Al ions were implanted with multiple energies up to 250 keV at elevated temperatures in n-type 4H SiC epitaxial layers to reach a surface concentration of 1 x 10(20) cm(-3). These samples were then annealed at temperatures between 1500 and 1950 degrees C. A similar 4H SiC epitaxial sample was implanted by MeV Al ions to lower doses and annealed only at 200 and 400 degrees C. After annealing, cross-sections of the samples were characterized by scanning spreading resistance microscopy (SSRM). The results show that the resistivity of high-dose Al implanted samples has not reached a saturated value, even after annealing at the highest temperature. For the MeV Al implanted sample, the activation of Al has not yet started, but a substantial annealing of the implantation induced damage can be seen from the SSRM depth profiles.

To obtain a better understanding of the damage annealing process and dopant defect incorporation and activation we have implanted epitaxially grown 4H-SiC layers with high doses of Al+ ions. Cross-sections of the samples are investigated by scanning spreading resistance microscopy (SSRM) using a commercial atomic force microscopy (AFM). The defects caused by the implanted ions compensate for the doping and decrease the charge carrier mobility. This causes the resistivity to increase in the as-implanted regions.The calculated profile of implanted ions is in good agreement with the measured ones and shows a skewed Gaussian shape. Implanted samples are annealed up to 400 degrees C. Despite these low annealing temperatures we observe a clear improvement of the sample conductivity in the asimplanted region. (c) 2006 Published by Elsevier B.V.

Results from the application of scanning spreading resistance microscopy (SSRM) for characterization of aluminum implanted 4H–SiC are presented. The implanted profiles are investigated electrically and morphologically as a function of post-implantation anneal conditions. The method is shown to be advantageous for measuring and optimizing the activation in many aspects with respect to existing alternative techniques: it provides information of the entire depth and Al concentration range, it is unaffected by annealing induced re-growth and/or surface roughening, and requires little sample preparation. The results indicate that the apparent activation and surface roughness do not saturate in the investigated temperature range of 1500–1650°C. Finally, an apparent activation energy for the process of 3 eV is estimated.

The formation of pn junctions in deep silicon pores has been studied for a new concept of X-ray imaging detectors. The sensitive part of the device is an array of CsI(Tl) columns formed by filling a silicon matrix of pores having pn junctions in their walls. Under X-ray illumination, the CsI(Tl) scintillator emits photons that are collected by the pn junctions. Relatively high signal collection efficiency is expected. However, the formation of pn junctions inside pore walls represents a challenging step in the detector fabrication. In this work pore matrices were fabricated in n-type silicon by deep reactive ion etching and by photo-electrochemical etching. The pn junctions were formed either by boron diffusion or deposition of boron doped poly-silicon. Various techniques were used to analyze the transverse depth profiles of boron atoms at different pore depths. The study shows successful results for pn-junctions formed both by diffusion and by poly-silicon deposition.

Ion implantation is considered a key technology for the realisation of silicon carbide electronic devices. Here we will give an overview of the field and present some recent results of ion implante ...

A strong contrast was observed in a conventional scanning electron microscope resembling the shape of stacking faults in epitaxial 4H silicon carbide layers. This contrast was seen at the location of degradation faults in bipolar diodes as well as far outside the device perimeter on the low n-type doped epitaxial layer. Photoluminescence measurements indicated that the nature of the observed features were stacking faults in the crystalline material. The intersections of the stacking faults with the sample surface was marked by long edges in the pattern of the step-bunching.

Free carrier distributions in 4H-SiC have been studied by scanning spreading resistance microscopy (SSRM). The SSRM signal is discussed in relation to the chemical impurity concentration measured by secondary ion mass spectrometry (SIMS). The method is demonstrated on a p-type Al epi-layer staircase structure with doping concentration ranging from 2x10(16)cm(-3) to 2x10(20) cm(-3) and an implanted acceptor dopant profile of shallow Al and deep B. The results of the epi-layer investigation are in good agreement with the macroscopic spreading resistance model R=rho/4r, and show a much higher dynamic range in 4H-SiC than conventional spreading resistance profiling. Measurements of the implantation profile reveal a resistance in the highly doped region (> 10(18) cm(-3)) of more than two orders of magnitude larger than that measured in the epi-layer with similar concentration. The observation may be attributed to remaining implantation-induced defects. The SSRM measurements also show a diffusion of B consistent with SIMS data. The investigation clearly demonstrates that the technique is a valuable method to determine high gradient carrier concentrations in p-type SiC, although the detailed mechanism of the spreading resistance current is presently not completely understood.

In boron implantation manufactured high-voltage 4H-SiC p(+)n diodes, a deep level was found by capacitance transient measurements to strongly influence the position of the pn junction at room temperature (RT). The trap effectively removed free carriers within the region of in-diffused boron making this region intrinsic. Scanning spreading resistance measurements confirmed the existence of this intrinsic region. The electrical measurements at elevated temperatures showed a decrease in depletion width compared to RT, attributed to the thermal activation of the deep boron level. It was shown that this level can cause an intrinsic region either as a deep acceptor or as a hole trap and these measurements could not determine the electronic nature of the level.

Epitaxially-grown Al-doped 4H-SiC has been studied by scanning spreading resistance microscopy. The measured current shows good quantitative agreement with the chemical Al concentration in the range 2×1016 to 2×1020 atoms cm−3. Simulations of the sample temperature distribution using finite element calculations predict a maximum temperature exceeding 750 K within 100 nm of the contact region at 7.5 V dc bias for an Al doping of 1020 cm−3. The heating causes a significant increase in the ionization of the dopants relative to that at room temperature. Due to the strong voltage dependence, the effect can be avoided by operating below 5 V dc bias where the temperature rise is shown to be negligible for all dopant concentrations.

High-temperature chemical vapour deposition (HTCVD) in a vertical chimney reactor was used to grow thick low-doped epitaxial layers of 4H silicon carbide. These layers were used as drift layers in a combined process to manufacture both bipolar and unipolar high-voltage diodes. The resulting diodes were characterized electrically in order to gain knowledge about the electric quality of the HTCVD epitaxial layers to assess the high-voltage properties of this material.

Three different methods for measuring the depth distribution of dopants in 4H-SiC have been investigated: (I) Spreading Resistance profiling (SRP), (2) Scanning Capacitance Microscopy (SCM) and (3) ...

A brief survey is given of some recent results on doping of 4H- and 6H-SiC by ion implantation. The doses and energies used are between 10(9) and 10(15) cm(-2) and 100 keV and 5 MeV, respectively, ...

Diodes have been manufactured of 4H SiC and investigated under static forward and reverse bias. Current-voltage characteristics have been correlated with electroluminescence and Electron Beam Induced Current (EBIC) measurements. During reverse bias structural defects give rise to early microplasma breakdown, but in most cases the reverse current density saturates after a sudden increase of several orders of magnitude. After saturation of the activated defect, the device can withstand an increased bias up to the point where planar breakdown occurs. Electroluminescence from micropipe defects is shown to consist of 6 individual microplasmas, having turn-on voltages separated by less than 0.1 V.