Читать книгу Phosphors for Radiation Detectors - Группа авторов - Страница 26

In this section, common characterization techniques for ionizing radiation induced luminescence phenomena are introduced. Transmittance and absorbance measurements are common characterization techniques for phosphors, because we can obtain the absorption bands of materials by this measurement. The PL spectrum is important when considering the emission origin. Figure 1.11a represents a typical methodology to measure transmittance and absorbance. At first, we measure the light transmission intensity of the sample holder, and we define its intensity as 100% transmittance at measured wavelength. Then, we place a sample on the holder, and measure the light transmission intensity of sample + holder. The ratio of the with‐without sample is the transmittance. The transmittance can be understood as Equation (1.26), and if we can plot it in log‐scale corrected by the thickness of the sample, we can draw the absorbance plot. The absorbance is in arbitrary units, and some literature is confused about the absorbance and absorption coefficient (cm⁻¹). It must be understood that if readers need to evaluate the absorption coefficient quantitatively, reflectance must be measured. Without the correction of reflectance, it is impossible physically to determine the absorption coefficient. In some special cases where the reflectance is close to zero, we can treat the absorbance with the rough absorption coefficient. In the case of undoped materials, the wavelength rapidly drops to transmittance to 0% (or rapid increase of absorbance to infinite value), and we can notice the wavelength (energy) as a band‐gap of each material. Some literature use the term “band‐gap” to absorption bands due to dopant in luminescence center doped materials. Such an expression is wrong, at least in solid state physics, although most readers can understand what authors of the paper mean. The transmittance obtained by Figure 1.11a is called in‐line transmittance, and if we use an integration sphere to collect a non‐straight line of light from the sample, we can measure diffuse transmittance. In ionizing radiation detector uses, diffuse transmittance is close to practical conditions, because we generally use reflectors to collect every photon generated by irradiation to photodetectors.

Figure 1.11 Schematic drawing of transmittance (absorption) (a) and PL measurement (b) setups.

Figure 1.11b represents a setup of PL excitation and emission spectra by using the integration sphere. Generally, PL can be observed upon UV–VIS excitation from Xe‐lamp, and we can select which luminescence center (or electron transition) can be induced by selecting the excitation wavelength using a grating. If we use the integration sphere, we can collect all the PL photons. If we define A and I as absorption, which is evaluated by the intensity of excitation light with and without the sample and PL emission intensity, respectively, we can deduce Q = I/A experimentally. The detectors for PL photons are generally grating and photodetector. PL quantum yield is a quantitative value, but generally PL is a qualitative evaluation with an arbitrary unit, since PL intensity depends on the geometry of sample setting. In common PL measurement, we do not use the integration sphere.

Figure 1.12 represents a typical experimental configuration of PL decay measurement. The excitation light was emitted from a pulsed light source such as a lamp or an LED. In the lamp, a grating is used to select excitation wavelength, and selected excitation photons are irradiated to the sample. The timing of the excitation is injected as a start signal to time to amplitude converter (TAC). PL photons from the sample are also filtered by the grating to select the target emission, and then collected by the photodetector. The emitted PL photons are reduced to a single photon by filters, pinhole, etc., and this single photon signal is put into the TAC as a stop signal. A timing window is set for the target timing range (e.g., several tens ns for few ns phenomenon), and only the coincidence event within the timing window is recognized as a target signal. The TAC has a function to convert the timing difference of start and stop signals to a pulse signal. After accumulation of pulse signals many times, we can obtain a PL decay curve. Such a technique is called time correlated single photon counting (TCSPC). Generally, we analyze the PL decay curve by an approximation of sum of exponential functions as

(1.71)

where A_i, t, are τ_i are intensity of each decay component, time, and decay time (lifetime) of each component, respectively. Sometimes, we deconvolute the excitation pulse (instrumental response function, IRF), and it should be noted that perfect deconvolution is generally difficult. After obtaining PL quantum yield Q and decay time τ, we can use convenient equations such as

(1.72)

(1.73)

(1.74)

where k_f and k_nr denote rate constants of radiative and non‐radiative transitions, respectively. In this way, we can evaluate radiative and non‐radiative rate constants quantitatively.

Figure 1.12 Schematic drawing of PL decay setup.

PL is a phenomenon at localized luminescence centers, and except for some cases like the semiconductor scintillator, we do not consider the movement and interactions of carriers. In other words, PL is a phenomenon within a band‐gap. But scintillation is accompanied with interactions of carriers, and sometimes PL and scintillation offer very different emission spectra. As a basic scintillation property, we generally measure radioluminescence spectrum to obtain the scintillation emission intensity to select adequate photodetector for applications. A typical setup is illustrated in Figure 1.13. The most common setup uses an X‐ray generator as an excitation source, and if readers have another radiation source, they can use that instead of the X‐ray generator. We sometimes use α‐rays instead of X‐rays [89] in laboratory‐level experiments, and sometimes use other radiation sources at large facilities [90] by taking our experimental setup there. Scintillation photons are generally collected via an optical fiber, since a direct hit of ionizing radiation causes radiation damage to grating and photodetector, and to avoid this, we generally set them apart from the radiation source. In some literature, light yield is calculated by an integrated area of radioluminescence emission spectrum. It must be noted that such an evaluation is not correct in most cases, as explained in Section 1.3.3. Figure 1.14 shows a geometry of transmission‐type measurement, and in some experiments, we use a reflection type measurement similar to the PL configuration.

