Room-temperature semiconductor detectors for in vivo monitoring of internal contamination.

In vivo monitoring of low-energy X-ray and gamma-ray emitters has always been a difficult task, primarily because of lack of accuracy and the high detection limits of classical techniques. Various types of PIN diodes (diodes with a large intrinsic zone) were tested in the Radiation Protection Department of the Studie Centrum voor Kernenergie, Centre d'étude de l'Energie Nucléaire (Mol, Belgium) in the measurement of radioactive body burden by direct methods. Current research is oriented toward the use of room-temperature diodes for the detection of low-energy photons escaping the body. In this paper, a new counting technique that involves a portable jacket containing the diodes is described. The system uses silicon diodes and is used out of shielding room in order to be near the contamination. With this method rapid analysis and long counting times are possible, stress is reduced, and medical treatment can be optimized. CdZnTe detectors were also evaluated for this measurement technique but this type of detector is better adapted for counting inside a shielding room. The improvement of the accuracy of the measurement, taking into account the effect of the ribs, is described here, as well the associated electronics necessary for this type of counting.


Introduction
Measurement of low-energy 7-ray emitters in the body by the direct method always has been difficult because of the strong attenuation of the photon flux in different tissues. For this reason measurement is possible only in a limited number of organs and tissues e.g., in the lungs, the thyroid, and in wounds. A good assessment is not possible using standard techniques. The use of a large detector with a high efficiency such as a phoswich (a NaI [Tl] optically associated with a CsI [Tl]) is only valuable when the contamination in the lungs is homogeneously distributed and distribution is similar to the activity distribution in the phantom This paper is based on a presentation at the International Conference on Radiation and Health held 3-7 November 1996 in Beer Sheva, Israel. Abstracts of these papers were previously published in Public Health Reviews 24 (3)(4): 205-431 (1996). Manuscript received at EHP 11 March 1997; accepted 30 June 1997.

cot@sckcen.b
Abbreviations used: Bq, becquerel; CZT, CdZnTe; HPGe, hyperpure germanium; keV, kiloelectron volt. used for calibration. Indeed, a detector that only measures background will increase the detection limits without bringing information. For example consider two small diodes placed on a lung; only one diode sees a signal S1. The other measures a zero net signal (S2 = 0). If we consider only diode 1 for the calculation, the result will be correct. If we consider both diodes, which together have a higher efficiency and which measure the same net signal (S1+ 0), the final result will underestimate the burden. The purpose of this study is to show that several small diodes can be used to answer different questions related to internal contamination. A second goal of the research is to show that two types of available diodes are best suited for in vivo measurements: room-temperature silicon diodes, which allow certain types of measurement of 241Am in the lungs out of a shielding room by long-duration counting, and CdZnTe (CZT) diodes, which can be used in groups to reach detection limits obtained with classic hyperpure germanium (HPGe) detectors to assess 24'Ar contamination of the lungs. These diodes are useful for in vivo detection of 1251, 103Pd, 109Pd, 133Ba, and even 137Cs.

