Gamma Spectrometry Forum

Hello everyone,

I'm really new to gamma spectroscopy and would like to share the results of my first few experiments. I welcome and appreciate any suggestions that you may have to improve my setup and operations. I wasn't sure if this was the right topic to post, please correct me if I'm wrong.

My detector is a refurbished Scionix 38B57 from iRad driven by a GS-USB-PRO. I house the detector in an improvised lead pit that I made with lead scraps. The shielded background count rate is 13.4 cps, a little over 4x reduction of the unshielded background.

I don't have a good calibration source yet, so I use a thoriated lantern mantle to do all my calibrations. I use PRA for all my measurements. At first I started acquiring spectra at 750 V and fairly short pulses, because that seemed the best way to enhance high energy peaks that were otherwise difficult to see. I soon discovered that this was not good at all, because it made the detector very non-linear above 900-1000 keV. Fortunately I found some posts on this forum that helped me understand what was going on, and I started experimenting with lower voltages and longer pulses. I read an advice from Steven to keep the K-40 1461 keV peak around 50 arbitrary units for best results, so I tried to do that.
I found that the best results in terms of linearity come with the pulse length at maximum setting and voltages of 650 V or below. Lower voltages make the detector more linear, but I had to raise the output volume to compensate. At 575 V I have a fairly good linearity up to 2600 keV, although the volume setting is at maximum and I had to add a 1.25x boost setting in PRA to keep 1461 keV around 50 arbitrary units.

Below are the results of my linearity experiments with various voltages. Each point in the plots is a major peak in the spectrum, the first plot has labels with the energies. The non-linearity always seems to appear beyond the 1588 keV Ac 228 peak.
Here is the thoriated mantle spectrum acquired at 575 V. For the interpretation I followed the spectra posted in this forum (thanks!) and elsewhere, along with some published energy lists. I'm not so sure about the 1588 keV Ac 228 peak, because I couldn't find a good published reference for that. Also, I'm not sure how to interpret the peaks at 2124 and 2303 keV. The first may be a single escape peak (2615 - 511 = 2104, well within the uncertainty), but I have no clue about the second. There might be a a couple of bumps at 209 and 277 keV from Ac 228 and Tl 208, but that seemed an over-interpretation to me.

Steve,

Thanks for posting, and well done on the linearity plots and setup, great spectrum.

I believe you have correctly identified the escape peak at around 2103, and the next little hump is the compton edge if I am not mistaken.

Steven

Thanks for the feedback, Steven.

Yesterday I tried to bring down the voltage to 550 V, linearity has improved a little bit although I had to up the gain 2x. This time I also added the 2104 keV single escape peak to the calibration, it looks like the linearity is very good all the way up to this energy.

I agree, very nice linearity..

Not surprising that the gain is low when you are running the detector at 550V.

I agree, very nice linearity..

Not surprising that the gain is low when you are running the detector at 550V, not sure what else you can do. How many mega ohm is your divider?

The divider was refurbished by iRad, it should be 120 MOhm.

That might explain the low gain, it's 10X what I normally use in my detectors. Such high impedance dividers are used to extend battery life on portable meters.

Steven

That makes sense. Gain is not really a problem anyway, the spectrum looks the same at low voltage and linearity is improved, so it's a win for me.

Thank you for the information regarding this question of calibration and linearity. I always wondered if there was a way of improving overall resolution Anyways - I tried your settings today with my 38B57 from iRad on 550V with a 10x gain on Theremino MCA / 15x on PRA24. I had excellent results!! ~ Similar results on my DIY CsI/Na probe at 800V with same gain settings.
It appears that this voltage setting with a higher software gain on both programs both greatly improves linearity with the energies below 100KeV as well as the clarity of the spectrum!