Figure 1.13 Common setup of X‐ray induced radioluminescence spectrum measurement.

Figure 1.14 Common setup of γ‐ray induced scintillation decay curve measurement (delayed coincidence method).

Scintillation decay time is generally measured by γ‐ray or X‐ray excitation. Figure 1.14 presents a typical setup for a γ‐ray induced scintillation decay curve measurement. The basic concept is the same as that in PL decay curve measurements but the excitation source is different. Generally, we use the ²²Na radioisotope as an excitation source, which emits two 511 keV γ‐rays in 180° opposite directions. One of these two γ‐rays hits a fast scintillator, and creates a start signal through a constant fraction discriminator (CFD). At the opposite side, the other γ‐ray hits the target scintillator for measurement, and emits scintillation photons. Then, the number of these emitted scintillation photons are reduced to a single photon, and a signal pulse corresponding to the single photon is output from photodetector (typically, PMT). After some delay, the signal is injected into TAC as a stop signal through adequate delay and CFD. The process after TAC is the same with PL decay, and the analysis method is also the same. This typical experimental methodology is called delayed coincidence method (DCM). The difference with PL is that DCM generally does not resolve wavelength because of the low signal intensity and frequency. If we resolve the wavelength, a very long measurement time will be required.

Although DCM with a 511 keV γ‐ray source is the most common way to evaluate scintillation decay time, it contains several technical disadvantages. Because the energy of γ‐ray (511 keV) is high and has a high penetrative power of materials, detection efficiency is not high and requires a long time for the measurement to be made. In addition, low detection efficiency makes it difficult to measure a slow component, and generally, a component slower than several μs is difficult to measure. For example, measurements of emissions from Eu³⁺, Tb³⁺ and most transition metal ions, are almost impossible, although they are used for integration‐type detectors. The lack of the wavelength resolution is also a problem because identification of the scintillation emission origin is sometimes difficult. In most cases, we can guess the emission origin by PL decay, but decays of scintillation and PL sometimes show a large discrepancy.

In order to solve these problems, some groups have developed pulse X‐ray‐based instruments for scintillation decay measurement. Pioneering work was done by the Lawrence Berkeley Laboratory (USA) with the cooperation of Hamamatsu Photonics (Japan) [91]. The key technology is a stable pulse X‐ray source which can be used at laboratory level, and if such a source is developed, we can use the same data processing system with PL decay and DCM. In this pulse X‐ray tube, the origin of the excitation is a laser diode or LED, and emits visible photons with 10–10⁶ Hz frequency, depending on the time range of measurement. Emitted visible photons hit a multi‐alkali photocathode of the X‐ray tube, and are converted into electrons. These electrons are accelerated by high voltage bias, and are lead to a tungsten target by a strong electric field. Then, bremsstrahlung X‐rays are generated and irradiated to the sample through a Be window. In this way, we can make a pulse X‐ray by the timing of laser diode or LED and use it as a start signal.

We now introduce some applications of this pulsed X‐ray tube for scintillation decay measurement. The timing resolution of an X‐ray tube is several tens ps, which is determined by the traveling time of accelerated electrons, and it is enough to measure the rise time of most scintillators [92]. Following these pioneering works, a pulse X‐ray equipped streak camera system was developed, and enabled us to observe the wavelength resolved scintillation decay curve [93]. In this system, the detection part consists of a spectrometer chamber, streak camera, and two‐dimensional CCD, and a two‐dimensional image of wavelength vs. time can be observed. By the streak system, one of the problems of DCM, non‐wavelength resolution, can be solved. Up until the development of the streak camera system, every study used a pulse laser diode as the root of the excitation source, and it was enough to measure fast‐timing events. However, it is not sufficient to measure slow events, especially in the ms time range. One of the solutions for ms measurement is to replace the excitation source to common LED, which has higher power than the pulse‐type laser diode but slower speed. The pulse LED type system was developed in 2014 [94]. This system enabled us to measure slow (ms) scintillation decay as well as the X‐ray induced afterglow. Because the signal intensity per unit time is very low in slower scintillation, the system does not offer the function of wavelength resolution, and the detection part is one PMT. When we compare with typical measurements by DCM, the measurement time dramatically decreases to a few minutes for one sample. At present, no significant difference of excitation energy between 511 keV (DCM) and several tens keV (pulse X‐ray tube) has been found.

