Current Deep Level Transient Spectroscopy Research Articles

Development of current-based microscopic defect analysis methods and associated optical filling techniques for the investigation on highly irradiated high resistivity silicon detectors

Current based microscopic defect analysis methods such as current deep level transient spectroscopy (I-DLTS) and thermally stimulated current (TSC) have been further developed in accordance with the need for the defect analysis of highly irradiated ( Φ n > 10 13 n/cm 2) high resistivity silicon detectors. The new I-DLTS/TSC system has a temperature range of 8 K ≤ T ≤ 450 K and a high sensitivity that can detect a defect concentration of less than 10 10/cm 3 (background noise as low as 10 fA). A new filling method using different wavelength laser illumination has been applied, which is more efficient and suitable than the traditional voltage pulse filling. It has been found that the filling of a defect level depends on such factors as the total concentration of free carriers generated or injected, the penetration length of the laser (laser wavelength), the temperature at which the filling is taking place, as well as the decay time after the filling (but before the measurement). The mechanism of the defect filling can be explained by the competition between trapping and detrapping of defect levels, possible capture cross section temperature dependence, and interaction among various defect levels in terms of charge transferring. Optimum defect filling conditions have been suggested for highly irradiated high resistivity silicon detectors.

Quantitative electron-beam tester for defects in semiconductors (CL/EBIC/SDLTS system)

A pulsed electron beam testing system was developed for quantitative analysis of optical and/or electrical properties of defects in semiconductors. This system can control not only the electron beam but also detect and record the absolute values of light intensity: cathodoluminescence (CL), electron-beam-induced current (EBIC), and scanning deep level transient spectroscopy (SDLTS). Several improvements were made to enable accurate quantitative measurement. An optical system with an ellipsoidal mirror of low magnification was designed to make uniform collection efficiency of CL over a wide specimen area. A photon counting system and a lock-in amplifying system were implemented to measure absolute value of light intensity. A quantitative EBIC or SDLTS system was also incorporated to this system. The combined system of CL/EBIC/SDLTS can both image the distribution of defects and study their characteristics.

High resolution method for the analysis of admittance spectroscopy data

There are several experimental methods which give information about the thermal relaxation times of the deep levels in a semiconductor. Analyzing the temperature dependence of the relaxation times, the activation energy and the cross section of the corresponding deep levels can be determined. An essential problem of such methods is the identification of the relaxation times in the measured signal. In the context of time-dependent measurements such as photoinduced current transient spectroscopy and deep level transient spectroscopy, Tikhonov regularization was recently proposed as a high resolution method for this purpose. In this contribution it is proposed to apply Tikhonov regularization in order to identify the thermal relaxation times in admittance spectroscopy data. The method is tested and discussed using simulated data. Finally, admittance spectroscopy data of a GaAs diode are analyzed. The results demonstrate that the resolution of an ordinary admittance spectroscopy setup can considerably be improved by the application of Tikhonov regularization.

Deep centers and electroluminescence in 4HSiC diodes with a p-type base region

We have performed electrical and optical measurements to characterize the processes of radiative recombination of carriers for 4H-SiC p-n + junctions. We have detected three levels in the band gap, using capacitance and current deep-level transient spectroscopy, which are responsible for the bands observed in the electroluminescence spectra. We assume that the HK 1 center with an activation energy of E v + 0.229 eV and a hole capture cross-section of 8 × 10 −13 (300/ T) 3 is related to the Al acceptor level.

Measurement and modeling of drain current DLTS in enhancement SOI MOSFETs

A new approach for current deep-level transient spectroscopy (DLTS) in enhancement-mode MOSFETs is presented. The novelty of this approach is three-fold: (1) it improves the modeling of the transient drain current by directly calculating the transient variation of the inversion charge instead of that of the threshold voltage; (2) it accounts for the mobility field dependence and series resistance effects through a new current dependent mobility law; and (3) it applies a combination of two already existing DLTS measurement procedures in order to avoid any temperature scanning. Measurements were carried out on n-channel enhancement mode MOSFETs fabricated on bulk Si as well as on SIMOX substrates. Deep-level traps were identified in SIMOX devices and the trap parameters (density, time-constant and trap level) were determined using the proposed method.

