Search

Systematic measurements of dark noise spectra of CdZnTe x- and γ-ray spectrometers, correlated with the dc I-V characteristics and detector technology, are reported. The dark noise of two innovative CdZnTe spectrometer configurations are studied: metal-semiconductor-metal (MSM) resistive detectors with three terminals as well as heterostructure PIN detectors with thermally evaporated n+ CdS and p+ ZnTe contacts, which are fabricated on high pressure Bridgman CdZnTe (Zn=10%) crystals. The two innovative CdZnTe spectrometer configurations presented here exhibit very low dark (leakage) current. By reducing the dc value of the dark (leakage) current below 1 nA, shot noise becomes the dominant noise mechanism and the contribution of 1/f noise becomes negligible. The use of non-injecting contacts (evaporated gold) for the MSM detectors and the operation of the PIN detector in the reverse bias mode prevent generation-recombination noise which becomes dominant with injecting contacts (for example MSM detectors with evaporated indium and titanium contacts) or when operating the PIN detector in the forward bias mode. Surface leakage is reduced by applying surface passivation but is eliminated only by using the three terminal MSM configuration which exhibits simple shot noise instead of the suppressed shot noise observed in the two terminal MSM spectrometers. The noise measurements are useful for optimizing detector technology.

The effect on energy resolution of dark noise currents experimentally observed in state of the art CdZnTe spectrometers is quantitatively analyzed. Expressions for the energy resolution determined by shot noise, generation-recombination noise, and 1/f noise are discussed in turn and are compared with the linewidth due to incomplete charge collection. The effect of the dark noise upon energy resolution is directly measured by monitoring the broadening of peaks provided by a pulser, due to biased, non-irradiated CdZnTe spectrometers that are connected in parallel. The noise characteristics of the spectrometers under study were measured separately and their dominant noise mechanism is characterized. In this manner, the effect on resolution of the dark noise is correlated not only with the dc value of the dark current but also with the noise spectral behavior which is determined by the noise mechanism.

This study characterizes, for the first time, contacts to CdZnTe radiation detectors by measuring the dark noise spectra as a function of the applied bias. The noise currents are correlated with the dc dark current-voltage characteristics of CdZnTe x-ray and gamma-ray detectors. In order to identify and separate the role of the contacts in the overall performance, the measured noise phenomena is correlated with detector configuration and contact design as well as the growth method of the CdZnTe crystals, contact technology, and passivation. Several contact technologies (electroless gold, and a number of evaporated metallic contacts including gold, indium, zinc, titanium, aluminum, and platinum contacts) are compared. Contacts to CdZnTe crystals grown by high pressure Bridgman are compared with contacts to CdZnTe crystals grown by modified Bridgman. Contacts of resistive detectors as well as of Schottky detectors are reported. Large area symmetric contacts are compared with small area pixelized contacts. The role of the metallization used for contacts, the role of surface effects and passivation, and the role of contact design are discussed.

The charge collection efficiency of CdZnTe radiation detectors with two different configurations: aSchottky diode detector and aresistive detector are compared. The average charge collection efficiencies for three different directions of irradiation (negative electrode, positive electrode and perpendicular to the electric field) are calculated. The mobility-lifetime product of the CdZnTe substrates is evaluated from the dependence of the measured spectra upon detector bias voltage. The measurement of the average charge collection efficiency is based on monitoring the shift of the peak channel with bias voltage in an experimental setup which is well calibrated. Two types of radiation are used:gamma photons from several radioactive sources andalpha particles from an241Am source. The models for the evaluation of mobility-lifetime product from the measured data for the two types of detector configurations as well as for the two types of radiation sources, are compared and discussed. The CdZnTe (Zn = 10%) substrates under study are obtained commercially and are grown by the high pressure Bridgeman method. The mobility-lifetime products and specific resistivity of the two types of detectors are evaluated and compared. A lower resistivity material has a narrower depletion region and behaves like a thinner detector thus exhibiting better collection efficiencies. Therefore, medium resistivity material which is completely inadequate for resistive detectors can still yield high performance Schottky detectors. The preferred direction of irradiation, i.e. from the negative electrode, is possible only in the case of n-type material which is reverse biased by negative voltages applied to the Schottky gate. The mobility-lifetime products that are derived on both the resistive detector (with specific resistivity of ≈1.1010 ω.cm) and the Schottky diode (with specific resistivity of ≈1.106 Ω.cm) are μnτn ≈-4.10−4 cm2V−1 and μpτp≅ 8.10−5 cm2V−1.

In this paper, we report on the design, fabrication, and performance of the first CdZnTe Schottky photodiode arrays for radiation detection. High pressure Bridgman-grown CdZnTe substrates with bulk resistivities in the range 108 to 1010 ohm-cm were used. CdZnTe Schottky photodiodes were formed with In and Ti/Au contacts. Diode arrays with pixel sizes from 1000 × 1000 (im to 100 × 100 urn were fabricated. The diode’s I-V characteristics exhibited low leakage current and high bulk resistivity; leakage current decreased as diode pixel size was reduced. Response of these detector arrays to high energy photons was uniform and their energy resolution improved with smaller pixel size.

The technology to control the interface charge density between CdTe substrates and CdTe epilayers grown by metalorganic chemical vapor deposition is studied. The interface charge is determined by the modified built-in potential derived from capacitance-voltage characteristics of Schottky contacts formed on the epilayers. Novel ultraviolet (UV) photon assisted and photo thermal surface pretreatments that control the interface between p-type CdTe substrates and CdTe epilayers are reported. The substrates are exposed to UV radiation provided by a high pressure Hg lamp operating at 600 W with a wide emission spectrum between 190 and 300 nm. The UV photon assisted surface pretreatment with hydrogen is compared with additional surface pretreatments: thermal pretreatment with hydrogen (without UV photons) and UV photo thermal pretreatment with hydrogen. The UV photon assisted and the UV photo thermal surface pretreatments with hydrogen reduce the interface charge density to a practically negligible value. In addition, the p-type doping level of the substrate is reduced considerably in a layer of few microns adjacent to the interface.

Photon assisted metalorganic chemical vapor deposition (MOCVD) of HgTe by the reaction of mercury and diethyltelluride (DETe) on CdTe substrates is reported. The growths were done at 380–450°C and subatmospheric pressure, with helium as the carrier gas. A simple kinetic model for the MOCVD growth is presented. The model yields the growth rate as a function of the gas-phase partial pressures of the constituents in various limiting cases, and is corroborated by experimental results. The apparent activation energy for growth without photons (33–49 kcal/mol) exceeds the activation energy of pyrolysis of DETe (∼25 kcal/mol) and is significantly reduced in photon assisted growth (18–28 kcal/mol). The effect of photon assisted growth on morphology, thickness uniformity, and location along the susceptor is also presented and discussed.

A kinetic model for the metalorganic chemical vapor deposition (MOCVD) growth of CdTe over a wide temperature range is presented. The model yields the growth rate as a function of the gas-phase concentrations of the constituents. The model is corroborated with experimental results obtained by the MOCVD growth of CdTe at 380° C. The major features of the model are the observed two-step surface-controlled pyrolysis and surface saturation, leading initially to a growth rate that increases with the square root of the concentrations of the reacting species and subsequently to a decrease of the growth rate as the concentrations increase. At even higher concentrations, an additional increase of growth rate is observed and modeled.