Linear x-ray detector array made on bulk CdZnTe for 30~100 keV energy

Journal of Electronic Materials, Jun 1997 by Yoo, S S, Jennings, G, Montano, P A

The uniformity of the linear array dark current was measured for 16 test elements. A bias of 100 V was applied to the bottom common electrode, and the current was measured from each element. During the measurement, the adjacent elements were grounded in order to avoid charging effect, which results in surface leakage current from neighboring elements. Excellent uniformity of both dark and generation currents was obtained as shown in Fig. 4. The average dark current for 175 x 800 (mu)m2 area element is 120 pA with 10% standard deviation. The generation current, obtained by exposure to an optical light source, was increased by nearly two and half orders of magnitude. However, with a negative bias, a slight increase of the generation current due to hole collection was observed.

The energy spectrum was measured for a test element, and the result is shown in Fig. 5. The photopeaks at 122 and 136 keV are distinctively seen in the spectrum. However, high counts in the energy below 55 keV energy severely distort the energy spectrum that is known to be a charge sharing effect and commonly seen in small size microstrip detectors.16,17 The carriers are generated in the entire volume ofthe CdZnTe by the 122 and 136 keV photons of Co^sup 57^. In the case of planar type radiation source, such high energy photons are often randomly projected in slanted angle onto elements, thus resulting in successive photon interaction along the elements. Then the generated charges would be distributed into several elements. Even if the photon interaction occurs in a single element, the created charges tend to diffuse into adjacent elements of the small pitch size array. Similarly, escaped x-ray photons are known to contribute to the charge sharing.

In spite of the increased counts in the low energy of the spectrum, the 14.4 keV photopeak was observed. An aluminum filter was inserted between the source and the array detector. A 1.5 mm thick aluminum foil should greatly attenuate the 14.4 keV photons, whereas the attenuation should be very small for the high energy photons. As a result, the count ratio of the 14.4 keV photopeak should decrease without modification of the spectrum at higher energy. As seen in the figure, the measured spectrum show a slight decrease in counts at 14.4 keV.

The photopeak full width at half maximum (FWHM) at 122 keV was fitted using a Gaussian distribution function. In spite of the presence of a ghost peak near 110 keV, the Gaussian function was well fitted to the photopeak within a minimal deviation, and we obtained 5.7% FWHM at the photopeak. This value is unexpectedly low considering the quality of the CdZnTe material used for the array. In Fig. 6, the energy resolution was measured using the same material but with a large area detector. At 14.4 keV, the photopeak was fitted with a symmetric Gaussian function and a 4.8 keV FWHM was obtained. At this energy, the photon interaction occurs near the negatively bias electrode. Therefore, hole trapping becomes trivial, and the resolution is limited by the trapping of electrons drifted to the opposite electrode. However, at 122 keV, the measured photopeak is asymmetric and deviates from the 4.8 keV FWHM Gaussian function in the low energy slope. In this case, low mobility hole trapping limits the energy resolution.