Many radiation detectors produce an electrical signal after each interaction of a particle or photon. The signal generated by the detector passes through a series of electronic circuits, each of which performs a function such as signal amplification, signal processing, or data storage. A detector and its associated electronic circuitry form a detection system. There are two fundamental ways that the circuitry may process the signal—pulse mode and current mode. In pulse mode, the signal from each interaction is processed individually. In current mode, the electrical signals from individual interactions are averaged together, forming a net current signal.

There are advantages and disadvantages to each method of handling the signal. GM detectors are operated in pulse mode, whereas most ionization chambers, including ion chamber survey meters and the dose calibrators used in nuclear medicine, are operated in current mode. Scintillation detectors are operated in pulse mode in nuclear medicine applications, but in current mode in direct digital radiography, fluoroscopy, and x-ray computed tomography (CT).

In this chapter, the term interaction typically refers to the interaction of a single photon or charged particle, such as the interaction of a γ-ray by the photoelectric effect or Compton scattering. The term event may refer to a single interaction, or it may refer to something more complex, such as two nearly simultaneous interactions in a detector. In instruments that process the signals from individual interactions or events in pulse mode, an interaction or event that is registered is referred to as a count.

The term spectroscopy, literally the viewing of a spectrum, is commonly used to refer to measurements of the energy distributions of radiation fields, and a spectrometer is a detection system that yields information about the energy distribution of the incident radiation. Most spectrometers are operated in pulse mode, and the amplitude of each pulse is proportional to the energy deposited in the detector by the interaction causing that pulse. The energy deposited by an interaction, however, is not always
the total energy of the incident particle or photon. For example, a γ-ray may interact with the detector by Compton scattering, with the scattered photon escaping the detector. In this case, the deposited energy is the difference between the energies of the incident and scattered photons. A pulse height spectrum is usually depicted as a graph of the number of interactions depositing a particular amount of energy in the spectrometer as a function of energy (Fig. 17-2). Because the energy deposited by an interaction may be less than the total energy of the incident particle or photon and also because of random variations in the detection process, the pulse height spectrum produced by a spectrometer is not identical to the actual energy spectrum of the incident radiation. The energy resolution of a spectrometer is a measure of its ability to differentiate between particles or photons of different energies. Pulse height spectroscopy is discussed later in this chapter.

The efficiency (sensitivity) of a detector is a measure of its ability to detect radiation. The efficiency of a detection system operated in pulse mode is defined as the probability that a particle or photon emitted by a source will be detected. It is measured by placing a source of radiation in the vicinity of the detector and dividing the number of particles or photons detected by the number emitted:

This equation can be written as follows:

Therefore, the detection efficiency is the product of two terms, the geometric efficiency and the intrinsic efficiency:

where the geometric efficiency of a detector is the fraction of emitted particles or photons that reach the detector and the intrinsic efficiency is the fraction of those particles or photons reaching the detector that are detected. Because the total, geometric, and intrinsic efficiencies are all probabilities, each ranges from 0 to 1.

The geometric efficiency is determined by the geometric relationship between the source and the detector (Fig. 17-3). It increases as the source is moved toward the detector and approaches 0.5 when a point source is placed against a flat surface of the detector, because in that position one half of the photons or particles are emitted into the detector. For a source inside a well-type detector, the geometric efficiency approaches 1, because most of the particles or photons are intercepted by the detector. (A well-type detector is a detector containing a cavity for the insertion of samples.)

The intrinsic efficiency of a detector in detecting photons, also called the quantum detection efficiency (QDE), is determined by the energy of the photons and the atomic number, density, and thickness of the detector. If a parallel beam of monoenergetic photons is incident upon a detector of uniform thickness, the intrinsic efficiency of the detector is given by the following equation:

where µ is the linear attenuation coefficient of the detector material, ρ is the density of the material, µ/ρ is the mass attenuation coefficient of the material, and x is the thickness of the detector. This equation shows that the intrinsic efficiency for detecting x-rays and γ-rays increases with the thickness of the detector and the density and the mass attenuation coefficient of the detector material. The mass attenuation coefficient increases with the atomic number of the material and, within the range of photon energies used in diagnostic imaging, decreases with increasing photon energy, with the exception of absorption edges (Chapter 3).