Characterization of trapping states in polycrystalline-silicon thin film transistors by deep level transient spectroscopy

Polycrystalline-silicon (poly-Si) thin film transistors (TFTs) have been assessed using deep level transient spectroscopy (DLTS) based on the measurement of current transients due to the thermal emission of carriers trapped at deep levels in the poly-Si. Measurements were made on fully hydrogenated TFTs and the DLTS signal was found to vary continuously between 77 K and room temperature, without showing the characteristic peaks normally associated with point defects in single crystal Si. This demonstrates the existence of a continuous distribution of states through the band gap rather than discrete monoenergetic states. Analysis of the DLTS signal suggests the density of states increases rapidly toward the conduction band edge indicating the presence of a relatively high density of tail states. After annealing the TFTs at 450 °C the current DLTS signal was found to increase due to an increase in trap state density, which occurs because Si–H bonds are thermally unstable at this temperature. The increase in signal was over the whole temperature range of the measurement showing that hydrogen passivates states up to at least 0.1 eV below the conduction band as well as states nearer mid-gap. The effects of varying the fill voltage on the measurement and the time dependence of the current transient were also investigated and added weight to arguments that the current transient was due to the thermal emission of carriers trapped at deep levels.

Drain Current DLTS Spectra and GaAs Substrate Crystal Effect on Low-Frequency-Oscillations of Si-Implanted Mesfets

ABSTRACTWe report drain-current (Id) deep level transient spectroscopy (DLTS) spectra and liquid-encapsulated-Czochralski-technique (LEC) GaAs crystal effect on low-frequency-oscillation (LFO) of wide gate (400-μm) Si-implanted GaAs metal- semiconductor field- effect- transistors (MESFETs). In the range of this experiment we could not find distinguishing DLTS peaks surely to be linked with Id-LFO of the MESFETs. Stoichiometric-melt growth LEC-boules showed relatively large Id-LFO phenomena. As-rich-melt growth LEC-boules showed tolerance to Id-LFO. We conclude that Id-LFO is not directly linked to deep centers themselves but interaction between deep centers and potential profiles and electrons. Stability of potential profile or band profile depends on “pinning’ center, which affects Fermi-level or quasi-Fermi-level stability. ‘Pinning’ center such as EL2s of ‘LEC GaAs crystals” is essential.

New model for the characterization of bulk traps by current deep level transient spectroscopy in metal-oxide-semiconductor transistors

New models for the characterization of volume traps in enhancement-mode field-effect transistors by current deep level transient spectroscopy are proposed. First, a simple model that is based on the use of the experimentally measured transconductance of the device is studied. Consequently it includes, in principle, the dependence of the mobility with gate bias and therefore allows us to avoid any approximation in the gate voltage dependence of the mobility. Different improvements of this first model are then proposed, compared, and discussed with respect to previous works. Finally a more complete model that requires less approximations is derived. These different approaches are finally used for the characterization of a deep trap in a bulk enhancement-mode field-effect transistor.

Current deep-level transient spectroscopy in MOS structures with a boxcar integrator and an arbitrary gate-width-data analysis

New formulas are given describing the current correlation signals of bulk traps and interface states in metal-oxide-semiconductor structures. The presented formulas describe the correlation signals for analog measurement systems utilizing a single-channel or a two-channel boxcar integrator. These formulas take into account both the so-called Lambda effect and an arbitrarily chosen gate width of a boxcar integrator. A comparative analysis has been made of a potential sensitivity and selectivity of these systems as a function of a gate width.

A deep-level transient spectroscopy study of high electron mobility transistors subjected to lifetime stress tests

Current and capacitance deep-level transient spectroscopy (DLTS) measurements have been used to examine the effect of accelerated lifetime stress tests on AlGaAs/GaAs high electron mobility transistor structures. The epilayers for these devices were grown by molecular beam epitaxy on semi-insulating GaAs substrates, and they were processed to have 0.7-μm long gates. The major defect detected by both the capacitance and current measurements on unstressed control devices is the DX center in the AlGaAs. The current DLTS tests on stressed samples reveal an additional defect, which is not found in capacitance DLTS measurements on the same devices. The new current DLTS line appears to be a minority-carrier trap, which is not expected as the trap filling pulse was not set to inject minority carriers. These features can be explained if the defect is associated with surface states located between the gate and either the source or the drain.

