Fourier Transform Deep Level Transient Spectroscopy Research Articles

Profiling of Border Traps at GeSn and High-K Oxide Interface

GeSn has emerged as an interesting channel material for beyond 14 nm node devices, in CMOS as well as TFET configurations [1-3]. However, the advantages of sub-60mV/decade subthreshold swing (SS) due to the increased band to band tunneling in small direct-bandgap tensile-strained GeSn and higher carrier mobilities are dwarfed by the presence of traps in the gate stack. For GeSn devices, Al or Hf based high-k dielectrics have been recognized as candidates for advanced gate stacks [4-6], where the presence of the interlayer oxide (IL) has shown to decrease the interface state density, Dit [7]. However, the border traps, known to be present in the vicinity of the GeSn/Oxide interface within a few nm, can exchange charge carriers with the underlying GeSn epitaxial layer via direct tunneling of injected charge, [8-9] thus reducing the mobility and increasing SS. A detailed characterization of these border traps has not been reported so far for GeSn gate stacks. In this work, two techniques, Fourier transform deep level transient spectroscopy (FT-DLTS) and temperature derivative admittance spectroscopy (TDAS) have been used to investigate the distribution of border traps in GeSn MOS capacitors.A 40 nm nominally thick compressively strained, un-intentionally doped Ge0.092Sn0.08 layer is grown using chemical vapor deposition on a thick Ge buffer layer on Si. The GeSn surface is baked at 5000C for 30 min to remove native oxide. A complete de-oxidation is confirmed by RHEED analysis. Plasma oxidation at 1500C is then carried out for 1 nm of GeSnOx IL formation, followed by 9 nm of Al2O3 deposition at -500C. Ni dots are deposited for gate contact, followed by forming gas anneal.A broad peak at higher temperatures (~235 K) was observed in conventional FT-DLTS measurements, when MOS capacitors are pulsed from depletion (Ur=2V) to accumulation (Up=-1.5V) as shown in Figure 1a. This peak originates from traps present at the oxide interface and is absent when the device is pulsed from deep depletion to depletion (positive pulse bias). The peak corresponds to a maximum in energetic distribution of border traps, found to be present at (ΔEa = 0.36± 0.02 eV) from the valence band edge (figure 1b). The capture kinetics is studied by means of pulse filling measurement using isothermal DLTS, which exhibits an s-shaped profile. A capture cross-section, σp of ~5e-17 /cm2 is estimated. With an assumption of all states being unoccupied or occupied at the beginning of the trap filling process, the density of border trap, the tunneling depth of the traps in the oxide and the trap energy are determined. Similar results are obtained for Al2O3/GeO2/GeSn capacitors.Figure 1: (a) Conventional FT-DLTS Spectrum at quiescent reverse bias (Ur) of 2 V, trap filling pulse duration (Tp) of 100 µs, emission period width (Tw) of 51.2 ms and varying filling pulse amplitude(Up). The solid curve shows de-convoluted peaks with a Gaussian distribution for the curve corresponding to Up = -1 V. The peak at ~235K corresponds to border traps with a background distribution. The peaks at lower temperature represent bulk traps in GeSn/Ge epitaxial layers[10]. (b) Temperature derivative Admittance spectrum [11], at gate bias(Vg) of -1V. An Arrhenius plot of ln(ωpeak/T2) vs 1/kT is used to obtain activation energy of 0.36 eV from the valence band edge where ωpeak is the frequency at peak position and k is the Boltzmann constant. In conclusion, FT-DLTS and TDAS of MOS capacitors on GeSn epitaxial layers are used to study the frequency and temperature dispersion observed in the accumulation region of capacitance curves. The results are described by means of the assumption of spatially and energetically distributed states, present in the depth of the oxide which interact by tunneling and thermally activated transitions of carriers.[1] S. Wirths et al., Appl. Phys. Lett. 102, 192103(2013). [2] S. Gupta et al., VLSI Technology (VLSIT), 2012 Symposium on. IEEE, 2012.[3] Y. Yang et al., Electron Devices Meeting (IEDM), 2012 IEEE International. IEEE, 2012.[4] Y. Yang et al., Electron Devices, IEEE Transactions on 60.12 (2013): 4048-4056.[5] S. Gupta et al., Electron Devices Meeting (IEDM), 2012 IEEE International. IEEE, 2012.[6] X. Gong et al., VLSI Technology (VLSIT), 2012 Symposium on. IEEE, 2012.[7] M. Zhao et al., Appl. Phys. Lett., APL 102.14 (2013): 142906.[8] H. D. Lakhdari et al., Physical Review B, 38.18 (1988): 13124.[9] P. Van Staa et al., J. Appl. Phys., JAP 54.7 (1983): 4014-4021.[10] S. Gupta et al., ECS Transactions 53.1 (2013): 251-258.[11] V. Jian Li et al., J. Appl. Phys., JAP 109.8 (2011): 083701.