Materials and Methods
In a previous study (1), different types of silicon PIN diodes (diodes with a large intrinsic zone) were compared for their efficacy in whole-body counting. Research in this study was limited to two types of diodes: a 500-pm thick Hamamatsu S3590-05 silicon PIN diode (Hamamatsu, Japan) and a 1-mm thick CZT detector (model eV 180-9-8-4-SD from eV Products, Saxonburg, PA). The objective was to determine the parameters of the counting devices (e.g., volume of the diode, thickness of the depleted zone, bias voltage, diode containment) that would optimize the counting capabilities of these two diodes in specific contamination cases. The experimental work was aided by a home-made humanoid torso phantom containing a true thoracic cage with two foam volumes to simulate the lungs. A photograph of the torso, with the front Lucite plate removed to show the thoracic cage, is presented in Figure 1. Although manipulation of this device is difficult, especially when radioactive charges must be changed, its advantages are imp6rtant: It has a true (corrected) skeleton; the cartilage is represented (with leather); it contains 40K homogeneously distributed; and it is possible to simulate all types of radioactive charge distribution. The absorption parameters of this phantom were controlled by loading it with different charges of 241Am homogeneously distributed and by measuring the phantom with a germanium measuring device that is currently used for routine measurements (2). Correction of this phantom was required because the skeleton portion does not contain blood or bone marrow and does not absorb a 60-kv photon in the same way that a living one does. For this reason, the ribs have been covered with 0.2 mm tin foil to give them the correct absorption properties. Figure 2 displays a vertical scan of the torso measured with the CZT detector after the placement of tin foil on each rib. This graphic, compared to the calculated scan (1), indicates that the phantom is correctly adapted to simulate an actual thoracic cage. The abscissas of this graph represent the vertical displacement in meters with respect to an arbitrarily chosen level. The peaks on each curve represent an interrib position where absorption is minimal.
Comparison of the various curves corresponding to different energy zones indicates the importance of diffusion of the photons in tissues of the body. This is why it is necessary to consider a broad range of energy and why the resolution of the diode is of secondary importance. The bias voltages were chosen to reach full depletion of the junctions (125 and 10OV, respectively, for silicon and CZT instead of the recommended 100 and 60V). This technique allows a higher efficiency and a stable response and does not affect the resolution or the lifetime of the semiconductor even if an increase of the leakage current intensity is observed. Table 1 presents detection limits for two cases: 125I in the thyroid and 24A'Am in the wound. Measurements were performed with three different silicon diodes: a Hamamatsu S3590-05 diode with a 0.81-cm2 active area and a 500-pm depletion layer; a passivated implanted planar silicon diode PD-16-300-AB with a 4.6-cm2 area and a 300-pm depletion area, kindly lent by Canberra Semiconductor, Olen, Belgium; and a SFH217 PIN diode (Siemens, Nixdorf, Germany) with a 0.0095-cm2 area and a 200-pm depletion thickness). Even when they are used as a single diode both detectors are suitable for special types of counting, e.g., wound and thyroid, thus meeting the recommendations of the International Committee on Radiological Protection (3) and various regulatory requirements. The CZT diode was also tested for measurement of contaminated wounds, as described later in "Applications." Lung Measurement As mentioned previously (1), the counting of lungs contaminated with 241'Am must be conducted with an array of diodes to improve efficiency and keep the counting duration within an acceptable limit for human controls (maximum 1 hr) when counting in a counting room. The results of this study can be summarized in two points. First, the counting of lungs with CZT diodes can bring the detection limits equivalent to germanium detectors in a 1-hr counting time if the measurement is performed inside a shielding room and if the number of CZT diodes is increased to 48 units. Second, when silicon PIN crystals are used, measurements could be achieved outside of a shielding room. Thus the counting time can be raised to 24 hr or more (not unreasonable if we consider the use of an electrocardiogram in real time). In such cases the number of diodes could be reduced to about 50 U ( Table 2) to reach the detection limits required by law in certain types of contamination (i.e., pure or high concentration of 241Am).

Wound and Thyroid Measurements
Use ofthe Detector Jacket with Small Diodes The accuracy and efficiency of the measurement can be improved by using an  (1). By using small diodes instead of largevolume detectors it is possible after counting to select the diodes that were actually involved in the counting and reject the diodes measuring only background; this avoids an increase in the detection limits and the consequent actions, as explained in "Introduction." Another advantage of counting outside of a shielding room is the ability to measure controls near the site of an accident in a room with a controlled background, thus reducing the psychological stress for the person waiting for results and subsequent remedial actions.

