|
|
|
|
Gamma Ray Interactions and the Gamma Ray SpectrumA typical pulse height spectrum measured with a 3x3” NaI(Tl) detector with a 137Cs source is shown in the figure. The photopeak, Compton maximum and backscatter peak are indicated. The lines around 30 keV are Ba X-rays emitted by the source.
Electromagnetic radiation can interact with matter via
If the matter is a detector, then we can record the energies and with adequate energy resolution create an energy spectrum of the measured radiation. The three main interactions of radiation with matter are directly related to the spectrum of energy data recorded by the detector. In the photoelectric effect, the gamma ray interacts with an electron around an atom and removes it from its orbit (ionization) creating a photoelectron. The gamma ray energy is completely absorbed in the process. In a detector, this interaction results in all of the energy being deposited in the detector and account being recorded in a channel corresponding to the full energy photopeak. Compton scattering is like a billiard ball collision in which the gamma ray interacts with an electron giving up part of its energy and being scattered. The electron is scattered in another direction and each of the scattering angles depends on the amount of energy transferred from the gamma ray to the electron. As only a part of the energy is deposited a count is recorded in a channel corresponding to less energy and contributes to the background of the spectrum. A gamma ray can have additional interactions occur in the detector allowing the radiation to deposit all of its energy. However, if it scatters out of the detector we get a count in the background continuum. Pair production occurs only if the gamma ray has energy above 1.02 MeV. In this process the gamma ray interacts near the nucleus of the atom and a complex set of interactions occur. First the gamma ray is converted into matter in the form of an electron-positron pair that has a rest mass of 1.02 MeV but also share the additional energy of the gamma ray. Next, some of the energy is deposited in the detector. Finally, the positron annihilates creating two 511 keV gamma rays that may or may not interact within the detector. As a result, a count can be added in the full energy photopeak, one of two possible escape peaks or in a background channel. In practice, all effects have a chance to occur, this chance being proportional to the energy of the radiation and the atomic number (Z-value) of the absorber (the scintillation material). From the perspective of a detector, the only issue is whether by one of multiple interactions, the gamma ray deposits all or a part of its energy in the detector. If all the energy is deposited we get a count in the full energy peak if a part of the energy is deposited we get a count in another channel that essentially adds to the background. The total detection efficiency (counting efficiency) of a scintillator depends on the size, thickness and density of the scintillation material. However, the photopeak counting efficiency, important for e.g. gamma-ray spectroscopy, increases with the Z 4.5 of the scintillator. At energies below 100 keV, electromagnetic interactions are dominated by the photoelectric effect. The absorption can be calculated from the attenuation coefficient for a certain scintillator (or absorber). Pulse Height Spectroscopy
The basic principle of pulse height spectroscopy is that the light output of a scintillator is proportional to the energy deposited in a crystal. The standard way to detect scintillation light is to couple a scintillator to a photomultiplier. Furthermore, a g-ray spectrometer usually consists of a preamplifier, a main (spectroscopy) amplifier and a multi-channel analyzer (MCA). The electronics amplify the PMT charge pulse resulting in a voltage pulse suited to detect and analyze with the MCA.
Gamma Spectroscopy and Nuclide Identification For systems ordered with the gamma spectroscopy option, a spectroscopy grade tube base is provided complete with high voltage, amplifier, a 1024 channel analog to digital converter (ADC) and histogram memory. The unit comes with its own multi-channel analysis software. The unit provides gain stabilization and auto-calibration on start up. A built-in K-40 source is be used to gain stabilize the system. A functional representation of the detector and associated electronics employed in gamma spectroscopic measurements is shown below. The results of the analysis depend directly upon the quality of the data collected by the hardware. No software method can make good data out of bad data. Therefore, the set-up of the hardware in the counting system is extremely important for quality results.