Application of isothermal current deep level transient spectroscopy to solar cells

The utility of isothermal current deep level transient spectroscopy (DLTS) techniques in directly probing solar cells is described and illustrated. A modified approach to processing the isothermal DLTS data is also presented. Specifically, it is pointed out that properly normalized isothermal data, whether derived from a current or capacitance transient, should conform to a single, temperature-independent curve.

Transient Current and Transient Capacitance Measurements of Defects in AlGaAs Hemts

AbstractThe defects present in AlGaAs high electron mobility transistor (HEMT) devices subjected to accelerated lifetime tests have been studied by transient current, and when possible, by transient capacitance techniques. By measuring transient source-drain currents following the application of a voltage pulse to the gate it is possible to perform deep level transient spectros-copy (DLTS) experiments on HEMT devices which are too small for the conventional capacitance DLTS. The capacitance and current spectra for both stressed and unstressed HEMTs contain the AlGaAs DX defect. The current DLTS spectra for stressed devices contain an additional feature which is not found in capacitance DLTS measurements on the stressed HEMTs. This additional current DLTS feature is anomalous in that the transient has a sign which is opposite to that expected for a majority carrier trap. The absence of the new defect in the capacitance DLTS suggests that the defect is located in the channel between the gate and either the source or the drain. The current DLTS line shape of the stressed induced defect depends upon the polarity and size of the source-drain voltage.

Bulk traps in ultrathin SIMOX MOSFET's by current DLTS

Bulk traps in very thin ( approximately 100-nm) SIMOX films have been studied by applying current deep-level transient spectroscopy (DLTS) to fully depleted, enhancement MOS transistors, fabricated in these films. The effect of states at both the front and back SiO/sub 2/-Si interfaces is eliminated by suitable biasing. Using this technique, a bulk trap with energy level 0.44 eV above the valence-band edge, capture cross section approximately 10/sup -17/ cm/sup 2/, and concentration approximately 10/sup 15/ cm/sup -3/, which is believed to be due to iron contamination, has been identified. >

Deep states in silicon-on-insulator substrates prepared by oxygen implantation using current deep level transient spectroscopy

Current deep level transient spectroscopy was applied using enhancement n-channel metal-oxide-semiconductor field-effect transistors fabricated in silicon-on-insulator substrates (prepared by oxygen implantation) to study the deep levels existing in the substrates. The current transients are not affected by the large series resistances which affect the measurement of capacitance transients on thin films. For the transistors used in this work a hole trap was found with energy ET=0.63 eV above the valence-band edge. The concentration and capture cross section of this state were estimated to be 1014 cm−3 and 10−16 cm2, respectively.

Deep Levels in Semi‐Insulating LEC GaAs Before and After Silicon Implantation

The deep trapping levels present before ion implantation of silicon into the semi‐insulating LEC starting material were investigated using optical transient current spectroscopy (OTCS). MESFET channel current deep level transient spectroscopy (DLTS) was used for the implanted material. With a silicon nitride layer used to encapsulate the for postimplantation annealing and with implantation directly into the , it was found that of seven or more deep levels seen in the semi‐insulating substrate prior to silicon implantation only the level believed to be EL12 remained. On implanting through a thin encapsulating layer and annealing under , only EL2 was found. With a silicon dioxide layer as an encapsulant, two traps remained and two apparently unreported levels appeared.

Capture-cross-section determination by transient-current trap-filling experiments

A detailed analysis is presented of capture-cross-section determination by transient-current (e.g., current deep-level transient spectroscopy) trap-filling measurements. Data for electron capture by the Au acceptor illustrates the technique. Calculations including spatially dependent capture and the competition of emission with capture are shown to match the experimental data. Results using the parabolic band-bending approximation are compared with those in which Poisson’s equation is solved exactly. The latter is found to fit the data slightly better. By curve fitting this calculation to trap-filling data, the electron capture cross section σn of the Au acceptor is found to be 7.1±0.3×10−17 cm2. The effects obtaining when the trap concentration NT cannot be assumed negligible compared to the shallow-dopant density ND are investigated, and found to be unimportant for the samples studied here (NT/ND≊0.3). It is demonstrated that σn cannot be obtained accurately from the initial slope of a semilog plot of the effective fraction of traps remaining empty at the end of the trap-filling pulse versus filling-pulse width, as is conventionally done for capacitance-transient measurements.