Nonstoichiometric deep levels in Mg-doped GaP epitaxial layers: Effects of pre-annealing of substrates

The effects of substrate pre-annealing on deep level density in Mg-doped GaP liquid-phase epitaxy (LPE) layers were investigated by photocapacitance measurement. With annealing under optimum phosphorus-vapor pressure, concentration of deep donor at E_C — 1.9–2.1 eV increased in undoped GaP substrate. Deep level densities in Mg-doped layers were also affected by pre-annealing of the substrates. Densities of dominant deep levels at E_V + 0.85 eV and E_V + 1.5 eV were an order of magnitude reduced and, in contrast, the level at E_C − 2.1 eV in Mg-doped layer increased with long pre-annealing. This level at E_C − 2.1 eV is supposed to be involved with P-rich-type nonstoichiometric defects, such as phosphorus interstitial atoms diffused from the surface of the substrate.     

Nonstoichiometric deep levels in Mg-doped GaP epitaxial layers

We have investigated the relation between deep levels in Mg-doped p-type GaP liquid phase epitaxy (LPE) layers and stoichiometry of the surface of the substrates by PHCAP measurement. Concentration of a deep donor level at E_C−1.9–2.1 eV is higher in an n-type undoped GaP substrate annealed with applying phosphorus vapor pressure of 20 kPa than in sample annealed beneath a carbon cover. Next, Mg-doped LPE layers are grown on substrates that have been pre-annealed under phosphorus vapor pressure just before the growth. The densities of deep levels at E_V+0.85 and E_V+1.5 eV in long-time (2 h) pre-annealing sample are greatly decreased, but a deep level at E_C−1.9–2.1 eV shows opposite tendency. The latter is thought to be identical to a deep level detected in the substrate, probably phosphorus interstitial atoms.     

Diffusion of nonstoichiometric defects in n-GaP crystals

Diffusion of nonstoichiometry-related point defects from a LEC-grown GaP substrate to a Te-doped GaP n-type epitaxial layer was investigated by means of photocapacitance. It was revealed that deep donor level at E_C−2.1 eV was introduced into GaP substrate with annealing under phosphorus vapor pressure. Thus, the 2.1 eV deep level is thought to be involved with excess P atoms such as interstitial phosphorus atoms. In Te-doped crystal, 2.1 eV level was detected and the density increased as the time of substrate annealing increased. By measuring PHCAP spectra of samples with different thickness of epitaxial layer, diffusion profile of the defects from the substrate interface was obtained. From this, the diffusion coefficient at 850°C is estimated to be 8×10^–11 cm²/s.     

Impurity states in cobalt-doped silicon

Cobalt was diffused into p⁺ pn⁺ silicon structures at 900° and 1150°C for 2−4 hours followed by various quenching conditions. Four primary hole traps and two electron traps associated with cobalt in these devices were observed. The hole traps are labeled H1(E_v + 0.22 eV), H2(E_v + 0.29 eV), H3 (E_v + 0.40 eV) and H4(E_v + 0.45 eV) while the electron traps labeled E1 and E2 are located at E_c − 0.36 eV and E_c − 0.44 eV, respectively. The concentrations, thermal emission rates, and the capture cross sections for the majority carriers at these defects are reported. The behavior of these defects under heat treatment and the emergence of secondary defects, H5(E_v +0.22 eV) and H6 (E_v +0.34 eV), will be discussed.     

Stoichiometry-dependent deep levels in undoped p-type Al0.38Ga0.62As grown by liquid phase epitaxy

Photocapacitance (PHCAP) and photoluminescence (PL) measurements were applied to unintentionally doped p-type Al_0.38Ga_0.62As grown by liquid phase epitaxy using the temperature difference method under controlled vapor pressure. PHCAP spectra revealed three dominant deep levels at E_v+0.9, E_v + 1.45, and E_v+1.96 eV, and a deep level at E_v+0.9−1.5 eV which was not neutralized by forward bias injection. These level densities increase with increasing arsenic vapor pressure and net shallow acceptor density. Furthermore, PL spectra reveal a deep level at 1.6–1.7 eV. The PL intensity of this deep level increases with increasing arsenic vapor pressure. These deep levels are thought to be associated with excess As.     

