The ability to operate at room temperature without cryogenic cooling has made cadmium zinc telluride (CdZnTe) nuclear detectors popular in X-ray detection and gamma-ray spectroscopy applications      . The absence of the need to cool the detector gives the advantage of reduced operational costs and the production of hand-held devices, when compared to detectors that need cooling, such as germanium-based detectors. The major application areas for CdZnTe detectors include national security, environmental protection.
Imaging, and astrophysics    . The fabrication of CdZnTe detectors involve cutting wafers from the ingot, mechanically polishing the wafers, chemical cleaning and treatment, and deposition of electrical contacts. The chemical cleaning process could be etching or chemo-mechanical polishing. These processes are used to remove surface defects that are left after the mechanical polishing of the wafers. The chemical treatment process involves passivation to make the wafer surfaces more chemically stable     .
While there have been many studies on the chemical treatment of CdZnTe detector, very little has been done on using the same chemical to compare the processes of etching and chemo-mechanical polishing  . In a recent study  , we compared the effects of chemical etching and chemo-mechanical polishing on the electrical properties of CdZnTe detectors using bromine-methanol-ethylene glycol. The chemo-mechanical polishing process was observed to contribute lesser surface current to the measured current while the chemical etching process gave a higher charge-carrier mobility—lifetime product and radiation detection energy resolution  . The new investigation in the present study covers the effects on the tellurium and tellurium-oxide species on the surface of the CdZnTe wafers. The major differences between the previous study  and the present investigation include the chemical used, and the measurements of Te and TeO2 species on the surfaces of the wafers. The previous study used bromine-methanol-ethylene glycol  . In the present work, we used a solution of hydrogen bromide in hydrogen peroxide and ethylene glycol mixture. The Te and TeO2 species on the surfaces of the wafers were measured using X-ray Photoelectron Spectroscopy (XPS).
2.1. Preparation of Samples
The CdZnTe used in this study was obtained from eV Products. It was grown by the vertical Bridgman method    . A special cutting machine equipped with a diamond wire saw was used to cut two samples (Sample-1 and Sample-2) from a larger CdZnTe wafer. The samples were polished mechanically with 800-grit silicon carbide abrasive paper. Subsequent mechanical polishing was carried out with 1000-grit and 1200-grit silicon carbide abrasive papers. Further smoothening of the surfaces was accomplished through polishing on MultiTex pad with alumina (Al2O3) powder and distilled water. Successive polishing with decreasing sizes of alumina powder from 3.0 µm to 0.1 µm was carried out. The dimensions of Sample-1 and Sample-2 after mechanical polishing are 4.2 × 3.8 × 1.7 mm3 and 4.0 × 3.4 × 1.6 mm3 respectively.
Electrical contacts are needed for some of the experiments: current-voltage (I-V) measurements, charge transport measurements, and detector response measurements. For these experiments gold electrical contacts were deposited on the opposite large sides of each sample. The electroless deposition method was used for the gold contacts by pipetting drops of 5% gold chloride (AuCl3) solution on each planner surface. After reaction with the surface to form the contacts, excess gold chloride solution is removed from the surfaces using a felt paper to absorb the solution  .
2.2. Characterization Instruments and Experiments
After mechanical polishing, the samples were thoroughly rinsed with distilled water and dried using compressed air. This is followed by characterization experiments that include XPS, current-voltage (I-V) measurements, and detector response measurements. The first set of measurements were used to collect the data on the samples before chemical etching and chemo-mechanical polishing processes.
Prior to the chemical etching and chemo-mechanical polishing, the gold contacts were removed by mechanical polishing using alumina powder from 3.0 µm to 0.1 µm. After rinsing in distilled water and drying with compressed air, Sample 1 was chemo-mechanically polished on a special felt pad using the same chemical solution. Sample 2 was chemically etched by dipping and stirring it in a solution of hydrogen bromide in hydrogen peroxide and ethylene glycol mixture for about 2 minutes. The chemo-mechanical polishing lasted for about 1.5 minutes for each of the two large surfaces. The set of characterization experiments were then carried out for the samples.
The XPS equipment is an RHK Technology UHV 7500 system. It has an ultrahigh vacuum chamber with a pressure below 8 × 10−10 Pa. The system is equipped with an Al/Mg X-ray source. We scanned for energy peaks of cadmium (Cd), tellurium (Te) and tellurium oxide (TeO2).
A customized aluminum box, equipped with a Keithley picoammeter and voltage source was used for the I-V measurements. We applied voltages from −100 V to 100 V and recorded the corresponding current. A special sample holder by eV Products was used in the experiment for recording the response of the CdZnTe samples to the 59.5 gamma line of sealed 241Am radiation source. The sample holder is made of brass and has a beryllium window. The sample holder is connected to a multichannel analyzer (MCA) via a pre-amplifier and an amplifier. A computer displays and record the datafrom the MCA. The 241Am spectra for the two samples were recorded at an applied voltage of 100 V.
3. Results and Discussion
3.1. XPS Analysis of TeO2 on CdZnTe Surfaces
The dominant surface species on CdZnTe wafers are cadmium (Cd), tellurium (Te) and tellurium oxide (TeO2). This study focused on Te and TeO2 species. The XPS results are in Figure 1 show the Te and TeO2 peaks. The Te elemental states for the 3d3/2 and 3d5/2 doublets binding energies appear approximately at 583.5 eV and 573.1 eV respectively  . The TeO23d3/2 and 3d5/2 doublets appear approximately at 586.5 eV and 576.1 eV respectively  . These binding energies were reported by Egarievwe et al.  with an estimated error of ±0.4 eV, and by Bahl et al.  with and estimated error of ±0.2 eV.