A gas-filled detector (Fig. 17-4) consists of a volume of gas between two electrodes, with an electric potential difference (voltage) applied between the electrodes. Ionizing radiation forms ion pairs in the gas. The positive ions (cations)
are attracted to the negative electrode (cathode), and the electrons or anions are attracted to the positive electrode (anode). In most detectors, the cathode is the wall of the container that holds the gas or a conductive coating on the inside of the wall, and the anode is a wire inside the container. After reaching the anode, the electrons travel through the circuit to the cathode, where they recombine with the cations. This electrical current can be measured with a sensitive ammeter or other electrical circuitry.

There are three types of gas-filled detectors in common use—ionization chambers, proportional counters, and GM counters. The type of detector is determined primarily by the voltage applied between the two electrodes. In an ionization chamber, the two electrodes can have almost any configuration: they may be two parallel plates, two concentric cylinders, or a wire within a cylinder. In proportional counters and GM counters, the anode must be a thin wire. Figure 17-5 shows the amount of electrical charge collected after a single interaction as a function of the electrical potential difference (voltage) applied between the two electrodes.

Ionizing radiation produces ion pairs in the gas of the detector. If no voltage is applied between the electrodes, no current flows through the circuit because there is no electric field to attract the charged particles to the electrodes; the ion pairs merely recombine in the gas. When a small voltage is applied, some of the cations are attracted to the cathode and some of the electrons or anions are attracted to the anode before they can recombine. As the voltage is increased, more ions are collected and fewer recombine. This region, in which the current increases as the voltage is raised, is called the recombination region of the curve.

As the voltage is increased further, a plateau is reached in the curve. In this region, called the ionization chamber region, the applied electric field is sufficiently strong to collect almost all ion pairs; additional increases in the applied voltage do not significantly increase the current. Ionization chambers are operated in this region.

Beyond the ionization region, the collected current again increases as the applied voltage is raised. In this region, called the proportional region, electrons approaching the anode are accelerated to such high kinetic energies that they cause additional ionization. This phenomenon, called gas multiplication, amplifies the collected current; the amount of amplification increases as the applied voltage is raised.

At any voltage through the ionization chamber region and the proportional region, the amount of electrical charge collected from each interaction is proportional to the amount of energy deposited in the gas of the detector by the interaction. For example, the amount of charge collected after an interaction depositing 100 keV is one tenth of that collected from an interaction depositing 1 MeV.

Beyond the proportional region is a region in which the amount of charge collected from each event is the same, regardless of the amount of energy deposited by the interaction. In this region, called the Geiger-Mueller region (GM region), the gas multiplication spreads the entire length of the anode. The size of a pulse in the GM region tells us nothing about the energy deposited in the detector by the interaction causing the pulse. Gas-filled detectors cannot be operated at voltages beyond the GM region because they continuously discharge.

Because gas multiplication does not occur at the relatively low voltages applied to ionization chambers, the amount of electrical charge collected from a single interaction is very small and would require huge amplification to be detected. For this reason, ionization chambers are seldom used in pulse mode. The advantage to operating them in current mode is the almost complete freedom from dead-time effects, even in very intense radiation fields. In addition, as shown in Figure 17-5, the voltage applied to an ion chamber can vary significantly without appreciably changing the amount of charge collected.

Almost any gas can be used to fill the chamber. If the gas is air and the walls of the chamber are of a material whose effective atomic number is similar to air, the amount of current produced is proportional to the exposure rate (exposure is the amount of electrical charge produced per mass of air). Air-filled ion chambers are used in portable survey meters and can accurately indicate exposure rates from less than 1 mR/h
to tens or hundreds of roentgens per hour (Fig. 17-6). Air-filled ion chambers are also used for performing quality-assurance testing of diagnostic and therapeutic x-ray machines, and they are the detectors in most x-ray machine phototimers. Measurements using an air-filled ion chamber that is open to the atmosphere are affected by the density of the air in the chamber, which is determined by ambient air pressure and temperature. Measurements using such chambers that require great accuracy must be corrected for these factors.

In very intense radiation fields, there can be signal loss due to recombination of ions before they are collected at the electrodes, causing the current from an ion chamber to deviate from proportionality to the intensity of the radiation. An ion chamber intended for use in such fields may have a small gas volume, a low gas density, and/or a high applied voltage to reduce this effect.