Electrically Active Defects in Electron Irradiated P-Type 6H-SiC

An isochronal annealing in the 100-1200 oC temperature range was carried out on 116 and 400 keV electron irradiated Al-doped 6H-SiC epitaxial layers. Electrical characterization of the epilayers, performed by Fourier-Transform Deep Level Transient Spectroscopy, revealed the presence of six levels in the band gap, in the 0.1-1.6 eV energy range. Their nature is discussed in the light of previous experimental and theoretical works found in the literature.

Minority carrier injection limited current in Re/4H‐SiC Schottky diodes

AbstractCurrent–voltage–temperature, capacitance–voltage, and Fourier transform‐deep level transient spectroscopy (FT‐DLTS) measurements have been employed to gain insights into the conduction mechanism in Re/4H n‐silicon carbide (SiC) Schottky diodes. A power law dependence of current on voltage has been observed in the Re/4H‐SiC Schottky diodes at large forward bias. DLTS studies of the diodes reveal the presence of a deep trap at an energy $E_{\rm c} - E_{\rm t} = 0.6\,{\rm eV}$. The electron trap center is tentatively associated with the prominent Z1 defect level, and the obtained trap density and capture cross‐section values correlate well with available literature values. Excess capacitance has been observed under forward bias suggesting that minority carrier injection is the mechanism responsible for the observed characteristics.

Thermal stability of deep levels between room temperature and 1500 °C in as-grown 3C-SiC

We report on the thermal stability of deep levels detected in as-grown bulk 3C-SiC. The investigation was performed by Fourier-transform deep level transient spectroscopy and an isochronal annealing series was carried out in the 100–1500 °C temperature range. We found three traps located between 0.14–0.50 eV below the conduction band edge minimum (EC). The shallower trap anneals out at temperatures below 1200 °C while the others display a high thermal stability up to at least 1500 °C. The nature of the former trap is discussed in detail on the basis of its annealing behavior and previous theoretical data found in the literature.

Deep levels investigation of AlGaN/GaN heterostructure transistors

Abstract In this work we investigate AlGaN/GaN HEMTs structures grown by metalorganic chemical-vapor deposition on SiC substrates. They consist of a 22 nm thick undoped AlGaN barrier with Al mole fraction x=0.24 on top of a 1.7 μm unintentionally doped GaN buffer layer. Structures with large gate are analyzed by Fourier transform deep level transient spectroscopy (FT-DLTS) between 100 and 450 K. The dominant traps are found to be electron traps with activation energies of 0.12, 0.15, 0.21, 0.40, 0.49 and 0.94 eV. The detailed analysis of DLTS spectra shows a strong dependence on the electric-field in the space charge region of the Schottky contact. Poole–Frenkel mechanism is observed and confirmed by the reverse current characteristic analysis. The nature of the deep traps is also discussed.