Figure 1(a) show that more TeO2 are formed after chemical etching as indicated by the Te3d3/2O2 and Te3d5/2O2 peaks. In contrast, the Te3d3/2O2 and Te3d5/2O2 peaks slightly decreased for the chemo-mechanical polishing process as shown in Figure 1(b). The importance of these results is in the effect of TeO2 on surface leakage current and noise in CdZnTe detectors.
The formation of TeO2 reduces the surface leakage current   , and hence reduces the detector noise  thereby improving the energy resolution of the detector  . The etching process using the same chemical (solution of hydrogen bromide in hydrogen peroxide and ethylene glycol mixture) resulted in the formation of more TeO2 compared to the chemo-mechanical polishing process. Figure 1(a) is from the same set of data as the XPS spectra reported in  where Sample-2 was used in a previous experiment as Wafer-1.
3.2. Leakage Current
The I-V curves for the CdZnTe wafers are shown in Figure 2. The electrical resistivity for each sample is of the order of 1010 Ω-cm. The resistivity ρ was calculated from the slope of the I-V curve, the thickness of the wafer L, and the area A of the gold contacts as follow:
where R is resistance and ΔI/ΔV is the slope.
An increase in current for each applied voltage was observed for the two wafers that a freshly processed (I-V measurement were made immediately after each process: chemicaletching and chemo-mechanical polishing). The chemo-mechanical polishing process increased the leakage current more that the chemical etching. This could be seen more clearly from Figure 3. These results are for measurements taken immediately after processing (about 1 to 2 hours). The currents after the surface treatments usually decrease towards that of the mechanical polishing after several days   . The measured current comes from the combination of the bulk current and the surface current. The bulk
(a) Chemical etching (b) Chemo-mechanical polishing
Figure 1. XPS spectra showing Te and TeO2 peaks: (a) Chemically etched after mechanical polishing (from the same data set repoted in ), and (b) Chemomechanically polished after mechanical polishing.
(a) Chemical etching (b) Chemo-mechanical polishing
Figure 2. Current-voltage characteristics of the samples: (a) Chemically etched (from the same data set reported in  ), and (b) Chemomechanically polished, after mechanical polishing.
current depends on the CdZnTe material. The surface current depends on the surface composition and the detector-contact interface  . The observed changes in the measured current come from the changes in the surface current because of the chemical treatments  , in this case, chemical etching and chemo-mechanical polishing.
3.3. Detector Resolution
The spectral response measurements are shown in Figure 4 and Figure 5 for the 59.5 keV gamma line of 241Am. The energy resolution immediately after chemical etching went from 21% to 32% full-width-at-half-maximum (FWHM). The FWHM immediately after chemo-mechanical polishing increased from 18% to
Figure 3. Increase in current, ΔI, for the chemical etching (Sample-1) and chemo-mechanical polishing (Sample-2).
(a) Mechanically polished (b) Etched after mechanical polishing
Figure 4. Response of Sample-1 to the 59.5-keV gamma line of 241Am at 100 V: (a) mechanically polished, and (b) chemically etched after mechanical polishing (from the same data set reported in  ).
49%. The increase in FWHM for the freshly chemically etched sample is 52% compared to 172% for the freshly chemo-mechanically polished sample. Thus, the etching process is more detrimental to the energy resolution than the chemo-mechanically polishing. This observation is similar to results reported for CdZnTe detectors treated with bromine-methanol-ethylene glycol  . The reduction in energy resolution agrees with the increase in surface current which causes increase in detector noise. Increase in detector noise lowers the energy resolution. It is expected that the energy resolutions will improve as the treated surfaces stabilize and the measured leakage currents reduces.
(a) Mechanically polished (b) Etched after mechanical polishing
Figure 5. Response of Sample-2 to the 59.5-keV gamma line of 241Am at 100 V: (a) Mechanically polished, and (b) Chemo-mechanically polished after mechanical polishing.
The effect of etching and chemo-mechanical polishing processes on CdZnTe nuclear detectors was studied using the same chemical solution (hydrogen bromide in hydrogen peroxide and ethylene glycol mixture). The XPS experiments showed that the chemical etching process resulted in the formation of more TeO2 on the detector surfaces compared to chemo-mechanical polishing. The electrical resistivity of the detector, obtained from the slope of the I-V curve, the wafer thickness, and the gold contact area is of the order of 1010Ω-cm. It was observed that the chemo-mechanical polishing process increased the leakage current more that chemical etching.
The etching process, for freshly treated surfaces, is more detrimental to the energy resolution than chemo-mechanically polishing. The currents after the surface treatments are expected to decrease towards that of the mechanical polishing after several days   . This should lead to improved energy resolutions of the CdZnTe wafers. Further studies on aging of the treated samples for longer periods are needed to understand the long-term effects of these processes.
This research was funded by the US Department of Homeland Security, Domestic Nuclear Detection Office, under competitively awarded contract/IAA award number 2012-DN-077-ARI065-05; the US Department of Energy, Office of Defense Nuclear Nonproliferation Research and Development, DNN R&D (NA-22); the US Nuclear Regulatory Commission through award number NRC-27-10-514; and the National Science Foundation (NSF) HBCU-UP Program through award number 1818732.