitivity.

I'm just curious how well RadiaCode 101 & RaysID work as spectrometer? The highest-spec RaysID looks promising for food/water inspection on the go?


User avatar
isoenzyme
Posts: 37
Joined: 20 Feb 2019, 17:36
Location: Utah (USA)
Contact:

Re: Mobile spectrometer

Post by isoenzyme » 19 Mar 2022, 19:17

Greetings ChrisqMalibu:
I own and have used the RadiaCode 101 and quite like it. Like you I don't expect an iOS solution any time soon; it works reasonably well with Windows 10 and Windows 11 on my Surface Go for mobile use (though it collects data without being connected to either a phone or computer and the data can be downloaded later). Plans are to upload some data to this site so you (and everyone else) can see the spectra. It comes "pre-calibrated" for energy so you can't calibrate it yourself (which I don't entirely like), but the spectra I've collected seem pretty believable.
Christopher Lloyd
Salt Lake City, Utah (US)


Setup: Modified GS-STANDUP-20 with GS-USB-PRO
Primary Detector: GS 2" x 2" CsI(Tl)

ChrisqMalibu
Posts: 29
Joined: 13 Mar 2022, 08:00
Contact:

Re: Mobile spectrometer

Post by ChrisqMalibu » 20 Mar 2022, 11:39

Thanks, Chris.

What I know so far is that RadiaCode 101 spec from https://radiascan.com/products/detector ... ogle-maps/
"The energy resolution (FWHM) of the cesium-137 662 keV line is not worse: 12%+/-1%" <== This would make it not very good for low-activity Cs-137 identification in food/water per below. Still, with shielding & giving it a lot of time, it may be ok (at least for just detection, not identification). Have you tested RadiaCode 101 against GS + probe to see if it would miss low activity nuclides? I'm waiting for my Cs-137 source, Atom fast 8850 to come in to find out myself. I plan to wrap Cs-137 source with lead sheet by sheet until Atom Fast can no longer detect it. Then, I'll use GS+probe to see if it can see anything. If it can, at least I would know the lower limit for Atom Fast (& know when not to trust it for food/water inspection 100%). I'm going to also get RadiaCode 101 & RaysID for testing as well. If you have everything to test this with RadiaCode 101, please let us know.

RaysID.com claims much larger (than Atom Fast which is already larger than RadiaCode 101) 5cm³ CsI/Tl scintillation crystal with a version that can handle FWHM 662keV better than 8.5% resolution mentioned at https://raysid.com/price.html

== FWHM better than 8.5% <== I asked to buy one of these but nothing is available yet. iOS app is planned for end of 2022.
These devices use our most successful sensors.
== FWHM 8.5%-11%
Such resolution is enough to identify most izotopes(comparing to 8.5% version) and check food for Cs-137.
== FWHM 11%-15% <== This would possibly be where RadiaCode 101 is. They claim it may miss even detection on low activity which is no good!
Some izotopes can't be identified (comparing to 8.5% version). Cs-137 can be detected but on quite high activity.
== FWHM >15%
Izotope identification is not possible or not evident. It is still very sensitive and fast scintillation counter with full functionality except izotope-identification.

I'm waiting to hear back about the resolution of Atom Fast 8850. But at this points, I plan to wait for the most expensive RaySID. I have also purchased Atom Fast 8850 which should be on the way to me from Moscow. I'll have something much better than my Radex RD1212-BT to take with me all the time (so slow & its iOS app is terrible).

I'll send Atom Fast 8850 to my brother (I also gave him GS + 1.5" NaI detector so he'll have a better spectrometer) if RaysID turns out much better at food/water inspection as claimed. For portability, I can't complain comparing to below viewtopic.php?f=5&t=932. I'll use GQ GMC-600 Plus & RaysID for rapid detection. If nothing found in both, I'll use GS + probe. This should be good enough to catch just about any low-level contamination. I'll improve it with better probe & shielding as well.
"2 x 2 " CsI crystal and a Photonis XP6242 PMT, and the result came out very nice. Resolution at 662 was around 5.8%" <== This would be a fun project to try.

kotarak
Posts: 56
Joined: 07 May 2020, 05:46
Location: Brookfield, CT, USA
Contact:

Re: Mobile spectrometer

Post by kotarak » 21 Mar 2022, 10:58

I hope you do differentiate between resolution and efficiency. If your main goal is to detect low level radioactivity, the efficiency (or the ability of the crystal to interact with high-energy photons (gamma) and produce low energy photons (light) is more important than FWHM.
In an eventual nuclear accident, the fission products are well known and it will not be difficult to identify them even if the resolution is not that great.

