Ever since Bern and Delft Universities discovered LaBr3:Ce as scintillators, numerous groups have increased our understanding of their properties [1,2]. In terms of their commercial availability, Saint-Gobain Crystals has made a great deal of progress.
Drawing primarily from results presented or published elsewhere, this summary reports LaBr3:Ce detectors’ performance, extending results to the 3” long, 3” diameter (“3x3”) crystals [3,4,5]. It does not try to perform a comprehensive review – the reader must remember that the general literature is partially compiled on the Saint-Gobain website, www.crystals.saint-gobain.com.
At the end of this summary there is a table of scintillator properties. It shows that LaBr3:Ce crystals have significantly more rapid decay times and superior timing properties than Nal(Tl) for energies near 1 MeV, as well as emitting 60% more light than them.
Relative efficiency and energy resolution are covered in this summary as a function of gamma-ray energy, highlighting a comparison of LaBr3:Ce and Nal(Tl) detectors for the 3”x3” size. Performances are also assessed against temperature and count rate, intrinsic background, and coincidence resolving time.
Energy Resolution Versus Energy
The fantastic energy linearity and high light output of LaBr3:Ce crystals determines their energy resolution.
In the results displayed in this subsection, 3” long, 3” diameter LaBr:Ce and Nal(Tl) detectors were analyzed alongside one another. These packages were both integrated, meaning they were directly coupled with a 3” diameter photomultiplier. For the LaBr3:Ce detector, this was a Photonis XP5300B, while the Nal(Tl) was paired with an ETI 9305.
The source was on axis with the detector, referred to as being “end-on.” Once the source was altered from one isotope to a different one, the distance was altered in order to attain counting rates that were reasonable (I.e., a few thousand per second). This same distance was used for each of the detectors in order to allow for a direct comparison of the two detectors.
The comparison begins with the two detectors’ responses to 137Cs (662 keV). Figure 1 compares the 3”x3” detector spectra. Both the barium Ka X-ray at 32 keV and the source’s gamma ray at 662 keV are shown. On the energy scale, spectra are normalized to 662 keV.
Also reported in Figure 1 are the areas beneath the 662 keV photopeak, which was set at 100% for Nal(Tl). This produced a relative efficiency of 118% for the LaBr3:Ce unit, mainly as a result of its enhanced density. The peak for Nal(Tl) is near 32 keV, which is marginally higher on the energy than that for the BrilLanCe 380 detector. This is because Nal(Tl) produces marginally more light per keV at lower energies than higher ones, as it is non-linear.
Figure 1. Comparison of 3”x3” spectra for 137Cs (662 keV) LaBr3:Ce detector (red) and NaI(Tl) (blue)
The two detectors’ responses for 60Co are shown in Figure 2, where the well-known lines at 1173 and 1332 keV are visible. The BrilLanCe 380 unit delivers 2.1% more energy resolution at 1332 keV than the Nal(Tl) unit, which only gives 5.4%. Consequently, it has 43% more efficiency.
Additionally, in the BrilLanCe 380 curve, there is a line in the 35 keV region. This is a result of the emission of Ba X-rays from 138La background which will be discussed further in this article.
The LaBr3:Ce detector achieves an energy resolution of 1.6% at 2615 keV (208Tl in the thorium decay chain), which compares to 4.5% for Nal(Tl). This makes the former 65% more efficient, as Figure 3 shows.
Once more, spectra have been normalized at the highest energy line of 2615 keV, and, as a result of the linearity variations, energy offsets are seen between the two materials at lower energies. The LaBr3:Ce package’s improvement in spectral resolution becomes especially visible in the multiple energy spectrum. Once again, the line visible just under 1500 keV in the LaBr3:Ce spectrum is a result of background and is discussed further below.
