Advances in High-Energy-Resolution CdZnTe Linear Array Pixel Detectors with Fast and Low Noise Readout Electronics

Abstract

Radiation detectors based on Cadmium Zinc Telluride (CZT) compounds are becoming popular solutions thanks to their high detection efficiency, room temperature operation, and to their reliability in compact detection systems for medical, astrophysical, or industrial applications. However, despite a huge effort to improve the technological process, CZT detectors' full potential has not been completely exploited when both high spatial and energy resolution are required by the application, especially at low energies (<10 keV), limiting their application in energy-resolved photon counting (ERPC) systems. This gap can also be attributed to the lack of dedicated front-end electronics which can bring out the best in terms of detector spectroscopic performances. In this work, we present the latest results achieved in terms of energy resolution using SIRIO, a fast low-noise charge sensitive amplifier, and a linear-array pixel detector, based on boron oxide encapsulated vertical Bridgman-grown B-VB CZT crystals. The detector features a 0.25-mm pitch, a 1-mm thickness and is operated at a -700-V bias voltage. An equivalent noise charge of 39.2 el. r.m.s. (corresponding to 412 eV FWHM) was measured on the test pulser at 32 ns peaking time, leading to a raw resolution of 1.3% (782 eV FWHM) on the 59 keV line at room temperature (+20 °C) using an uncollimated ²⁴¹Am, largely improving the current state of the art for CZT-based detection systems at such short peaking times, and achieving an optimum resolution of 0.97% (576 eV FWHM) at 1 µs peaking time. The measured energy resolution at the 122 keV line and with 1 µs peaking time of a ⁵⁷Co raw uncollimated spectrum is 0.96% (1.17 keV). These activities are in the framework of an Italian collaboration on the development of energy-resolved X-ray scanners for material recycling, medical applications, and non-destructive testing in the food industry.

Keywords: CZT; CdZnTe; Gamma-ray spectroscopy; X-ray spectroscopy; nuclear microelectronics; semiconductor radiation detectors.

Figure 1

Geometry of the CZT linear array. An inner guard ring is present that can be connected to a dedicated read-out channel to allow charge sharing corrections.

Figure 2

I–V characteristic of a pixel of the CZT linear array at +25 °C.

Figure 3

Micrograph of a SIRIO charge sensitive preamplifier (650 × 600 µm²) with the bonding-pad pinout. In the inset, a detail of a CZT pixel wire bonded to the preamplifier using a small drop of conductive glue (dark brown circle).

Figure 4

Room temperature energy spectrum acquired with ${}^{55}$Fe calibration source.

Figure 5

Best room temperature spectrum of a ${}^{241}$Am source acquired using Amptek PX5 DPP at the optimum peaking time of 1 µs.

Figure 6

Room temperature spectrum acquired with the ${}^{57}$Co radioactive source. The presence of long left-hand-side tails, as highlighted in the inset of the ${}^{57}$Co is minimized when low-energy photons are absorbed.

Figure 7

Detail of the room temperature energy spectrum acquired with the ${}^{241}$Am radioactive with 32 ns peaking time.

Figure 8

Room temperature energy spectra acquired in a 7-h measurement session, with 1-h acquisition steps.

Figure 9

Measured relative centroid shift and FWHM degradation on the 59.5 keV photo-peak over 7-h acquisition time.

Figure 10

Line widths at room temperature of the SIRIO-CZT system with a ${}^{241}$Am calibration source on the 59.5 keV and 13.9 keV lines with fast peaking-time sweep performed using Dante DPP. Solid blue and green lines are measured FWHM values, while the respective dashed lines are the expected FWHM values summing the pulser FWHM and Fano noise contribution. The estimated series and parallel noise contributions on the pulser FWHM are shown separately with dashed black lines and quadratically summed on the red dashed line. The shot noise associated with the detector current (21 pA) is shown on the bottom-right, estimated evaluating the slope (V/µs) of the ramp at the output of the CSA in absence of incoming photons [45].

Figure 11

Percent contribution of Fano noise (pink), electronic noise (green), and measured excess contribution (purple) on the line FWHM for different photon energies.