Deep Traps in AlGaN/GaN Heterostructure Field-Effect Transistors Studied by Current-Mode Deep-Level Transient Spectroscopy: Influence of Device Location

Across-wafer variation of deep traps in AlGaN/GaN heterostructure field- effect transistors (HFETs) has been studied using current-mode deep-level transient spectroscopy (I-DLTS). It is found that applying a bias to the gate corresponding to open-channel or nearly pinched-off operation while cooling from 550 K has significant effects on the threshold voltage of the device and spectral features of I-DLTS at T <~400 K. To compare I-DLTS spectra acquired under different measurement conditions (e.g., different filling pulse heights or widths), attention should be paid to the gate bias applied while cooling. We find that the spectral features observed across the wafer can be divided into three categories: (i) a dominant trap A ₁ at ~360 K; (ii) a prominent hole-like trap H ₁ at ~480 K; and (iii) for T < 340 K, several traps varying in magnitude, including A, B, and C ₁, and a hole-like trap H ₀. Based on I-DLTS, measured as a function of filling pulse height and filling pulse width, we suggest that: (i) trap A ₁, with E _T = 1.1 eV to 1.2 eV and σ = 4 × 10⁻¹³ cm² to 2 × 10⁻¹² cm², is associated with extended defects, such as threading dislocations; (ii) traps A, B, and C ₁, located mainly in the two-dimensional electron gas (2DEG) channel region, could be related to point defects; and (iii) the hole-like trap H ₁, with E _T = 1.2 eV to 1.3 eV and σ = 8 × 10⁻¹⁷ cm² to 5 × 10⁻¹⁵ cm², may be related to surface states. Note that traps A (~0.67 eV), B (0.58 eV to 0.61 eV), and C ₁ (0.44 eV to 0.49 eV) are commonly observed in GaN layers grown by various techniques. For all of the HFETs near the center of the wafer, trap A ₁ dominates the DLTS spectra, whereas trap B becomes prominent for some of the HFETs near the edge of the wafer.     

Electrical properties of undoped bulk ZnO substrates

Undoped bulk ZnO crystals obtained from Tokyo Denpa show either resistive behavior [(5×10⁴)−(3×10⁵) Ohm cm) or low n-type conductivity (n ⋍10¹⁴ cm⁻³) with mobilities in the latter case of 130–150 cm²/V sec. The variation in resistivity may be related to the thermal instability of Li that is present in the samples. The Fermi level is pinned by 90-meV shallow donors that are deeper than the 70 meV and hydrogen-related 35-meV shallow donors in Eagle Pitcher and Cermet substrates. In all three cases, 0.3-eV electron traps are very prominent, and in the Tokyo Denpa material they dominate the high-temperature capacitance-frequency characteristics. The concentration of these traps, on the order of 2×10¹⁵ cm⁻³, is about 20 times higher in the Tokyo Denpa ZnO compared to the two other materials. The other electron traps at E_c −0.2 eV commonly observed in undoped n-ZnO are not detected in conducting Tokyo Denpa ZnO samples, but they may be traps that pin the Fermi level in the more compensated high-resistivity samples.     

Nature of E c −0.37 eV centers and the formation of high-resistivity layers in n-type silicon

The DLTS and Van der Pauw methods are used to investigate the production of E _c −0.37 eV centers responsible for the formation of high-resistivity layers in n-type Si irradiated with electrons and annealed in the temperature range 80–320 °C. An analysis of the experimental data leads to a conclusion as to the composition of the E _c −0.37 eV centers ([V-O-C]) and to the conclusion that their formation is stimulated by a flux of interstitial atoms away from the interface into the interior of the semiconductor during annealing accompanied by the reactions: 1) I+C_s→C_i,C_i+[V-O]→[V-O-C] (dominant reaction); 2) I+V ₂→V,V+[C-O]→[V-O-C]. Fiz. Tekh. Poluprovodn. 31, 993–997 (August 1997)     