Figure 2. Comparison of 3”x3” spectra for 60Co LaBr3:Ce detector (red) and NaI(Tl) (blue)
Figure 3. Comparison of 3”x3” spectra for the Thorium decay chain. LaBr3:Ce detector (red) and NaI(Tl) (blue)
BriLanCe 380 detectors’ advantages carry on into low energies. This is shown in Figure 4, which displays responses to 57Co. The 136 keV line is clearly resolved from the 122 keV line by the BrilLanCe 380 detector, while this is not the case for the Nal(Tl). Also visible in the BrilLaCe 380 spectrum is the background from Ba X-ray lines in the 35 keV region.
In order for the survey to be complete, a 133Ba spectrum is shown for the two detectors in Figure 5. There is a visibly enhanced separation of the lines near 350 keV for the LaBr3:Ce detector. Prominent lines marginally above 30 keV are visible for both detectors, as Cs Ka X-rays are emitted by the source.
Figure 4. Comparison of 3”x3” spectra for 57Co LaBr3:Ce detector (red) and NaI(Tl) (blue)
Figure 5. Comparison of 3”x3” spectra for 133Ba LaBr3:Ce detector (red) and NaI(Tl) (blue)
This section’s results are summarized in Table 1, which tabulates relative efficiency and energy resolution for numerous of the presented energies. At all of these energies, BrilLanCe 380 detectors’ advantages over Nal(Tl) are visible.
Table 1. Summary 3”x3” Detector Response vs. Energy Resolution and Relative Efficiency
Figure 6 displays the well-behaved nature of the energy resolution against energy. This data is overlaid on points extracted from smaller detectors and earlier work. As expected statistically for linear detectors, the energy resolution follows the square root of energy faithfully. It is possible to conclude from the data in Figure 6 and Table 2 that Nal(Tl) does not track this scheme, even though it is not demonstrated.
The LaBr3:Ce scintillator is over 10 times as fast as the Nal(Tl) and has 1.6 times the light output. In the pmt, this can produce non-linear effects. The instantaneous charge pulse is approximately 25 times that of Nal(Tl), as is shown by a simple calculation. This is based on a 1.6 pH ratio multiplied by the more rapid time factor of 250 ns/16 ns, specifically: 25X = 1.6X(250/16).
The non-linearity is manifested in two different ways. Firstly, there is a better-than-expected FWHM peak at that energy. Secondly, compared to what is expected from a linear extrapolation, the position of higher energy peaks will be at a lower pulse height.
Whether or not this is happening can be verified by an approximate 100 V decrease in HV and an observation of an improvement in linearity and similar marginal decrease in FWHM. Saint-Gobain Crystals is selecting pmts with improved linearity properties, as a means of minimizing these undesirable effects. Frequently, these are 8-stage pmts.
Figure 6. Energy resolution as a function of energy
Response Versus Temperature
LaBr3:Ce crystals’ remarkable properties are preserved with increases in temperature, as displayed in Figure 7. Light output is also drastically higher at enhanced temperatures than for the other crystals tested. At room temperature, and in similar high temperature, rugged packages which are useful in oil well logging, LaBr3:Ce emits 160% pf Nal(Tl)’s light output, as recent testing has confirmed.
Response Versus Rate
Given that the difference in decay time between LaBr crystal and Nal(Tl) is tenfold, there is a performance to high rate expected, which is shown in Figure 8. For these measurements, 1”x1” crystals were used, as was the same 8575 photomultiplier for each detector, proceeded by a constant fraction discriminator and a timing filter amplifier.
By changing source position and strength, rate was adjusted. The expected difference between materials is verified in this demonstration, however detailed results remain highly dependent on the chosen electronics configurations.
Figure 7. Response of scintillator at temperature with Photomultiplier tube held at constant temperature.15
Figure 8. Response versus Rate
Coincidence Resolving Time
The LaBr3:Ce crystals have fantastic timing properties, as anticipated from the high light output and decay time properties. In the Table of Scintillation Properties at the bottom of this summary, this is suggested by the figure of merit column. In column three, the figure of merit (FOM) is the square root of decay time over light output. This is an indication of the anticipated timing performance compared to other scintillators.
LaBr3:Ce crystals are rapid enough to make timing results dependent upon light propagation times and consequently on the size of the crystal, as shown in Figure 9. Different groups have demonstrated that decay time is just a rudimentary indicator of rise time, which is a more illuminative parameter. This latter parameter is, however, extremely dependent on the photomultiplier choice and is harder to measure.
Figure 9a shows representative coincidence resolving times (CRT) for numerous sizes of LaBr3:Ce detectors. The data was obtained with the use of two Photonis XP20Y0 photomultipliers (PMTs). In order to be measured, the PMT serving as the STOP channel was coupled to the crystal. The other PMT was coupled to a 4x4x5 mm LaBr3:Ce crystal, a dedicated START crystal.
Previously, this crystal had been measured using a “single channel” CRT value of 115 ps. This term denotes a value which would be measured against a channel which was infinitely fast. A 22Na source was used to obtain CRT data. In order that events which led to a 511 keV photopeak in each of the channels were counted, the system was gated in a specific manner.
LaBr3:Ce has a somewhat superior CRT, as demonstrated in Figure 9a. The dependence upon crystal size is also visible, increasing as light transit times and crystals become larger. As the Photonis XP20Y0 is a tube of 51 mm diameter, the 76 mm point is a special case. This is because the Photonis CP20Y0 does not entirely cover the crystal.
A reflective annulus was placed to cover the area of the window which was not covered by the PMT when the larger diameter crystal was measured. Although this geometry’s effect on CRT has not been quantified, it is hypothesized that it increases its value.
The single channel CRT, which was measured for a number of geometries with a standard PMT which had a plano-plano face plate, is shown in Figure 9b. As it is possible to maintain the fantastic energy resolution of LaBr3:Ce crystals with standard PMTs, this is a crucial point.
It is worth noting that the CRT is critically dependent upon the PMT. For instance, a much worse performance is delivered by the XP2060 38 mm PMT than the larger PMTs.
Figure 9a. Coincidence Resolving Time (CRT) of LaBr3:Ce detectors as a function of the crystals’ longest dimension measured with fast timing PMTs that have plano-concave face plates.
||PMT Size (mm)
* Diameter and length of right cylindrical crystal.
** CRT is the Coincidence Resolving Time (single channel)
*** These are standard PMTs with plano-plano photocathode face plates.
Figure 9b. Timing measured at 511 keV with LaBr3:Ce Integrated Detector
138La and 227Ac Background
With the same decay scheme shown in Table 2.9 for 66.45% of its decays, 138La is a naturally occurring radioisotope of La with an abundance of 0.09%. 138La produces excited 138Ba after undergoing electron capture (EC). This then subsequently decays by emitting a 1436 keV gamma.
The refilling of the electron shell constitutes a necessary by-product of electron capture. This results in coincident barium X-rays being emitted in the 35 keV region. The 33.6% remaining decays proceed by beta emission to 138Ce, which decays by the emission of a 789 keV gamma in coincidence, while the beta has an end-point energy of 255 keV.
Self-counting is easily used to measure the background spectrum. Such a spectrum is shown for a 1.5”x1.5” detector (38x38 mm) counted for approximately 3 days (278278 seconds) in a low background chamber in Figure 10.
If the self-counting spectrum is reviewed from left to right, firstly, a beta continuum at low energies for 138La is seen. This decays to 138Ce, in which the 789 keV gamma has entirely escaped the detector. This beta-only spectrum carries on to its 255 keV end-point.
From approximately 255-750 keV, the Compton continua from the 789 and 1436 keV gamma rays is shown by the spectrum. As we continue to higher energies, the 789 keV line is next. However, as it is in coincidence with the beta, it is smeared to an enhanced energy in a gamma plus beta continuum, which ends marginally above 1 MeV.
