The paper presents the damage accumulation in silicon carbide (SiC) as a function of the ion mass, the ion energy and the implantation temperature. A defect-interaction and amorphization model is used to analyse the dose dependence of defect production, as obtained by the various methods. The temperature dependence of the amorphization dose can be represented assuming a thermally enhanced annealing within the primary collision cascades. On the basis of such a model, a critical implantation temperature is obtained, which was found to vary with the ion mass and the implantation energy. The concurrent influence of implantation temperature and ion fluence on the resulting damage distribution in SiC is demonstrated. The damage annealing of ion implanted SiC is investigated for low, medium and high damage concentrations. The effect of the implantation temperature and the concentration of implanted atoms, both influencing the kind of defects obtained after implantation, on the annealing behaviour is analysed.

Ion-beam induced damage and annealing behaviour in SiC

Technological breakthroughs in growth control of SiC are reviewed. Step-controlled epitaxy by using off-axis SiC {0001} substrates to grow high-quality epitaxial layer is explained in detail. The introduction of substrate off-angles brings step-flow growth, which easily makes polytype replication of SiC at rather low temperatures. Off-angle dependence, rate-determining processes, and temperature dependence of growth rate are discussed. Prediction, whether step-flow growth or two-dimensional nucleation does occur, is given as a function of off-angle, growth temperature, and growth rate. Optical and electrical properties of undoped epitaxial layers are characterized. Impurity doping during the growth is explained. Recent progresses in peripheral technologies for realization of power electronic devices, such as bulk growth, epitaxial growth, ion implantation, MOS interface, ohmic contacts, are introduced. Finally application to high-power electronic devices is briefly described.

https://iopscience.iop.org/article/10.1143/JJAP.43.6835/pdf

Technological Breakthroughs in Growth Control of Silicon Carbide for High Power Electronic Devices

Silicon carbide semiconductor devices and methods of fabricating silicon carbide semiconductor devices have a silicon carbide DMOSFET and an integral silicon carbide Schottky diode configured to at least partially bypass a built in diode of the DMOSFET. The Schottky diode may be a junction barrier Schottky diode and may have a turn-on voltage lower than a turn-on voltage of a built-in body diode of the DMOSFET. The Schottky diode may have an active area less than an active area of the DMOSFET.

Silicon carbide MOSFETs with integrated antiparallel junction barrier Schottky free wheeling diodes and methods of fabricating the same

High-dose phosphorus-ion (P+) implantation into 4H–SiC (0001) followed by high-temperature annealing has been investigated. Annealing with a graphite cap largely suppressed the surface roughening of implanted SiC. The surface stoichiometry of implanted SiC was examined by x-ray photoelectron spectroscopy. Electronic behaviors of P+-implanted SiC are discussed based on Hall effect measurements. There is no significant difference in the sheet resistance between SiC annealed with a graphite cap and without a graphite cap. The sheet resistance (resistivity) takes a minimum value of 45 Ω/□ (0.9 mΩcm) at an implant dose of 6.0×1016 cm−2. The sheet resistance shows a weak temperature dependence.

/pdf/electronic-behaviors-of-high-dose-phosphorus-ion-implanted-ly2m7nep8w.pdf

Electronic behaviors of high-dose phosphorus-ion implanted 4H-SiC(0001)

Silicon carbide (SiC) devices have the potential to yield new components with functional capabilities that far exceed components based on silicon devices. Selective doping of SiC by ion implantation is an important fabrication technology that must be completely understood if SiC devices are to achieve their potential. One major problem with ion implantation into SiC is the surface roughening that results from annealing SiC at the high temperatures which are needed to activate implanted acceptor ions, boron or aluminum. This paper examines the causes and possible solutions to surface roughening of implanted and annealed 4H-SiC. Samples consisting of n-type epilayers (5 × 1015 cm−3, 4 µm thick) on 4H-SiC substrates were implanted with B or Al to a total dose of 4 × 1014 cm−2 or 2 × 1015 cm−2, respectively. Roughness measurements were made using atomic force microscopy. From the variation of root mean square (rms) roughness with annealing temperature, apparent activation energies for roughening following implantation with Al and B were 1.1 and 2.2 eV, respectively, when annealed in argon. Time-dependent activation and surface morphology analyses show a sublinear dependence of implant activation on time; activation percentages after a 5 min anneal following boron implantation are about a factor of two less than after a 40 min anneal. The rms surface roughness remained relatively constant with time for anneals in argon at 1750°C. Roughness values at this temperature were approximately 8.0 nm. Annealing experiments performed in different ambients demonstrated the benefits of using silane to maintain good surface morphology. Roughnesses were 1.0 nm (rms) when boron or aluminum implants were annealed in silane at 1700°C, but were about 8 and 11 nm for B and Al, respectively, when annealed in argon at the same temperature.