Electrically active defects in boron doped diamond homoepitaxial layers studied from deep level transient spectroscopies and other techniques

AbstractHomoepitaxial thin films of boron doped diamond are investigated with the help of Fourier transform deep level transient spectroscopy (FT‐DLTS) and high resolution isothermal transient spectroscopy (HR‐ITS) in order to determine the properties of hole traps. Conductivity studies and other characterisation techniques of defects are also used to bring complementary information. The main conclusions are as follows: (i) many hole traps can be found with activation energies in the range 0.9–1.6 eV; (ii) most of them show inwards decreasing concentrations, which passes below 1015/cm3 at a depth close to a few hundred nanometers from the surface; (iii) some of them are metastable, in the sense the emission rate distribution undergoes changes after thermal treatments performed at temperatures well below those used for growth and previous annealing. These two last facts suggest that hydrogen atoms may be involved in charged defects. In other samples, an irreversible decrease of the conductivity correlated with the increase of the cathodoluminescence (CL) A band demonstrates that a deep donor is connected with dislocations and acts as a very efficient compensating centre in the first 400 nm just below the surface but with a decreasing concentration deeper inside the homoepitaxial layer. The physical origin of these deep levels is discussed and believed to be likely connected with hydrogen atoms.

Capacitance Spectroscopy Study of Midgap Levels in n-Type SiC Polytypes

As-grown and 116 keV electron-irradiated n-type 3C and 4H-SiC epilayers were electrically characterized by means of Fourier-transform deep level transient spectroscopy (FT-DLTS). A total of four deep levels, in the 0.20-0.73 eV range, below the conduction band, have been detected. By considering the band gap offset between 4H and 3C polytypes, we found that the deepest level in 3C-SiC labeled K3 (Ec-0.73 eV) has an energy position close to the EH6/7 level in 4H-SiC. An electron-dose dependence study of K3 and EH6/7, reveals that these two centers display a similar dose dependence behavior, suggesting that they may be related to the same defect.

Electronic properties of deep defects in n-type GaN

The recurrent presence of deep levels in n-type hexagonal GaN films grown by metal organic chemical vapour deposition (MOCVD) is testified by Fourier transform deep level transient spectroscopy (FTDLTS) studies performed in various samples. Some of these electron traps are systematically detected with emission rates showing activation energies in the range 0.94–1.30 eV and apparent capture cross sections from 10 −15 to 10 −12 cm 2. These properties are deduced from an Arrhenius diagram where data coming from samples of different origins are close to the same straight line, assessing a common physical identity for the corresponding defects. However, isothermal spectroscopy and trap filling kinetics characteristics contradict these capture cross section values and even show nonmonotonic variations which must be ascribed to the emission and capture rates of interactive multiple deep levels. Fine structure of the energy levels, although it is never resolved by standard FTDLTS, is indeed revealed by a high resolution DLTS method (HRDLTS) which relies on a time domain treatment of the transients recorded over a large range of time. Multiple levels are well evidenced in this way and both their ionisation energy and capture cross sections turn out to be much less than those measured by standard FTDLTS, respectively in the ranges 0.26–0.90 eV and near 10 −17 cm 2 on the one hand or less than 10 −20 cm 2 on the other hand, suggesting that these defects can capture two electrons, leading to three charge states and negative charges at least when filled with 2 electrons. Correlation of the concentration of these defects both with the shallow donor concentration and dislocation density suggest that they are due to the doping impurities in dislocation sites.

