Resolution of several NaI(Tl) detectors

[Bruce Meagher]

I recently acquired a basic gamma spectroscopy system. Currently, it consists of:

1) Tracor Northern TN-7200 multichannel analyzer (supports both pulse height analysis and multichannel scaling)
2) A NIM bin containing an Ortec 485 amplifier, Ortec 556 HV power supply, and a Tennelec 812 pulser
3) Ortec 276 preamplifier/PMT base
4) Several NaI(Tl) detectors. See picture below.

After doing some basic calibration and testing with the pulser I wanted to take some rough energy resolution measurements of the different NaI(Tl) detectors. The detectors (along with the rest of the setup) are used, old, and of unknown condition. However, once I adjusted the threshold and zero on the pulse height analyzer I was able to use the pulser as an input to the preamp to verify the rest of the system. The system remained linear over the 1024 channels at several different amplifier gain settings (or at least to some small unmeasurable error with my current test equipment), and the counts closely match the pulses inputted.

The detectors: two Harshaw 12S12 3"x3" NaI(Tl); one Teledyne Isotopes NaI(Tl) 3"x3" with a through-well; (still debugging the other two)

The FWHM energy resolution results using a 5uCi Cs137 source disk, at channel 662, with 100 second live time were:

Harshaw #1: ~9%
Harshaw #2: ~12%
Teledyne: ~17%

I then ran the same test on a 0.1uCi Cs137 source over a 30 min time window with similar results. Finally I ran a resolution vs voltage test to see if I could significantly improve the resolution based on voltage. The results of this experiment didn't significantly change the above results (see table below). I'm sure there is an optimum voltage to run each detector, but at the resolutions I'm measuring it's in the noise right now.

I've read past posts where many of you are getting 7.5%-9% on your 3"x3" NaI(Tl) detectors. Besides disintegration/leakage and cracking what are some other causes for poor resolution? The pulses coming out of the detectors through the preamp look clean on the o-scope, and the count rates can be very high (if I put the spicy source close I can spike the % dead time meter). Do PMTs typically fail in a way that degrades resolution? These three detectors seem sealed pretty good. In your experience is there something that might be fixable to improve things if I crack these open, or will I just be able to potentially verify the root cause?

Carl, in a recent post you mention you measured the resolution of your 2"x2" NaI(Tl) detector at ~5% on a Cs137 source. Is there anything special you are doing with your system to get such a fantastic energy resolution out of an NaI(Tl) detector?

Any other thoughts on where or how I might be introducing errors in my measurements or setup?

Thanks,

Bruce

pic #1 setup; pic #2 Harshaw #1 with Cs137; pic #3 Harshaw #2 with Cs137; pic#4 Teledyne with Cs137

[Carl Willis]

Hi Bruce,

I don't see anything woefully amiss with your setup. I think this is the realm of tweaking more than hunting down a specific error. Here are some general suggestions:

-Give the equipment, including the HV supply and the dynode chain, time to warm up before use. This eliminates a lot of drift.

-Your method of calculating FWHM is an acceptable estimate for many purposes, but it will give you a number that's too high if there are many continuum counts underneath the Cs peak. It is best to write a calculation that subtracts local baseline and then fits a Gaussian to the peak.

-17% for the side well detector doesn't sound too bad. Crystal shape and light pathways have a major influence on performance, illustrated by that resolution that doesn't compare with the cylindrical crystals.

-You could get a result faster and possibly with less drift by turning down the gain. Try to get 10-20 channels only in your FWHM for the purposes of this measurement.

-Adjust the pole-zero function on the amplifier.

-Keep dead time under ~5% for a good resolution measurement. Position the source accordingly.

-Wiggle all cables, connectors, terminators, and the PMT socket during operation to look for gain shifts. A surprising number of noise problems come from bad contacts.

-You might try a different HVPS. I have had problems with the Ortec 556 supplies.

-Better luck: that's right, it's a crapshoot with NaI detectors. My ~5.5% unit is a 1.5x1.5" and it's sheer luck that it is as good as it is. I have gone through dozens of scintillators and generally I sell off ones that don't compete with what I have already. I'm pretty happy with the two I use right now.

-Carl

[Bruce Meagher]

Carl, thanks for taking the time to respond to my post. Your suggestions are extremely helpful.

On the warm up issue I have observed some drift vs time/temperature. In fact, I observe a drift on all the detectors from the morning over several hours. I've used the pulser to eliminate everything upstream of the preamp. My latest thinking is that the crystal is soaking overnight in the cold and when the temperature control comes on in the morning it takes time for the crystal mass to warm up to the room temp (which also fluctuates quite a bit over the day). I need to run an experiment to better understand detector output voltage vs a controlled temperature to see if this is really the cause or if I have some other issue. It's a little disconcerting to calibrate the setup with a source and an hour later the system has shifted. Doesn't inspire confidence.

When changing voltage or switching to a different detector is there a rule of thumb on the amount of time to let everything warm up?

With respect to the removing the baseline, with the hot Cs137 source the background is really low (count wise), but with my weaker source this makes a significant difference. The TN-7200 has a strip feature which comes in handy for this, but I really need to get my computer connected for better offline analysis.

I'll investigate your other suggestions to see if I can identify any issues.

Bruce

[Richard Hull]

Carl is correct in that with any surplus crystal and even some new ones, it is pretty much luck of the draw as to what you draw in the way of resolution. You will only be able to do so much in the way of tightening up your electronics and procedures. After that is done, you have what you have.

I have five different nice scintillator/PMT combos and no two are alike and, like any parent, I favor the one with the best resolution even if it is not my big 5X5 Bicron. the others are also-rans and backups.

Richard Hull

[Steven Sesselmann]

Bruce,

Recently I have bought quite a few new NaI detectors, and any reputable manufacturer will give you a test report with the detector stating it's performance.

An FWHM better than 7.5% for 662Kev is considered acceptable, once you get out around the 12% to 17% it's pretty ordinary for a NaI detector.

Steven


BeeResearch