Associated Electronics
To manage the signals coming randomly from several detectors and send them to a unique spectrometric device, special electronic circuits must be used. Currendy there is no circuit available that is used in conditions adequate for X-and y-ray spectrometry; however, different laboratories are developing such electronics in integrated circuit or in discrete form (4)(5)(6). Two other circuits were tested with positive results. The first circuit (made by Velleman, Brussels, Belgium, catalog no. K8000) is based on I2C interfaces (PCF8591 and PCF8574 from Philips, Eindhoven, Holland). It allows the selection of diodes one at a time among a whole series using a computer program named Select, which we wrote to follow the requirements of the measurement. This device, which has been positively tested, cannot be used for the counting of randomly appearing events (sparse readout), but it can be used to study metabolism by measuring the quantities of radioactive burdens in different organs.
The second circuit uses the mixer routers frequently operational in nuclear spectrometry (model 8222B from Canberra, Olen, Belgium). It contains four analog inputs able to receive signals from four silicon diodes. An example of spectra collected with this system is presented in Figure 3 (counting time = 24 hr), where three S3590-05 silicon PIN diodes have been connected. Figure 3, where the respective backgrounds are also displayed, indicates the different shapes of the collected spectra due to the different positions of the diodes on the phantom. Figure 3 also indicates that diffusion in the tissues produces an important deterioration of the spectra so that the resolution of these diodes is not of [Critical level Lc and detection limit LD are defined according to Currie's concepts: LC = 2.33XNB and LD = 2.71 + 4.65 NB where NB is the total count measured in the investigated zone of the background spectrum.]

Detecton of Contaminaton Hot Spots
After the acquisition time, the sum spectrum (summation of the different spectra) could indicate a light contamination or a null result whereas examination of each spectrum separately might reveal a high concentrated contamination. For this reason, the detector jacket also can be used for detection and localization of hot spots, which is not possible with a large phoswich crystal. Experimental measurements indicate that detection limits are LC/LD= 44/88 Bq for a collection time of 24 hr with silicon diodes and 15/30 Bq for the same counting time with CZT (for measurements outside of a shielding room). An example of spectrum of a hot spot with CZT is shown in Figure 4. In this case, the use of CZT outside of a shielding room is appropriate.

Applications
The devices described here can be used in industry workers to detect eventual high lung contamination by 24i'Am before using classic techniques in shielding rooms. The devices are also useful for the detection of 125i, 103Pd, '09Pd, 133Ba, and 137Cs in wounds or in organs when the organs are not too deep in the body (throat, lungs, liver, and bladder). The best application of these devices is the measurement of skin contamination when the contaminant is a low-energy photon emitter, which is the case in many applications of nuclear medicine e.g., diagnostics and therapy. The tracer can be labeled with one of the radionuclides mentioned, a valuable application in metabolism followup. A single silicon PIN diode is well adapted for counting 125I in thyroid burden or metabolism studies. In our laboratory we assessed the capabilities of the CZT diode measuring 24'Am in a wound using a finger model with high-density polystyrene plates of (C8H8)n-a material equivalent to muscle for absorption and diffusion of X-rays. A 2.5-mm deep incision was made in this plastic finger, in which 15 (± 0.5) Bq 241Am was placed. Measurement with an HPGe crystal resulted in a count of 12 Bq in 10 min, whereas the CZT diode detected 13.7 Bq 24'Am in 6 min. This count assumes that no other y-ray was present from 241Am in the 60 keV energy zone.This technique can be used advantageously instead of one described previously (7). Finally, because they do not require liquid nitro-  in environmental studies by using the first electronic circuit described in this paper.

Discussion and Conclusions
Room-temperature silicon diodes and CZT can be used to replace HPGe detectors in many cases for assessment of actinides in the lungs or other radioactive burdens in organs. The number of diodes should be chosen according to the type of measurement. Because of their size and weight the diodes can be attached to a detector jacket worn by the user. An important advantage of the detector jacket is the elimination of systematic error in the counting because the diodes are fixed on the torso. The close proximity of the diodes with respect to the measured organ increases the counting efficiency. The shielding room is not required with silicon diodes in the energy range of 20 to 100 keV. Long counting times are possible with the jacket; because the diodes are not sensitive to radon variations there are no fluctuations in the background rate in the considered energy range. The counting can start immediately after a suspected contamination, which is an important psychological aspect. Discrete square diodes are preferred to long strip detectors because they can be selected after counting to reduce the background and the detection limits.
The concept of a portable detection jacket can be applied to different types of room-temperature detectors, suggesting applications in radiation protection and nuclear medicine; but electronics for the random readout of an array of diodes has yet to be developed. An example of a design for a detector jacket is shown in Figure 5. Silicon and CZT as single detection elements provide interesting applications in the control of 1251 in the thyroid and 241Am (from mixed oxide fuel particles) in wounds.