Deep centres in bulk MOCVD n‐type hexagonal GaN thin films and near their interface

Deep level transient spectroscopy (DLTS) is performed in metal organic chemical vapour deposition (MOCVD) hexagonal GaN doped with silicon either intentionally or not. Capacitance transients are measured in junctions and analysed with the help of several techniques. In these samples, the main contribution due to deep levels appears as a prominent peak in the Fourier transform DLTS (FTDLTS) spectrum with an apparent ionisation energy of 0.95 eV and a capture cross section close to 5 × 10–15 cm2. However, these values must be considered as apparent ones. Isothermal transient spectroscopy (ITS) and a high resolution method based on the analysis of the transients recorded over five decades of time show that several sub-levels exist with ionisation energies between 0.40 and 0.76 eV, and capture cross sections in the range 5 × 10–20– 5 × 10–17 cm2. Capture kinetics confirms this fact, indicating that at least three charge states (either +, 0, – or 0, –, 2–) are involved. The possible origins of long-term instabilities and reversible evolutions of the current–voltage characteristics of the diodes which are also detected, are discussed in relation with these deep states.

Deep levels in MOCVD n-type hexagonal gallium nitride studied by high resolution deep level transient spectroscopy

Deep level transient spectroscopy (DLTS) is performed in MOCVD α-GaN either doped with two different concentrations of Si or unintentionally doped. Capacitance transients are measured in Schottky diodes made of an Au or Ni rectifying contact and an Al/Ti ohmic contact, both on the top of the samples. Only two peaks are detected in each sample in the energy range from the conduction band edge down to 1.1 eV below it, respectively close to 0.50 and 0.92–1.05 eV, by Fourier Transform DLTS (FTDLTS) with concentrations not exceeding 3×10 15 cm −3. These two results testify the high crystalline quality of the samples. The deeper level characteristics depend on the shallow impurity, either Si or the unintentional shallow donor, in deep states which comprise in fact a fine structure not evidenced by FTDLTS. A high resolution DLTS (HRDLTS) method is implemented to resolve this fine structure into several sub-levels which cannot be related to distinct chemical environments. The study of emission and capture kinetics confirms that at least three charge states (+,0,-) are involved. It is concluded that MOCVD α-GaN comprises deep centers which are stabilized in such a form with a concentration in the range of a few 10 14–10 15 cm −3.

Investigation of deep levels and precipitates related to molybdenum in silicon by DLTS and scanning infrared microscopy

We report on the properties of Si diffused with molybdenum (Mo) investigated using experimental techniques that included Fourier transform deep-level transient spectroscopy (FT-DLTS) and scanning infrared microscopy (SIRM). Samples were prepared using boron-doped float zone (1 0 0)Si wafers (100–400 Ω cm) and Mo was diffused into them by placing Mo powder onto the Si surface and annealing in a vacuum (8×10 −6 Torr) at temperatures between 400 and 800°C for 1–10 h. FT-DLTS measurements revealed that deep levels due to Mo ( E v+0.29 eV) were only formed in samples when Mo was diffused above a threshold temperature of 650°C. SIRM imaging showed the presence of Mo-related precipitates having a density of 2.3×10 7–5.8×10 9 cm −3 near the surface region and iron-related precipitates having a density of 1.2×10 7–1.1×10 8 at a depth of 30 μm. The precipitate size was found to be strongly dependent on diffusion temperature and ranged between 50 and 100 nm as calculated from the scattered light intensity. The minority carrier lifetime was found to decrease with increasing density of iron traps that were unintentionally incorporated during the diffusion process.

FTDLTS: a novel isothermal DLTS method using Fourier transforms

A novel isothermal method of analysing the deep impurity levels in semiconductors is proposed. In this method, named FTDLTS (Fourier transform deep-level transient spectroscopy), the transient junction capacitance, whose waveform is considered as a multi-exponential, is decomposed using two Fourier transforms. The FTDLTS spectrum is fitted by a sum of Gaussians using the least-squares method. The numerical calculus of Fourier transforms and the automatic computer-driven setting up of the FTDLTS method are described. On the basis of the same synthetic signals with and without noise, the FTDLTS spectra are compared with those of two conventional DLTS methods. Because of its better time constant resolution, the FTDLTS method allows correct characterization of deep centres with closely spaced levels. Application to an oxygen multi-implanted 650 degrees C annealed GaAs sample is described.