A few things that can help you with very low activities:
- detector with larger crystal - the larger the crystal, the higher is going to be your counting sensitivity. Radiacode and Raysid use very small crystals so they are not going to be very efficient. 63x63 mm crystals or larger have much better efficiency than the tiny CsI in those device.
- sample geometry - consider the use of a Marinelli beaker - for low level samples it will help a great deal if you situate the sample as close as possible and around the crystal. Alternatively, you can use a crystal with built-in well and place the sample inside the well.
- shielding - a good lead castle is paramount to analyzing low activity samples - consider at least 2"-3" of lead in a 4pi configuration. More lead is always better and copper lining or grading the shielding with different metals will help you suppress the XRF photopeak coming from the lead.
Andrey E. Stoev
Brookfield, Connecticut, USA

ChrisqMalibu
Posts: 29
Joined: 13 Mar 2022, 08:00
Contact:

Re: Mobile spectrometer

Post by ChrisqMalibu » 24 Mar 2022, 05:46

Thanks, Andrey! I'll only use GQ GMC-600 Plus & Atom Fast 8850 for rapid check. If nothing found, I'll use GS + 1.5" NaI probe to double check. This way, we won't have to use GS for everything.

I have families & relatives spreading all over Thailand. I'll give each locations at least a Radex 1503 (just as a small Geiger counter to share among the hoiusehold) & GQ GMC-500 Plus or 600 Plus to use its Food/water inspection (only good if it detects something obvious). If nothing was detected, they would need to send samples in to my 2 brothers which will eventually have the exact same setup as mine with all the equipment needed to detect low activity contamination.

I'll definitely looking into getting the 3" ones, will buy Marinelli beaker & will definitely build shielding with 300 lbs of lead for sure.

User avatar
isoenzyme
Posts: 37
Joined: 20 Feb 2019, 17:36
Location: Utah (USA)
Contact:

Re: Mobile spectrometer

Post by isoenzyme » 29 Mar 2022, 18:44

As for the "efficiency" of the RadiaCode I took some measurements using both my RadiaCode and my 2" CsI crystal that was shielded in my GS-STANDUP filled with Lead shot (and lined with copper and tin) and was surprised a bit by the results. Basically, I tested several Cs-137 sources (0.01, 0.05, 0.1, 0.25, 0.5, 1 and 2 uCi standards from Spectrum Techniques).
[attachment=2]RadiaCodeSetup.jpg[/attachment]
Simply placing each source on the RadiaCode and letting it measure for about 30-60 seconds one can record the activity (it reads it in either microR/hr or milliR/hr for the more active sources). Taking the same sources and testing them with the shielded detector (see viewtopic.php?f=10&t=492 to see the setup) I could measure the counts/second and get some fairly good statistics:
[attachment=1]GS Timecourse.pdf[/attachment]
I was expecting the shielded crystal (2" CsI(Tl)) to significantly outperform the wee-tiny crystal in the RadiaCode but was shocked to see it performing pretty well comparing them as Geiger Counters (and not spectrometers):
[attachment=0]Combined Cs-137.pdf[/attachment]
The units are different, and it's sometimes difficult to look at log-log plots for comparison, but both systems achieved < 10 nCi detection (between 6 to 8 nCi for each using 3*SD for the least-square fit). Overall, I'm pretty pleased with the performance (unless I've really messed up here).

I plan on testing them with my Uranium standards and will show screen shots of the spectra generated by the RadiaCode when time permits.
Attachments
Combined Cs-137.pdf
(34.99 KiB) Downloaded 17 times
GS Timecourse.pdf
(11.12 KiB) Downloaded 10 times
RadiaCodeSetup.jpg
Christopher Lloyd
Salt Lake City, Utah (US)


Setup: Modified GS-STANDUP-20 with GS-USB-PRO
Primary Detector: GS 2" x 2" CsI(Tl)

ChrisqMalibu
Posts: 29
Joined: 13 Mar 2022, 08:00
Contact:

Re: Mobile spectrometer

Post by ChrisqMalibu » 03 Apr 2022, 09:03

It may be interesting to check against Co-60 source @ https://youtu.be/MSxD964J8sY I understand RadiaCode 101 use cube sensor unlike Atom Fast 8850 giving it similar sensitivity from all directions (with smaller sensor, of course).

ChrisqMalibu
Posts: 29
Joined: 13 Mar 2022, 08:00
Contact:

Re: Mobile spectrometer

Post by ChrisqMalibu » 03 Apr 2022, 13:22

Is it possible that PMT & the rest may be the weakest link for most of the large crystal like yours? I understand RadiaCode 101, Atom Fast 8850 & RaysID are all using much smaller crystal (doped with Tl) with silicon PMT plus optimized pipelining & machine learning-type of processing to achieve something similar to computational photography. Shall we call this computational nuclear physics?

As you could see here RyasID mentions it very clear "Dose rate value is energy-compensated and sensitivity-compensated. There is no big overestimation for Am-241 or underestimation for Co-60" @ https://raysid.com/#specification <== This sounds like machine learning or knowledge engineeering-based approach, doesn't it?

Maybe it's Tl doping?

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest