In our research, we utilize highly efficient bismuth germanate scintillators (BGO). In the past, most researchers have not used BGO because of two reasons: 1) its poor energy resolution and 2) its temperature dependent light output.

Much work has been performed on the resolution problem in BGO in recent years by detector manufacturing firms. Many firms, such as Rexon and Saint Gobain-Bicron, now offer affordable BGO detectors with energy resolutions of 10% or better at 662 keV (nearly comparable to 6-7% for sodium iodide (NaI) scintillators). While the energy resolution for BGO cannot yet match that of NaI, the gamma-ray detector efficiency of BGO is much higher than any of the large volume commercially available scintillators. In Figure 1, we show a comparison of various scintillators (NaI, cesium iodide (CsI), gadolinium ortho-silicate (GSO), and BGO for the same volume (5 cm x 5 cm) in the energy region around 5 MeV. This figure clearly shows the BGO’s excellent photopeak efficiency. This efficiency is due ^[i] to its high Z (83) and its high density (7.13 g/cc compared to 3.67 g/cc for NaI). The efficiency is one of the keys in reaching the sensitivity limits required by many of our applications.

• Figure 1. Comparison of the efficiencies of various scintillators. See text for details.

When BGO’s temperature increases its light output decreases.  In a practical sense, this decrease in light output effectively changes the gain of the amplification system. When the gain of the amplifier is changed, the energy calibration of the spectrum is also changed. Our work required us to solve this problem since this makes detection of radionuclides difficult.

• Figure 2. Variation of gamma peak position as function of temperature for a BGO detector.

This temperature dependence is shown in Figure 2, wherein the position of a particular gamma-ray from a radioactive source was tracked while the BGO is heated and cooled. The position of the gamma ray peak changes roughly 1.5% for every degree Celsius.  This process is reversible i.e. cooling or heating the detector to its original temperature will return the peaks to their original position. A common method to compensate for this temperature-dependency is to utilize refrigerators to maintain a constant temperature on the crystal. For a portable system, this solution is not an option due to the increased bulk and weight.

• Melcor PT Series Thermoelectric Cooler.

In our work on the coal analyzer, we developed a cooler system based on thermoelectric coolers which use the Peltier effect to provide heating and cooling.  The Peltier effect occurs whenever current passes through the circuit of two dissimilar conductors; depending on the current direction, the junction of the two conductors will either absorb or release heat. The amount of heat pumped is in direct proportion to the current supplied. The Peltier effect is utilized to its maximum when thermocouples are made of material of different conductivity.

Coolers and heatsinks are mounted on an aluminum jacket. This jacket is placed around the detector. If the ambient temperature is constant, these Peltier effect coolers can keep the detector at constant temperature.

In portable systems, we utilize a small radioactive ¹³⁷Cs source (0.2 μCi, an exempt quantity) to correct for this deficiency. It is located on the crystal. Prior to data collection, a quick 10 second count of this source is performed. The position of the 662 keV gamma ray is established relative to a measurement previously performed in the laboratory. The amplifier is then adjusted to place the position of the ¹³⁷Cs peak into the position that it held in the laboratory. In the PELAN, this is performed automatically with no input required from the user.

The fact that the light output changes does not change the efficiency of the detector. It will, however, broaden peaks but the amount of broadening is small compared to the resolution of the detector.

• Figure 3. Graphs of temperature, relative humidity, and change in energy position of the 6130 keV gamma ray during the demonstration. Note: the left axis is both relativity humidity in % (lowermost plot) and degrees Farhenheit (uppermost plot). The x-axis denotes the spectrum number for the entire two-week demonstration.

A BGO detector was utilized in the PELAN system for two weeks for a demonstration for the U.S. Navy. Figure 3 shows the temperature readings (^oF) (uppermost curve) and relative humidity (%) during the demonstration. During the tests, the ambient temperature varied from 53^o F to 93^o F (10^o C to 33^o C). Relative humidity varied between 22% to 63%. The energy of the 6130 keV gamma ray from oxygen is also plotted on this graph to evaluate the effectiveness of the temperature compensation system. Without any correction, the energy should vary approximately 30% or nearly 2000 keV. As one can see, the variations in this energy are less than 1%! Thus, the automatic gain correction system seems to be working extremely well.

For more detailed information, please see the Publications portion of this web site.