Evidence for causality between GaN RF HEMT degradation and the EC-0.57 eV trap in GaN

The degradation of industry-supplied GaN high electron mobility transistors (HEMTs) subjected to accelerated life testing (ALT) is directly related to increases in concentrations of two defects with trap energies of E_C-0.57 and E_C-0.75 eV. Pulsed I-V measurements and constant drain current deep level transient spectroscopy were employed to evaluate the quantitative impact of each trap. The trap concentration increases were only observed in devices that showed a 1 dB drop in output power and not the result of the ALT itself indicating that these traps and primarily the E_C-0.57 eV trap are responsible for the output power degradation. Increases from the E_C-0.57 eV level were responsible for 80% of the increased knee walkout while the E_C-0.75 eV contributed only 20%. These traps are located in the drain access region, likely in the GaN buffer, and cause increased knee walkout after the application of drain voltage.     

Effect of composition on deep levels in heteroepitaxial GexSi1−x layers and evidence for dominant intrinsic recombination-Generation in relaxed ge layers on Si

Electron traps, hole traps, and the dominant recombination-generation (R-G) centers have been investigated with deep level transient spectroscopy and current-voltage/temperature measurements in heteroepitaxial Ge_xSi_1-x alloys with x ranging from 0.15 to 1, grown on graded Ge_y.Si_1−y/Si substrates. For all samples with compositions x < 0.85, which retain the Si-like conduction band structure, we detect a dominant electron trap and R-G center whose activation energy is ΔE = 0.5 eV, independent of composition. This energy agrees with that of electron traps previously reported for plastically deformed (PD) Si, suggesting a connection to the Si-like band structure. This 0.5 eV level dominates the reverse leakage current over a wide range of growth and annealing conditions for the 30% Ge samples, indicating that the electronic state at ΔE = 0.5 eV is a very efficient R-G center, as would be expected from its midgap position. Alternatively, for strain relaxed, pure Ge (< 1), we detect electron traps at E_c − 0.42 eV and E_c − 0.28 eV, in agreement with the literature on PD Ge and Ge bicrystals. These energies are significantly different from those observed for x < 0.85, and we conclude that these changes in activation energy are due to changes in the conduction band structure for high Ge content. Moreover, in contrast with the Si-like samples (x < 0.85), the reverse leakage current in the relaxed Ge cap layer is not controlled by deep levels, but is rather dictated by intrinsic, band-to-band generation due to the reduced bandgap of Ge as compared to Si-like alloys. Only for reverse bias magnitudes which incorporate a significant portion of the graded buffer within the depletion region do R-G centers dominate the reverse leakage current. These results confirm the high quality of the strain-relaxed, pure Ge cap region which was grown on a Ge_ySi_1−y/Si step graded heterostructure (where y was increased from 0 to 1) by ultra high vacuum chemical vapor deposition. Finally, we report for the first time, what is apparently the dislocation kink site state at E_c − 0.37 eV, in a Ge_xSi_1−x alloy.     

Effects of Illumination on Ar<Superscript>+</Superscript>-Implanted <Emphasis Type="Italic">n</Emphasis>-Type 6H-SiC Epitaxial Layers

Argon ions were implanted into n-type 6H-SiC epitaxial layers at 600°C. Postimplantation annealing was carried out at 1,600°C for 5 min in an Ar ambient. Four implantation-induced defect levels were observed at E_C-0.28 eV, E_C-0.34 eV, E_C-0.46 eV, and E_C-0.62 eV by deep level transient spectroscopy. The defect center at E_C-0.28 eV is correlated with ED₁/ED₂ and with ID₅. The defect at E_C-0.46 eV with a capture cross section of 7.8 × 10⁻¹⁶ cm² is correlated with E1/E2, while the defect at E_C-0.62 eV with a capture cross section of 2.6 × 10⁻¹⁴ cm² is correlated with Z1/Z2. Photo deep level transient spectroscopy was also used to study these defects. Upon illumination, the amplitudes of the deep level transient spectroscopy (DLTS) peaks increased considerably. Two emission components of Z1/Z2 were revealed: one fast and the other slow. The fast component could only be observed with a narrow rate window. In addition, a new defect was observed on the low-temperature side of the defect at E_C-0.28 eV when the sample was illuminated. [rl](Received ...; accepted ...)     

