Gamma Spectrometry Forum

Hi Guys,

I have just been building another GS-1010-LaBr3 for a client and was doing some testing today, thought I should share the results.

This build used a 1.5" Hamamatsu PMT and initially I struggled a bit with linearity, but it turned out I was running the voltage too high, I had to bring it down to 510 Volts before the linearity was acceptable.

Always fun to build detectors with these exotic crystals, but the internal noise from Lanthanum around 2000-2500 keV is a bit annoying.

I would like to try new alternatives like Strontium Iodide if I can get it, apparently similar resolution without the noise.


Steven

That's an amazing result!
I never got below 3.1% with my LaBr. Do you know how low it can go with an SBA PMT?
https://capescint.com/product-category/ ... -crystals/ This site sells SrI:Eu in case you're interested.
With your linearity check with interspec it would be interesting to mention the polynomial correction factors for your energy calibration

The difference between an standard and a UBA cathode pmt is around 0.5%
Luuk

GigaBecquerel wrote: ↑

29 Nov 2020, 21:47

I never got below 3.1% with my LaBr. Do you know how low it can go with an SBA PMT?
https://capescint.com/product-category/ ... -crystals/ This site sells SrI:Eu in case you're interested.
With your linearity check with interspec it would be interesting to mention the polynomial correction factors for your energy calibration

GBq,

Yes as it happened I had to lower the bias to 510V to get a linear spectrum, which meant I had to drive up the gain to max, this resulted in good linearity.

The Cs137 spectrum above was recorded before I corrected the voltage, which means the indicated 2.9% is understated by about 0.5%.

No polynomial function was used with Interspec, I calibrated three peaks in PRA using the interpolate method before exporting the data in text format.

Steven

This is really an extension of an earlier thread started under the Strontium Iodide [SrI2(Eu)] topic. But there was some discussion about my newly acquired LaBr3 detector. Steven made this for me in the Fall of 2020 and I have been playing with this and other detectors.

I will show some of the energy calibration data that I have obtained using various detectors, pulse acquisition hardware, and software. In particular I was looking at the energy calibration of the LaBr3 detector.

The astute observer will notice a bit of an unusual source used in the calibration set, namely I-131. My cat had to be treated for a thyroid issue and I-131 was used. When my cat came back from the vet I collected a small (~30 g) sample of her solidified urine (from litter box). This sample was double bagged and has been used as a source for the past month. I-131 has an 8 day half-life and I use a store and decay method with the balance of the litter box contents.

The first set of data shows four detectors used. All detectors are contained in a lead box comprised of 2-inch thick (~ 5 cm) lead. I did not use Cd, Sn, or Cu to block lead x-rays. The detectors used were NaI, CZT, SrI2(Eu), and LaBr3. In each instance I will provide in tabular form the data collected from each source, next will be a plot of the channel (or bin) vs energy with a polynomial fit. In the initial calibration the LaBr3 was used with the most recently purchased GS-USB-Pro and the Theremino v7.3 software. The LaBr3 energy calibration is quite similar to that of the Strontium Iodide detector which used entirely different hardware and software. The CZT detector used the Kromek Kspect software and the NaI detector (obtained from Spectrum Techniques) was used with a Spectrum Techniques UCS-30 MCA and the associated software.

In the next set of calibration data I swapped out the NaI detector for another similar NaI detector. The second NaI detector was a 38 mm x 38 mm Scionix detector refurbished by iRad (Tom Hall); I found this detector to work very well. It was used with a different GS-USB-Pro unit set-up for this detector and was used with Theremino.

The LaBr3 detector was used with the Spectrum Techniques UCS-30; a splitter was used as the UCS-30 has separate HV and signal connections. The voltage on the LaBr3 was reduced to 530 volts and the gain was adjusted to better position the photo peaks. The CZT and SrI detector configurations were unchanged.

For this set of calibration data the NaI response was more similar to that earlier of the LaBr3 detector. Whereas the LaBr3 detector with the UCS-30 was similar to that previously shown for the NaI detector and was quite linear.
Below is the result from using the LaBr3 detector on a small sample of monazite sand. First shown are the results from the earlier configuration of the LaBr3 with the GS-USB-Pro and Theremino and next are the results with the Spectrum Techniques UCS-30. I would say that in both instances the apparent resolution and peak separation is less than I had hoped to observe.

For the next set of energy calibration the LaBr3 detector was in an unshielded configuration. The data for the other detectors was unchanged from their earlier shielded results.

The LaBr3 detector was used with the GS-USB-Pro unit acquired with this detector. I reduced the voltage to 511 volts (couldn't quite get 510!) and the gain was turned to the maximum position and then the pot was backed off one full rotation. This was my initial attempt at using PRA v24; I had watched one of Steven's videos a while earlier.

I was surprised to see that the energy calibration curve was very similar to that obtained with Theremino.

I then went to Steven's most recent set of video tutorials, in particular the third on use of PRA. I realized that I had not made use of the highest data acquisition rate available and I should have used the boost gain rather than alter the number of bins per arbitrary unit. I used a Cs-137 source to acquire pulse shaping data. [As an interesting side note - it appears that better results are obtained if the source is positioned so that one obtains a count rate in the 200 to 500 counts per second range. I need to spend some more time experimenting with this.]

The photo below shows the placement of the source related to the unshielded LaBr3 detector. The results for the other three detectors are as before.
The resolution with Cs-137 was quite good as shown below. Not as good as Steven obtained, but I have noticed the effect of source placement relative to the detector and the count rate on the peak. More work is needed to examine this effect.
Following are the tabular and graphical results of this energy calibration. With PRA set up more consistent with the guidance from Steven a very linear energy calibration curve is observed.

The PRA results with the same monazite sand sample as before are shown. Clearly there is better peak separation.

Rob,

Nice set of data you have collected there..

Your findings are consistent with mine, best linearity was achieved with your detector at quite low voltages around 550V which means gain is quite low. That said I don't think we have to be too determined to use the whole 100 arb.u range, because in PRA we have 64,000 channels at our disposal (bin size 0.0015) , so even if our highest peak only makes it to 60 arb.u there is still plenty of resolution in the spectrum.

The actual LaBr3 crystal outputs plenty of light, so in your case I think the PMT could have done with a little more gain. PMT gain can vary quite a bit even within the same model number and batch. Some manufacturers label the PMT's with gain in mV at 662 keV. and it ranges from 250 mV to 500 mV.

In practical use absolute linearity is not usually critical, if you know the energy of the isotope you are looking for, you just set the voltage to a good gain and calibrate either side of the region of interest.

Steven