Ultimately, the 1436 keV gamma is seen, however it is displaced to a higher energy by roughly 37 keV to 1473 keV, as a result of the coincident capture of X-rays caused by the filling of the Ba K levels after K-electron capture. Likewise, the 1436 keV gamma plus 5 keV cause the hump near 1441 keV on the low energy side of the 1473 line, as a result of coincident capture of X-rays when the Ba L-level fills after L-electron capture.
In Figure 10, the Ba K X-rays at 37 keV are only seen partially, as a result of the MCA’s discriminator setting. The sum line at 1473 keV may be used as a calibration peak, and has a measurable and constant activity which can be subtracted within statistical limits in order to determine 40 K, even though it potentially interferes with detection of 40 K at 1441 keV.
Figure 10. Self-counting background spectrum for a LaBr3:Ce detector
Table 2. 138La decay scheme. (from 8th edition, Table of the Isotopes)9
Background data extended to energies above 1750 keV can be seen in Figure 11. The presence of low level alpha contaminants is revealed. These have been shown to result from 227Ac contamination.10 Detectors produced early in the program contained higher levels of these alpha emitters and subsequent process refinements reduced them to the point that 138La now produces the dominant background features.
Figure 11. Self-counting background count to 3000 keV
Table 3 summarizes overall background for a detector approximately 38x38 mm.
Table 3. Background Count Rates/cc From 1.5” x 1.5” Detector
||0-255 keV beta continuum
||790 keV – 1000 keV gamma + beta
||1468 gamma peaks
||Alphas above 1600 keV
Reports demonstrating LaBr3:Ce’s relatively good radiation hardness are becoming available11,12. Authors of these conclude that the material could be useful to space missions, as it copes well against protons and deals with gamma irradiation sufficiently.
There is a drop in light output (approximately 8%) reported in unpackaged crystals after 1 kGy under 60Co gamma exposure. The pulse height resolution at 662 keV also deteriorates from 3.0% to 3.8%. After this, the rate of deterioration is seen to substantially slow. Performance remains useful even after exposure to 111 kGy (82% light output and 4.8% FWHM).
These authors’ data shows that recovery with temperature and time is, at best, incomplete and slow. Details of the packaging are also important to radiation resistance, as related work suggests13. It is worth noting that this radiation resistance is significantly better than that seen with Csl(Tl) or Nal(Tl).
At least as rugged as Nal(Tl), LaBr3:Ce is a robust scintillation crystal. For systems used behind drill bits in oil well logging for measurements carried out during drilling operations, properly packaged detectors of both materials are suitable. These detectors are able to survive temperatures of 200 degrees centigrade, 30 g rms random vibration, and 1000 g shock.
The thermal expansion coefficient of LaBr3 is not isotropic, and it has a crystal structure which is asymmetrical. At first, this made crystal yields problematic with boules fracturing while they cooled to room temperature. Saint-Gobain Crystals has since improved the growth process significantly, and is now able to grow ingots sufficient to manufacture a detector which is 244 mm long and 97 mm in diameter. The material becomes quite robust once it is grown and cooled to room temperature.
Table of Scintillator Properties14
||Light Yield (photons/keV)
||1/e Decay time t(ns)
||Wavelength of maximum emission λm (nm)
||Refractive index at λm
||Thickness (cm) for 50% attenuation (662 keV)
Compiled and Edited by C. M. Rozsa, Peter R. Menge, and M. R. Mayhugh. Originally prepared for distribution by SaintGobain Crystals at IEEE NSS/MIC San Diego CA, November 2006.
1 E.V.D. van Loef , P. Dorenbos, C.W.E van Eijk, H.U. Gudel, K.W. Kraemer, Applied Physics letters, 77, 1467-1469 (2000).
2 E.V.D. van Loef , P. Dorenbos, C.W.E van Eijk, H.U. Gudel, K.W. Kraemer, Applied Physics Letters, 79, 1573-1575 (2001).