Germanium as a deep level in alxga1− xas grown by organometallic vapor phase epitaxy

Optoelectronic devices require materials which exhibit extremely low trap concentrations. The Al_xGa_1−xAs system has been used extensively for optoelectronic applications despite trap concentrations in the Al_xGa_1−xAs which limit the efficiency of the resulting devices. Deep level transient spectroscopy (DLTS) performed on Al_0.2Ga_0.8As layers grown by organometallic vapor phase epitaxy (OMVPE) has revealed three traps with concentrations >10¹³ cm⁻³ -E _c-E_t = 0.3, 0.5 and 0.7 eV. The dominant source of the 0.3 eV trap has proven to be a Ge impurity in arsine. SIMS analysis of Al_0.2Ga_0.8As samples show Ge as the only candidate for the impurity responsible for the 0.3 eV trap. DLTS and SIMS analysis performed on Al_0.2Ga_0.8As samples intentionally doped with Ge displayed a proportional increase in the 0.3 eV trap concentration with the Ge concentration and establishes that Ge is indeed the source of the 0.3 eV trap in Al_xGa_1−xAs. Comparison of C-V, SIMS and DLTS measurements performed on Al_xGa_1-xAs:Ge indicate that approximately 30% of elemental Ge incorporated created the 0.3 eV trap, DX_Ge.     

Characterization of electron traps in plasma-treated AlInAs

Electron traps have been investigated for molecular beam epitaxially grown n-AlInAs by isothermal capacitance transient spectroscopy measurement. Two traps, EOl(Ec−0.44 eV) and EO2(Ec−0.52 eV), appear for oxygen-plasma treated samples in addition to El(Ec−0.47 eV) and E2(Ec−0.69 eV) traps detected for control (as-chemically etched) sample. On the other hand, the EO1 trap is not detected for Ar plasma-treated sample. This suggests that the EO1 is generated by reaction of atomic oxygen with AllnAs and that the EO2 trap is induced by plasma damage. It is found that a reduction in the densities of the four traps occurs due to annealing subsequently after oxygen plasma treatment.     

Energy levels of vacancies and interstitial atoms in the band gap of silicon

Based on the analysis of the secondary processes of radiation-induced defect formation in Si crystals with charge-dependent selective traps for vacancies and interstitial atoms, the energy levels of vacancies and interstitial atoms were identified; these level were determined previously from the effect of the irradiation conditions on the annihilation rate of elementary primary defects. It is ascertained that the levels at ～E _c -0.28 eV and at ～E _c -0.65 eV in the band gap of Si belong, most likely, to vacancies; the levels at ～E _c -0.44 eV, at ～E _c -0.86 eV, and, presumably, at ～E _c -0.67 eV belong to intrinsic interstitial atoms.     

Space-charge-limited currents in an Se<Subscript>95</Subscript>As<Subscript>5</Subscript> chalcogenide glass-like semiconductor system containing EuF<Subscript>3</Subscript> impurities

It is established that charge carrier (hole) transport in the Al-Se₉₅As₅〈EuF₃〉-Te structure is effected by unipolar injection currents limited by space charges with the involvement of two capture trap groups. Shallow traps corresponding to charged intrinsic defects C ₁⁻ are related to broken selenium bonds. Deep traps corresponding to charged intrinsic defects P ₂⁻ are formed by arsenic atoms with broken coordination. It is shown that the EuF₃ impurity strongly affects the concentration of the capture traps, especially those localized near the Fermi level.     