Gas-filled detectors tend to have low intrinsic efficiencies for detecting x-rays and γ-rays because of the low densities of gases and the low atomic numbers of most common gases. The sensitivity of ion chambers to x-rays and γ-rays can be enhanced
by filling them with a gas that has a high atomic number, such as argon (Z = 18) or xenon (Z = 54), and pressurizing the gas to increase its density. Well-type ion chambers called dose calibrators are used in nuclear medicine to assay the activities of dosages of radiopharmaceuticals to be administered to patients; many are filled with pressurized argon. Xenon-filled pressurized ion chambers were formerly used as detectors in some CT machines.

Air-filled ion chambers are commonly used to measure the related quantities air kerma and exposure rate. These quantities were defined in Chapter 3. Air kerma is the initial kinetic energy transferred to charged particles, in this case electrons liberated in air by the radiation, per mass air and exposure is the amount of electrical charge created in air by ionization caused by these electrons, per mass air. There is a problem measuring the ionization in the small volume of air in an ionization chamber of reasonable size. The energetic electrons released by interactions in the air have long ranges in air, and many of them would escape the air in the chamber and cause much of their ionization elsewhere. This problem can be partially solved by building the ion chamber with thick walls of a material whose effective atomic number is similar to that of air. In this case, the number of electrons escaping the volume of air is approximately matched by a similar number of electrons released in the chamber wall entering the air in the ion chamber. This situation, if achieved, is called electronic equilibrium. For this reason, most ion chambers for measuring exposure or air kerma have thick air-equivalent walls, or are equipped with removable air-equivalent buildup caps to establish electronic equilibrium. The thickness of material needed to establish electronic equilibrium increases with the energy of the x- or γ-rays. However, thick walls or buildup caps may significantly attenuate low energy x- and γ-rays. Many ion chamber survey meters have windows that may be opened in the thick material around the ion chamber to permit more accurate measurement of low energy x- and γ-rays. Electronic equilibrium, also called charged particle equilibrium, is discussed in detail in more advanced texts (Attix, 1986; Knoll, 2010).

Unlike ion chambers, which can function with almost any gas, including air, a proportional counter must contain a gas with low electron affinity, so that few free electrons become attached to gas molecules. Because gas multiplication can produce a charge-per-interaction that is hundreds or thousands of times larger than that produced by an ion chamber, proportional counters can be operated in pulse mode as counters or spectrometers. They are commonly used in standards laboratories, in health physics laboratories, and for physics research. They are seldom used in medical centers.

Multiwire proportional counters, which indicate the position of an interaction in the detector, have been studied for use in nuclear medicine imaging devices. They have not achieved acceptance because of their low efficiencies for detecting x-rays and γ-rays from the radionuclides commonly used in nuclear medicine.

GM counters must also contain gases with specific properties, discussed in more advanced texts. Because gas multiplication produces billions of ion pairs after an interaction, the signal from a GM detector requires little additional amplification. For this reason, GM detectors are often used for inexpensive survey meters.

GM detectors have high efficiencies for detecting charged particles that penetrate the walls of the detectors; almost every such particle reaching the interior of a detector is counted. Many GM detectors are equipped with thin windows to allow beta particles and conversion electrons to reach the gas and be detected. Very weak charged particles, such as the beta particles emitted by tritium (³H, E_max = 18 keV), which is extensively used in biomedical research, cannot penetrate the windows; therefore, contamination by ³H cannot be detected with a GM survey meter. Flat, thin-window GM detectors, called “pancake”-type detectors, are very useful for finding radioactive contamination (Fig. 17-7).

In general, GM survey meters are very inefficient detectors of x-rays and γ-rays, which tend to pass through the gas without interaction. Most of those that are detected have interacted with the walls of the detectors, with the resultant electrons scattered into the gas inside the detectors.

The size of the voltage pulse from a GM tube is independent of the energy deposited in the detector by the interaction causing the pulse: an interaction that deposits 1 keV causes a voltage pulse of the same size as one caused by an interaction that deposits 1 MeV. Therefore, GM detectors cannot be used as spectrometers or precise dose-rate meters. Many portable GM survey meters display measurements in units of
milliroentgens per hour. However, the GM counter cannot truly measure exposure rates, and so its reading must be considered only an approximation. If a GM survey meter is calibrated to indicate exposure rate for 662-keV γ-rays from ¹³⁷Cs (commonly used for calibrations), it may overrespond by as much as a factor of 5 for photons of lower energies, such as 80 keV. If an accurate measurement of exposure rate is required, an air-filled ionization chamber survey meter should be used.

