Posted on by maxfomitchev

Cerium-doped lanthanum bromide / LaBr3(Ce) scintillators offer better resolution and greater stopping power when compared to conventional thallium-doped sodium iodine / NaI(Tl) scintillators.

E.g. a typical 1.5″ x 1.5″ LaBr3(Ce) crystal achieves a 2.5-3.5% FWHM @ 662 keV resolution compared to a 6-7% resolution characteristic of 2″ x 2″ NaI(Tl) crystals (1.5″ x 1.5″ lanthanum bromide crystal is roughly equivalent to a 2″ x 2″ sodium iodine in terms of gamma stopping power).

The advantages of LaBr3(Ce) scintillators over NaI(Tl) are:

  • 2-3x better resolution;
  • 2x smaller size at the same stopping power;
  • Fast (100 ns) peaking time allows for much faster counts (10x faster than sodium iodine);
  • exceptional linearity.

The drawbacks of LaBr3(Ce) compared to NaI(Tl) are:

  • Much higher price: a new 1.5″ x 1.5″ high-resolution (e.g. 2.5% FWHM @ 662 keV) lanthanum bromide gamma detector costs almost $10,000 (aftermarket detectors are almost non-existent);
  • Self-activity of lanthanum bromide produces peaks at 30 keV and 1.5 MeV, with the latter being especially unwelcome since it overlaps with the important 40K line.

Do the advantages outweigh the disadvantages? With the exception of 40K the answer is a definite ‘yes’. To put things in perspective, here are some examples of spectra captured using a Saint Gobain Brilance 38S38 lanthanum bromide detector (Fig. 1) compared to the same spectra captured using a Scionix 38B57 sodium iodine detector.Fig

Saint Gobain Brilance 38S38 LaBr3 Detector
Fig. 1. Saint Gobain Brilance 38S38 LaBr3 Detector with Scionix voltage divider.

The Brilance detector uses Hamamatsu R6231 photomultiplier, which is an 8-stage tube that requires an 8-stage divider (most 14-pin 2″ dividers are 10-stage). Following Hamamatsu R6231 recommended voltage distribution ratio and using the Hamamatsu PMT handbook I have implemented a simple divider circuit shown on Fig. 2.

Fig. 2. Voltage divider circuit.

I have also tried a stock Scionix VD14-E1-X35 voltage divider, which yielded similar results. Bicron AS20 divider specifically designed for LaBr3 detectors was not available for testing.

After a lot of tinkering I was able to achieve 3.3% resolution at 662 keV. The important thing to realize is that with lanthanum bromide scintillators even a minute amount of noise from power supply, divider, or cabling will dramatically reduce resolution. With average electronics the resolution is just a tiny bit better than with NaI.

To compute spectrum I relied on PulseCounter software, which I setup to sample at 20 MHz and tuned the proprietary edge detection algorithm for 150 ns rise time.

Fig. 3 shows Cs-137 spectrum captured with Brilance 38S38 compared to Scionix 38B57. LaBr3 resolution at 662 keV (3.6%) is ~2x better than NaI (6.8%). With LaBr3 detector a single-point calibration yields exceptional linearity: without trying I was able to match spectral peaks within 1 keV accuracy. This compares favorably to most NaI detectors where a 10 keV accuracy may be difficult to achieve even with multi-point calibration.

LaBr3 Cs-137 Spectrum Compared to NaI
Fig. 3. LaBr3 (blue) Cs-137 spectrum compared to NaI (red).

Fig. 4-6 compare Co-6, Co-57 and Ba-133 spectra captured using the two scintillators. Starting at about 100 keV LaBr3 clearly outperforms NaI in terms of spectral line resolution and the gap between NaI and LaBr3 performance widens with the increase in gamma energy.

Fig. 4. LaBr3 (blue) Co-60 spectrum compared to NaI (red).
Fig. 5. LaBr3 (blue) Co-57 spectrum compared to NaI (red).
Fig. 6. LaBr3 (blue) Ba-133 spectrum compared to NaI (red).

For energies under 100 keV there is very little difference in resolution, which is seen on uranium spectrum – Fig. 7.

Fig. 7. LaBr3 (blue) uranium spectrum compared to NaI (red).

Background spectrum illustrates self-activity of LaBr3, which is characterized by a 30 kV line and a peak around 1.4 MeV – Fig. 8. Background count rate was ~120 CPS.

Fig. 8. LaBr3 (red) background spectrum compared to NaI (blue).

Fig. 9 and 10 illustrate low-activity (< 200 CPS) thorium and lutetium-176 spectra captured with LaBr3 detector.

Fig. 9. Thorium spectrum captured with LaBr3 detector,
Fig. 10. Lutetium-176 spectrum captured with LaBr3 detector,

Conclusion

Unless you are counting potassium-40 or measuring low-activity samples (which generate count on par with the detector’s self-activity), cerium-doped lanthanum bromide detectors are beat sodium iodine scintillators by a healthy margin. Once you start capturing spectra in higher resolution it is hard to go back.

On the other hand LaBr3(Ce) scintillators require high-quality electronics, faster sampling rates and finely-tuned signal processing in order to achieve superior resolution. Noisy or unstable systems will essentially negate the LaBr3(Ce) resolution advantage.

High cost of LaBr3(Ce) is also a factor. Crystals larger than 1.5 x 1.5″ are cost-prohibitive for most applications.