In other ionizing radiations such as α‐ray, measurement is generally difficult. One easy but rough way is to use an oscilloscope for direct observation of the signal output from a detector, consisting of scintillator and photodetector by a self‐trigger mode. Although it may contain noise (e.g., thermal noise of photodetector), we can roughly observe scintillation decay time. Previously, most papers observed decay curves by such an oscilloscope measurement, but recently, a digitizer has been used to accumulate many decay curves. In such a case, we can accumulate a large amount of decay curves, and select correct decay curves by setting correct selection conditions of events. An accurate but technically difficult way is to use DCM by one scintillator sample and two photodetectors. If we take the case of α‐ray irradiation as an example, let us assume one cylindrical scintillator optically coupled with two PMTs by edge surfaces, and α‐ray is irradiated to the side of the scintillator. When the α‐ray is absorbed by the center part of the scintillator, half of the scintillation photons are transported to one PMT and the other remaining half to another PMT. We can use one of the two output signals as a start signal, and if we can reduce the number of photons in another side to a single photon, we can use it as a stop signal. In this way, we can measure scintillation decay under particle excitation.

The pulse height spectrum is one of the most important properties of scintillation detectors, because most of measurements in other scientific fields use integration‐type detectors, and pulse height (photon counting) type measurements is one of the original techniques in radiation and high energy physics. The detailed explanation is given in Chapter 12.

Typical experiments on dosimeters are described in Chapters 7 (TSL), 8 (OSL), and 9 (RPL). In most cases, dosimeter properties are examined by X‐ray or γ‐ray irradiation, since an accessibility of high energy photons are easier than other ionizing radiations. Especially, X‐rays are mainly used for tests of individual dosimeters, because energy and amounts of X‐rays can be controlled by bias voltage and tube current in a typical X‐ray generator. In evaluations of dosimeters, the irradiated dose range should be adequate for the purpose. The lower detection limit of commercial individual dosimeters is 1–100 μGy, and experiments should be conducted from these lower detection limits to ~10 Gy. In some papers, several kGy irradiations are done, although the aim is an individual dosimeter. On the other hand, if the purpose is to monitor radiation therapy, tests of several hundreds Gy to several kGy irradiation will be valid. Another important aspect is fading. If the aim is individual dosimeters, fading should be tested at about one month, and if the aim is an imaging plate typically used in a dental clinic, three to four days' tests will be adequate. In addition to these general aspects, in evaluation of TSL properties, heating rate affects peak temperatures in the glow curve. In practical applications, the heating rate is around 10 °C/s, while 0.1–1 °C/s rate is common for studies of TSL as a physical property. In actual materials, we must consider time of heat conduction, and if the heating rate is too fast, responses of materials cannot follow. With a fast heating rate, we generally observe a relatively higher glow peak temperature than the true value, although we do not have a way to know the exact true value. When we measure TSL with a very slow heating rate, an observed glow peak temperature will be close to the true value. Another detailed explanation of the measurement technique of TSL can be found in Chapter 7.

In characterizations of OSL and RPL, we generally use common systems of PL spectrum. In OSL, after the irradiation of ionizing radiation, we put the sample in a PL machine and stimulate the sample under the adequate stimulation wavelength. In some spectrofluorometers, the automatic measurement function of PL excitation spectrum is installed, and by this function, we can easily investigate the stimulation spectrum. Unlike the PL excitation spectrum, which generally shows a clear excitation band due to electron transitions of dopant, OSL stimulation spectrum generally shows a broad unclear feature. For OSL and TSL measurements, an automation type instrument (Risø TL/OSL DA‐20 reader) is widely used ([95]). Generally, this system equips β‐ray (⁹⁰Sr) as an irradiation source, LED as a stimulation source for OSL having typically a 470 nm emission wavelength, and a heater for TSL measurement. One of the disadvantages of this automated system is a lack of wavelength resolution, which equals to non‐determination capability on the emission origin, and some ideas are proposed to add a function of wavelength resolution [96].

In order to investigate OSL and RPL, some important aspects should be noted in actual experiments. Typically, the stimulation wavelength of OSL is longer than the emission wavelength, and resembles the situation of upconversion. In order to distinguish them, continuous stimulation is a simple way. If the phenomenon is OSL, the emission intensity decreases under the continuous stimulation, while in the case of upconversion, the intensity is constant. In addition, the stimulation wavelength of OSL is not necessarily longer than the emission wavelength, because we cannot deny the existence of a deep trap. In such a case, distinction of OSL and RPL becomes difficult. The continuous stimulation (excitation) is also a useful technique in this situation, and if the phenomenon is OSL, the emission intensity decreases with time, while that of RPL is constant against time. But in the special case where the trap depth and excitation band overlap, clear distinction is difficult. Other experimental aspects of OSL and RPL are introduced in Chapters 8 and 9, respectively.