This overresponse of a GM tube to low-energy x-rays and γ-rays can be partially corrected by placing a thin layer of a material with a moderately high atomic number (e.g., tin) around the detector. The increasing attenuation coefficient of the material (due to the photoelectric effect) with decreasing photon energy significantly flattens the energy response of the detector. Such GM tubes are called energy-compensated detectors. The disadvantage of an energy-compensated detector is that its sensitivity to lower energy photons is substantially reduced and its energy threshold, below which photons cannot be detected at all, is increased. Energy-compensated GM detectors commonly have windows that can be opened to expose the thin tube walls so that high-energy beta particles and low-energy photons can be detected.

GM detectors suffer from extremely long dead times, ranging from tens to hundreds of microseconds. For this reason, GM counters are seldom used when accurate measurements are required of count rates greater than a few hundred counts per second. A portable GM survey meter may become paralyzed in a very high radiation field and yield a reading of zero. Ionization chamber instruments should always be used to measure high intensity x-ray and γ-ray fields.

Scintillators are materials that emit visible light or UV radiation after the interaction of ionizing radiation with the material. Scintillators are the oldest type of radiation detectors; Roentgen discovered x-radiation and the fact that x-rays induce scintillation in barium platinocyanide in the same fortuitous experiment. Scintillators are used in conventional film-screen radiography, many direct digital radiographic image receptors, fluoroscopy, scintillation cameras, CT scanners, and positron emission tomography (PET) scanners.

Although the light emitted from a single interaction can be seen if the viewer’s eyes are dark adapted, most scintillation detectors incorporate a means of signal amplification. In conventional film-screen radiography, photographic film is used to amplify and record the signal. In other applications, electronic devices such as photomultiplier tubes (PMTs), photodiodes, or image-intensifier tubes convert the light into electrical signals. PMTs and image-intensifier tubes amplify the signal as well. However, most photodiodes do not provide amplification; if amplification of the signal is required, it must be provided by an electronic amplifier. A scintillation detector consists of a scintillator and a device, such as a PMT, that converts the light into an electrical signal.

When ionizing radiation interacts with a scintillator, electrons are raised to an excited energy level. Ultimately, these electrons fall back to a lower energy state, with the emission of visible light or UV radiation. Most scintillators have more than one mode for the emission of visible light or UV radiation, and each mode has its characteristic decay constant. Luminescence is the emission of light after excitation. Fluorescence is the prompt emission of light, whereas phosphorescence (also called afterglow)
is the delayed emission of light. When scintillation detectors are operated in current mode, the prompt signal from an interaction cannot be separated from the phosphorescence caused by previous interactions. When a scintillation detector is operated in pulse mode, afterglow is less important because electronic circuits can separate the rapidly rising and falling components of the prompt signal from the slowly decaying delayed signal resulting from previous interactions.

It is useful, before discussing actual scintillation materials, to consider properties that are desirable in a scintillator.

In all scintillators, the amount of light emitted after an interaction increases with the energy deposited by the interaction. Therefore, scintillators may be operated in pulse mode as spectrometers. When a scintillator is used for spectroscopy, its energy resolution (ability to distinguish between interactions depositing different energies) is primarily determined by its conversion efficiency. A high conversion efficiency is required for superior energy resolution.

There are several categories of materials that scintillate. Many organic compounds exhibit scintillation. In these materials, the scintillation is a property of the molecular structure. Solid organic scintillators are used for timing experiments in particle physics because of their extremely prompt light emission. Organic scintillators include the liquid scintillation fluids that are used extensively in biomedical research. Samples containing radioactive tracers such as ³H, ¹⁴C, and ³²P are mixed in vials with liquid scintillators, and the light flashes are detected and counted by PMTs and associated electronic circuits. Organic scintillators are not used for medical imaging because the low atomic numbers of their constituent elements and their low densities make them poor x-ray and γ-ray detectors. When photons in the diagnostic energy range do interact with organic scintillators, it is primarily by Compton scattering.

