【Domestic Papers】Investigation of Bias Temperature Instability in β-Ga₂O₃ UMOSFET with N-ion Implanted Current Blocking Layers

Abstract

      We report the first comprehensive bias temperature instability characterization of vertical β-Ga₂O₃ UMOSFETs featuring nitrogen-ion-implanted current blocking layers (CBLs). The device exhibits exceptional stability under negative bias stress at - 20 V. This result effectively validates the integrity of the CBL junction, confirming the absence of mobile ions or leakage-inducing defects in the implanted region. Conversely, positive bias stress reveals a threshold voltage (V_TH) instability dominated by electron trapping. Kinetic analysis yields a low time exponent (n=0.12) independent of stress voltage and a near-linear dependence on gate overdrive (γ=1.1), characteristic of tunneling into pre-existing traps. Furthermore, the invariant subthreshold swing and an extracted activation energy of E_A= 0.42 eV suggests that the instability is primarily attributed to phonon-assisted tunneling into border traps, with negligible generation of slow interface states. Recovery analysis based on a universal relaxation model reveals an incomplete restoration, confirming that the permanent component stems from deep traps with extremely long emission times.

Highlights

      For the first time, complete Bias Temperature Instability (BTI) reliability characterization is implemented on vertical β-Ga₂O₃UMOSFET with N-ion implanted current blocking layer, and the device stability difference under negative and positive bias stress is distinguished.

      The N-implanted CBL structure shows negligible threshold voltage shift under heavy negative bias stress of -20 V, verifying perfect junction integrity without mobile ions or leakage defects.

      The kinetic law of Positive Bias Temperature Instability (PBTI) is quantitatively analyzed: time exponent n=0.12 and voltage acceleration factor γ=1.1. Combined with activation energy of 0.42 eV, the dominant degradation mechanism is confirmed as phonon-assisted electron tunneling into gate dielectric border traps instead of massive generation of interface states.

      Multi-dimensional tests including double pulse sweep, temperature-variable stress and relaxation recovery separate recoverable and permanent threshold shift components, proving irreversible degradation originates from deep-level traps.

      This work identifies that ALD-Al₂O₃ gate dielectric defects are the primary reliability bottleneck, providing core optimization guidance for gate stack design of Ga₂O₃ power devices.

Conclusion

      In this work, we presented the first comprehensive reliability assessment vertical β-Ga₂O₃ UMOSFETs featuring CBL. First, the device demonstrates exceptional robustness under NBS (- 20 V), showing negligible threshold voltage shift. Consistent with the deep depletion nature of the β-Ga₂O₃ interface, this stability confirms that the CBL preserved the junction integrity without introducing mobile ions or leakage channels. Second, the primary reliability bottleneck is identified as Positive Bias Instability caused by electron trapping. Through rigorous kinetic analysis, we observed an invariant subthreshold swing, a low time exponent (n=0.12), and a near-linear dependence on gate overdrive (γ=1.1). These signatures suggest that the degradation is governed by phonon-assisted tunneling into pre-existing border traps, rather than the generation of slow interface states detectable within the measurement window. The incomplete recovery observed even after 10⁴ s verifies the presence of deep border traps and the propagation of the electron tunneling front into the dielectric. The findings clarify that future efforts to enhance the dynamic stability of vertical Ga₂O₃ power devices should prioritize the optimization of the gate dielectric quality.

Project Support

      This work is supported by the National Key Research and Development Program of China (No. 2024YFE0205200), the National Natural Science Foundation of China under Grant nos. 62404214, 62522411, U23A20358, 62234007, 62474170 and 61925110, the University of Science and Technology of China (USTC) Research Funds of the Double First-Class Initiative under Grant No. WK2100000055, the Project of 46th Research Institute of CETC under Grant no. WDZC202446007, the open research fund of Suzhou Laboratory under Grant no. SZLAB-1208-2024-ZD012.