3 Peter R. Menge, G. Gautier, A. Iltis, C. Rozsa, V. Solovyev to be published in Proceedings of 2006 Symposium on Radiation Measurements and Applications, Ann Arbor MI (2006)
4 A. Iltis , M. R. Mayhugh, P. R. Menge, C. Rozsa, O. Selles, V. Solovyev, III Workshop on Advanced Transition Radiation Detectors Proceedings, Ostuni, Italy Sept 7-10, 2005, to be published in Nucl. Instr. and Meth. A.
5 C. M. Rozsa, M. R. Mayhugh, P. R. Menge Presentation to the 51st Annual Health Physics Society Meeting, Providence Rhode Island, June 27, 2006. Available on our website. Search Rozsa or Health Physics.
6 Temperature data for NaI(Tl), CsI(Na), CsI(Tl), Plastic Scintillator (BC438), and BGO are from: “Characteristics of Scintillators for Well Logging to 225 °C” by C.M. Rozsa, et al. prepared for the IEEE Nuclear Science Symposium, San Francisco, October 1989. The PMT is held near room temperature. The temperature response for BrilLanCe 380 (LaBr3:Ce) and BrilLanCe 350 (LaCl3:Ce) were measured in mid 2005 in high temperature packages, again with the PMT held isothermally near room temperature. To overlay the temperature curves from these two eras, the relative pulse heights for BrilLanCe 380 (LaBr3:Ce) and BrilLanCe 350 (LaCl3:Ce) were measured by comparing 1” diameter x 1” long crystals in low temperature packages to NaI(Tl) of the same size also in a low temperature package. This result placed the curves at 130% and 75% of NaI(Tl) at room temperature, as shown. As mentioned in the text, later data shows BrilLanCe 380 (LaBr3:Ce) to be over 160% of NaI(Tl) at room temperature for more recent data all taken in high temperature Ti-sapphire packages. Other groups report similar room temperature light output, 160% NaI(Tl) or greater.
7 G. Bizarri, J. T. M. de Haas, P. Dorenbos, and C. W. E. van Eijk, Phys. Stat. Sol. (a) 203, No. 5, R41– R43 (2006)
8 For details search High Count Rate or Note 519 for the work by Vladimir Solovyev at detectors.saint-gobain.com.
9 Table of Isotopes, Eighth Edition. Richard B. Firestone, Virgina S. Shirley, Ed. John Wiley & Sons (1996)
10 T.W. Hossbach, W.R. Kaye, E.A. Lepel, B.S. McDonald, B.D. Milbrath, R.C Runkle, L.E. Smith. Nuclear Instruments and Methods in Physics Research, Section A, 547, 2-3, pp 504-510, August 1, 2005
11 Gamma-Ray Induced Radiation Damage in LaBr3:5%Ce and LaCl3:10%Ce Scintillators. W. Drozdowski, P. Dorenbos, A. J. J. Bos, S. Kraft, E. J. Buis, E. Maddox, A. Owens, F. G. A. Quarati, C. Dathy, and V. Ouspenski, IEEE Transactions on Nuclear Science, Vol. 54, No. 4, August 2007, 1387
12 Effect of Proton Dose, Crystal Size, and Cerium Concentration on Scintillation Yield and Energy Resolution of LaBr3 :Ce W. Drozdowski, P. Dorenbos, A. J. J. Bos, J. T. M. de Haas, S. Kraft, E. Maddox, A. Owens, F. G. A. Quarati, C. Dathy, and V. Ouspenski IEEE Transactions on Nuclear Science, Vol. 54, No. 3, June 2007
13 Gamma Ray Induced Radiation Damage in Ø1” × 1” LaBr3:5%Ce Winicjusz Drozdowski, Pieter Dorenbos, Adrie J.J. Bos, Alan Owens, Francesco, G.A. Quarati (Private communication via preprint)
14 The 2007 version reflects revised density and attenuation coefficients for the BrilLanCe materials.
15 The 2009 version reflects new temperature response data.
This information has been sourced, reviewed and adapted from materials provided by Saint-Gobain Crystals.
For more information on this source, please visit Saint-Gobain Crystals.