There are also many inorganic crystalline materials that exhibit scintillation. In these materials, the scintillation is a property of the crystalline structure: if the crystal is dissolved, the scintillation ceases. Many of these materials have much larger average atomic numbers and higher densities than organic scintillators and therefore are excellent photon detectors. They are widely used for radiation measurements and imaging in radiology.

Most inorganic scintillation crystals are deliberately grown with trace amounts of impurity elements called activators. The atoms of these activators form preferred sites in the crystals for the excited electrons to return to the ground state. The activators modify the frequency (color) of the emitted light, the promptness of the light emission, and the proportion of the emitted light that escapes reabsorption in the crystal.

No one scintillation material is best for all applications in radiology. Sodium iodide activated with thallium [NaI(Tl)] is used for most nuclear medicine applications. It is coupled to PMTs and operated in pulse mode in scintillation cameras, thyroid probes, and γ-well counters. Its high content of iodine (Z = 53) and high density provide a high photoelectric absorption probability for x-rays and γ-rays emitted by common nuclear medicine radiopharmaceuticals (70 to 365 keV). It has a very high conversion efficiency; approximately 13% of deposited energy is converted into light. Because a light photon has an energy of about 3 eV, approximately one light photon is emitted for every 23 eV absorbed by the crystal. This high conversion efficiency gives it a very good energy resolution. It emits light very promptly (decay constant, 250 ns), permitting it to be used in pulse mode at interaction rates greater than 100,000/s. Very large crystals can be manufactured; for example, the rectangular crystals of one modern scintillation camera are 59 cm (23 inches) long, 44.5 cm (17.5 inches) wide, and 0.95 cm thick. Unfortunately, NaI(Tl) crystals are fragile; they crack easily if struck or subjected to rapid temperature change. Also, they are hygroscopic (i.e., they absorb water from the atmosphere) and therefore must be hermetically sealed.

PET, discussed in Chapter 19, requires high detection efficiency for 511-keV annihilation photons and a prompt signal from each interaction because the signals must be processed in pulse mode at high interaction rates. PET detectors are thick crystals of high-density, high atomic number scintillators optically coupled to PMTs. For many years, bismuth germanate (Bi₄Ge₃O₁₂, often abbreviated as “BGO”) was the preferred scintillator. The high atomic number of bismuth (Z = 83) and the high density of the crystal yield a high intrinsic efficiency for the 511-keV positron annihilation photons. The primary component of the light emission is sufficiently prompt (decay constant, 300 ns) for PET. NaI(Tl) was used in early and some less-expensive PET scanners. Today, lutetium oxyorthosilicate (Lu₂SiO₄O, abbreviated LSO), lutetium yttrium oxyorthosilicate (Lu_xYSiO₄O, abbreviated LYSO), and gadolinium oxyorthosilicate (Gd₂SiO₄O, abbreviated GSO), all activated with cerium, are used in newer PET scanners. Their densities and effective atomic numbers are similar to those of BGO, but their conversion efficiencies are much larger and they emit light much more promptly.

Calcium tungstate (CaWO₄) was used for many years in intensifying screens in film-screen radiography. It was largely replaced by rare-earth phosphors, such as gadolinium oxysulfide activated with terbium. The intensifying screen is an application of scintillators that does not require very prompt light emission, because the film usually remains in contact with the screen for at least several seconds after exposure. Cesium iodide activated with thallium is used as the phosphor layer of many indirect-detection thin-film transistor radiographic and fluoroscopic image receptors, described in Chapters 7 and 9. Cesium iodide activated with sodium is used as the input phosphor and zinc cadmium sulfide activated with silver is used as the output phosphor of image-intensifier tubes in fluoroscopes.

Scintillators coupled to photodiodes are used as the detectors in CT scanners, as described in Chapter 10. The extremely high x-ray flux experienced by the detectors necessitates current mode operation to avoid dead-time effects. With the rotational speed of CT scanners as high as three rotations per second, the scintillators used in CT must have very little afterglow. Cadmium tungstate and gadolinium ceramics are scintillators used in CT. Table 17-1 lists the properties of several inorganic crystalline scintillators of importance in radiology and nuclear medicine.