A Wilkinson ADC takes the analog signal that passes through the input gate and charges a capacitor. The capacitor is discharged at a fixed rate by a constant current source to produce a linear ramp. The time it takes the capacitor to discharge is measured, and this time is proportional to size of the analog pulse. Wilkinson ADCs are frequently described by their clock speed, which is related to their throughput. A 100-MHz Wilkinson has a clock that produces 100 million pulses per second. The time required to convert an event is proportional to its height, so high-energy events are converted more slowly than low- energy events. The conversion time is also dependent on the conversion gain of the ADC. Successive-approximation ADCs work by a very different principle. They do a series of successive binary comparisons, the results of each comparison sets a bit in a binary number to a 0 or a 1. The voltage of the amplifier pulse is compared to a reference voltage. If the pulse is smaller than the reference voltage, we set the bit to 0 if it is larger we set the bit to 1. With each successive guess, the reference voltage is increased or decreased by ˝ of the previous voltage depending upon whether the pulse was larger or smaller than the reference voltage. As an example, we can pick a value from 0 to 7 and try to guess it using successive approximation. Let’s say the number is 5; the first guess is whether the number is bigger than 4, it is so we set the highest bit to a 1. As 5 is bigger than or equal to 4, we make the second guess 6 = 4+2. Now we ask is 5 bigger than 6? As the answer is no, we make the second bit a 0. For the third guess we ask whether the number is bigger than or equal to 5. As the answer is yes, we set the last bit to 1. In binary notation (base 2), the number 101 is equal to 5 in decimal value (base 10). Each extra guess doubles the resolution of the conversion gain so 10 guesses would allow a value from 0 to 1023 to be uniquely determined. As the number of guesses is constant, the dead time for each conversion is fixed. A typical conversion time for a successive approximation ADC is 10 microseconds. It is difficult to compare throughput of the two types of ADC directly, but assuming an average pulse of 5 V were to be used with both types of ADCs, the performance of a 10-µs fixed conversion time, successive-approximation ADC would be about equal to a 450-MHz Wilkinson type. channel 4000 / 450 • 106 channels/second = 9 • 10-6 seconds Originally, successive-approximation ADCs were not often used in spectroscopy applications because of their inherent non-linearity. In 1963, Cottini and Gatti described an averaging method that resulted in modern circuit designs capable of differential and integral linearity as good as Wilkinson ADCs. Multi-channel Analyzer A functional block diagram of a typical MCA is shown in the figure to the left. The architecture of the typical gamma spectroscopy system has changed over time. Today the analyzer can be a one-box traditional MCA, with display screen and control pad, or a NIM module that includes the ADC and histogram memory of the data. Display and control of the latter type system is accomplished through the use of a personal computer. Functionally, they do the same things: 1. Convert the analog voltage into a digital number, 2. Add 1 to the appropriate data memory channel, 3. Keep track of the live time and real time, 4. Allow display of the data, 5. Allow manipulation of the data, and 6. Allow transfer of data to computer for analysis. Sometimes, the ADC is external to the analyzer. In this case, additional cabling is required. The inputs to the ADC and analyzer include: 1. The linear energy signal, 2. The Busy output from the amplifier, and 3. The inhibit/pileup rejector signal from the amplifier.
A typical analyzer functionality is shown in this slide. The energy signal from the amplifier is split, with one part going to a delay and the other going to an SCA. If the analog signal is between the upper and lower discriminators, a logic pulse is sent to a linear gate which is opened to pass the pulse through. If the signal is outside the limits, then the signal is not presented to the ADC. The ADC also has a gate, an input gate, which is open when the ADC is available to process pulses. When the ADC is busy, this gate is closed even if the pulse passed the first linear gate operated by the SCA. These unprocessed pulses occur during the dead time of the ADC. To account for this dead time and the dead time from the amplifier, a fixed frequency clock generates pulses only when these logic signals are not being applied.
Gain
This figure illustrates the effect of count rate on peak shift. Temperature is another parameter that can affect the peak shift. Obviously, correcting the source of the problem is better than correcting the problem electronically. However, for long counting times, harsh operating environments, or rapidly changing count rates, the use of a digital stabilizer may be beneficial. Usually a two-point stabilization is employed. The high channel provides gain stabilization, while the low channel provides zero stabilization. A window or region of interest is defined around the two channels. This method adjusts the gain and zero by examining whether incoming counts fall in the lower or upper half of the region of interest. The principle of peak sensing is illustrated in the figure to the right. This method works best at medium to high count rates; however, at very low count rates, this method overcorrects and causes a deterioration in the resolution. These successive over-corrections in both directions are sometimes referred to as hunting. Nuclide Identification Assuming the hardware is working properly and has been calibrated and gain stabilized, the normal operations of the system are quite simple. Once radioactive materials are determined to be present by gross counting, the gamma ray detector is positioned at a reasonable distance from the source and a spectrum is acquired for a preset time. For measurements in the field, counting times from 30 to 60 seconds should be sufficient to collect enough data to get an accurate analysis. The analysis software matches the whole spectrum acquired to reference spectra taken for a variety of radionuclides of interest. The library of spectra can be modified by the supervisor as required, but the libraries provided typically contain all of the radionuclides of interest for defense and law enforcement purposes. The table below presents the nuclides included in the “Customs” library provided with the gamma spectroscopy option in Nucsafe systems.
For some TRMS and all CRMS systems, the gamma ray nuclide identification is automated and is keyed on the occupancy sensor. A spectrum is automatically acquired when the occupancy sensors detects the presence of a vehicle or object and is analyzed immediately after the occupancy sensor indicates the vehicle or object has moved past the system. Results are immediately displayed to the screen. |
|