Surface roughening in ion implanted 4H-silicon carbide

Phosphorus and boron ion implantations were performed at various energies in the 50 keV–4 MeV range. Range statistics of P+ and B+ were established by analyzing the as-implanted secondary ion mass spectrometry depth profiles. Anneals were conducted in the temperature range of 1400–1700 °C using either a conventional resistive heating ceramic processing furnace or a microwave annealing station. The P implant was found to be stable at any annealing temperature investigated, but the B redistributed during the annealing process. The implant damage is effectively annealed as indicated by Rutherford backscattering measurements. For the 250 keV/1.2×1015 cm−2 P implant, annealed at 1600 °C for 15 min, the measured donor activation at room temperature is 34% with a sheet resistance of 4.8×102 Ω/□. The p-type conduction could not be measured for the B implants.

Phosphorus and boron implantation in 6H–SiC

A1 and B implantations were performed into n-type 6H-bulk SiC and epitaxial layers at both room temperature and 850°C. Annealings were performed in the temperature range of 1100–1650°C in a SiC crucible. For single-energy implants, the implant gettered to the 0.7Rp location for annealing temperatures ≥1400°C. For the 850°C implanted samples the RBS yield in the annealed material is comparable to the yield in the as-grown material, indicating a good lattice recovery. A maximum activation of 18% for Al-implanted samples was observed. PN junction diodes were made using Al-implanted material.

PN junction formation in 6HSiC by acceptor implantation into n-type substrate

Multiple-energy Al implantations were performed with and without C or Si coimplantations into 6H-SiC epitaxial layers and bulk substrates at 850°C. The C and Si co-implantations were used as an attempt to improve Al acceptor activation in SiC. The implanted material was annealed at 1500, 1600, and 1650°C for 45 min. The Al implants are thermally stable at all annealing temperatures and Rutherford backscattering via channeling spectra indicated good lattice quality in the annealed Al-implanted material. A net hole concentration of 8 × 1018 cm−3 was measured at room temperature in the layers implanted with Al and annealed at 1600°C. The C or Si co-implantations did not yield improvement in Al acceptor activation. The co-implants resulted in a relatively poor crystal quality due to more lattice damage compared to Al implantation alone. The out-diffusion of Al at the surface is more for 5Si co-implantation compared to Al implant alone, where 5Si means a Si/Al dose ratio of 5.

Al, Al/C and Al/Si implantations in 6H-SiC

Elevated temperature (ET) multiple energy N, P, and N/P implantations were performed into p-type 6H-SiC epitaxial layers. For comparison, room temperature (RT) N and P implantations were also performed. In the N/P coimplanted material a sheet resistance of 2.1×102 Ω/□ was measured, which is lower compared to the values measured in N or P implanted material of the same net donor dose. The RT P implantation resulted in heavy lattice damage and consequently low P electrical activation, even after 1600 °C annealing. After annealing the Rutherford backscattering yield either coincided or came close to the virgin level for ET implantations and RT N implantation, whereas for RT P implantation the yield was high, indicating the presence of high residual damage. Vertical n-p junction diodes were made by selective area ET N, P, and N/P implantations and RT N and P implantations using a 2.5 μm thick SiO2 layer as an implant mask. The diodes were characterized by capacitance–voltage and variable temperature current–v...

Material and n-p junction properties of N-, P-, and N/P-implanted SiC

Elevated temperature (700°C) N ion implantations were performed into 6H-SiC in the energy range of 50 keV-4 MeV. By analyzing the as-implanted depth distributions, the range statistics of the N+ in 6H-SiC have been established over this energy range. Annealing at 1500 and 1600°C for 15 min resulted in Rutherford backscattering spectrometry scattering yields at the virgin crystal level, indicating a good recovery of the crystalline quality of the material without any redistribution of the dopant. A maximum electron concentration of 2 × 1019 cm−3, at room temperature, has been measured even for high-dose implants. The p-n junction diodes made by N ion implantation into a p-type substrate have a forward turn-on voltage of 2.2 V, an ideality factor of 1.90, and a reverse breakdown voltage of 125 V with nA range leakage current for -10 V bias at room temperature. By probing many devices on the same substrate we found uniform forward and reverse characteristics across the crystal.

Jason A. Gardner

Papers

Phosphorus and boron implantation in 6H–SiC

PN junction formation in 6HSiC by acceptor implantation into n-type substrate

Al, Al/C and Al/Si implantations in 6H-SiC

Material and n-p junction properties of N-, P-, and N/P-implanted SiC

Elevated temperature nitrogen implants in 6H-SiC