Nitrogen Annealing As a Sustainable Method for Interface Trap Passivation in 4H-SiC Mosfets

Silicon Carbide (4H-SiC) has emerged as a leading wide band gap semiconductor for high-power, high-temperature applications 1 . 4H-SiC metal-oxide semiconductor-field-effect transistors (MOSFETs) have lower power dissipation compared to silicon, allowing for low-noise and high-efficiency all-electric vehicle drives, fast-charging stations, solar inverters, and more. While these devices provide substantial advancements for next-generation energy efficient power systems, 4H-SiC may also offer additional functionality in the form of integrated circuits (ICs) at high temperatures (>300 °C). Because of its high noise immunity and low static power consumption, lateral complementary-metal-oxide-semiconductor (CMOS) IC technology in 4H-SiC is desirable for large-scale integration 2 . This technology necessitates the use of both n- and p-channel MOSFETs that can operate at high temperatures. Despite the advances of 4H-SiC MOSFETs, the high density of interface states (D it ) at the 4H-SiC/SiO 2 interface prevents reaching full potential resulting in high channel resistance and low mobility. Alternatives to nitric oxide (NO) annealing, the most common method adopted to reduce D it in 4H-SiC 3,4 , are actively sought due to its toxicity and relatively expensive cost. Annealing in pure nitrogen (N 2 ) 5,6 at high temperatures (1400 °C-1600 °C) has been recently demonstrated promising results for 4H-SiC MOSFET processing. In this work, we report D it measurements consistent with [6] and attempt to correlate the nitrogen areal densities of the near interfacial regions with the D it for high temperature N 2 annealing processes compared to NO. In our study, metal oxide semiconductor capacitors were fabricated on p- and n-type 4H-SiC epitaxial layers. Gate oxides were thermally grown at 1150 °C for 10 h in dry O 2 resulting in a ~ 60 nm thick oxide layer. Selected samples are then annealed in flowing N 2 at high temperatures (1400 °C, 1 h; 1450 °C, 1 h; and 1500 °C, 30 minutes or 1 h) or NO (1175 °C, 2 h). X-ray photoelectron spectroscopy (XPS), carried out after etching the oxide, indicates that the amount of nitrogen at the interface due to high temperature N 2 annealing is ~ 4 × higher than NO annealed devices. Simultaneous high frequency (100 kHz)- low frequency CV was performed to extract interface trap densities (D it ) for each process and compared at room temperature (27 °C) with reference 1175 °C, 2 h NO annealed samples. The comparison reveals that, N 2 annealing at 1500 °C for 30 minutes with a flow rate of 3 LPM results in D it values comparable to NO annealing across the bandgap. Moreover, nitrogen annealing is more effective in reducing D it near the valence band than NO annealing, while the opposite is true close to the conduction band-edge, consistent with previous reports [6] and observed in atomistic models of these interfaces using the density functional theory 7 . Nitrogen annealing also decreases the positive fixed charges at the interface of p-type 4H-SiC and SiO 2 , as evidenced from the flat band voltage comparison. The oxide breakdown voltage for the devices made with 1500 °C N 2 annealing was similar to that of NO annealed devices. XPS analysis of the N 2 annealed devices, their behavior under high temperature and bias, and their potential to substitute NO will be further discussed. The authors gratefully acknowledge the support from the National Renewable Energy Laboratory/ US Department of Energy sub-contract NREL-AHL-9-92362-01. References: 1 T. Kimoto and J.A. Cooper, Fundamentals of Silicon Carbide Technology: Growth, Characterization, Devices and Applications (John Wiley & Sons, 2014). 2 D. Liu and C. Svensson, IEEE J. Solid-State Circuits 29 , 663 (1994). 3 G. Liu, B.R. Tuttle, and S. Dhar, Appl. Phys. Rev. 2 , 021307 (2015). 4 S. Das, T. Isaacs-Smith, A. Ahyi, M.A. Kuroda, and S. Dhar, J. Appl. Phys. 130 , 225701 (2021). 5 A. Chanthaphan, T. Hosoi, T. Shimura, and H. Watanabe, AIP Adv. 5 , 097134 (2015). 6 K. Tachiki and T. Kimoto, IEEE Trans. Electron Devices 68 , 638 (2021). 7 L. Wang, S. Dhar, L.C. Feldman, and M.A. Kuroda, Phys. Status Solidi B 259 , 2100224 (2022). Figure 1