Thermally stimulated current measurements in N-type LEC GaP

Thermally stimulated current measurements have “been performed on n-type LEC GaP substrates. Nine deep donor states with energy level ranging from 0.27eV to 0.90eV have been observed. The most prominent peak at O.36eV has been associated with the incorporation of common donors Te, Se and S and it is found that the concentration of this 0.36eV trap increases as the square of the net doping concentration (N_D-N_A). Association of this O.36eV trap with the commonly observed 1.72eV luminescence band and the S-pits in LEC crystals is discussed.     

Deep level studies in MBE GaAs grown at low temperature

The photo-induced current transient spectroscopy (PICTS), thermoelectric effect spectroscopy (TEES) and thermally stimulated current (TSC) spectroscopy have been used to characterize the deep levels in the GaAs materials grown at low temperature by molecular beam epitaxy. At least five hole traps and five electron traps have been identified by the TEES measurement employing a simplified sample arrangement. We have studied the behavior of various traps as a function of the growth temperature and the post-growth annealing temperature. Some of the shallower hole traps were annealed out above 650‡ C. Electron traps atE _c- 0.29 eV andE _c- 0.49 eV were present in the material, and have been identified as M3 and M4, respectively. The dominant electron trap, atE _c- 0.57 eV, is believed to be associated with the stoichiometric defect caused by the excess As in the material, and our data show evidence of forming a defect band by this trap. A possible model involving As precipitates is proposed for this trap atE _c-0.57 eV.     

Effect of iodine impurity on relaxation of photoexcited silver chloride

The time and temperature dependences of relaxation of excited AgCl and AgCl:I crystals is studied by the method of photostimulated flash of luminescence. The presence of iodine impurity in silver chloride gives rise to hole recombination (luminescence) centers and hole traps in the band gap. It is shown that the main contribution to the decrease in the concentration of electrons localized at deep traps is made by the recombination of electrons with holes released thermally from shallow localization levels (iodine-related centers). Estimation of activation energy for the relaxation process showed that these energies for the AgCl and AgCl:I samples under study are the same within the experimental error and are equal to E _{rel 1} = 0.01 ± 0.0005 eV for the initial stage of relaxation and E _{rel 2} = 0.09 ± 0.005 eV for the final state. This fact indicates that the majority of hole traps involved in the relaxation process in AgCl are related to iodine impurity. In the course of thermal relaxation in AgCl, relocalization of nonequilibrium charge carriers from shallow levels to deep levels is observed. The depth of the corresponding trap is E _arl = 0.174 ± 0.03 eV.     

Shallow and Deep Centers in As-Grown and Annealed MgZnO/ZnO Structures with Quantum Wells

Shallow and deep centers in ZnO(P)/MgZnO/ZnO/MgZnO/ZnO(Ga) structures grown by pulsed laser deposition on sapphire were studied before and after annealing in oxygen atmosphere at high temperatures of 850°C to 950°C. In both as-grown and annealed structures, microcathodoluminescence spectra in the near-bandgap region demonstrate a blue-shift by 0.13 eV compared with bulk ZnO films, indicating carrier confinement in the MgZnO/ZnO/MgZnO quantum well (QW). Annealing strongly decreases the concentration of shallow uncompensated donors from ~10¹⁷ cm⁻³ to ~10¹⁶ cm⁻³ and makes it possible to probe the region of the QW by capacitance–voltage (C–V) profiling. This profiling confirms charge accumulation in the QW. The dominant electron traps in the as-grown films are the well-known traps with activation energies of 0.3 eV and 0.8 eV. After annealing, the electron traps observed in the structure have activation energies of 0.14 eV, 0.33 eV, and 0.57 eV, with the Fermi level in the n-ZnO(P) pinned by the 0.14-eV traps. The annealing also introduces deep compensating defects that decrease the intensity of band-edge luminescence and produce a deep luminescence defect band at 2.2 eV. In addition, a defect vibrational band becomes visible in Raman spectra near 650 cm⁻¹. No conversion to p-type conductivity was detected. The results are compared with the data for the structures successfully converted to p-type, and possible reasons for the observed differences are discussed.