1 INTRODUCTION

Gallium nitride is currently considered to be one of the most promising wide-band materials of semiconductor electronics. A large band gap Eg = 3.4 eV and breakdown field strength Ei ~ 3 MV/cm, which is an order of magnitude higher than the Ei value in silicon (~0.3 MV/cm), provide an opportunity to design GaN Schottky barrier diodes (SBDs) with blocking voltage Ub in excess of 1 kV and a near-unity ideality factor η [1–3].

The band gap of GaN is only slightly greater than Eg 4H-SiC (3.34 eV), which has already found wide application (see, e.g., [4]). However, GaN has several important potential advantages over silicon carbide: a higher electron mobility; a direct band gap, which allows one to construct efficient optoelectronic devices based on GaN; and the applicability of GaN/AlGaN heterostructures in devices with a two-dimensional electron gas with a high mobility.

The resistance of semiconductor devices to various types of irradiation (specifically, proton and electron irradiation) often dictates the feasibility and conditions of use of such devices in electronic systems of nuclear reactors, particle accelerators, and space and aviation electronics. The resistance of SiC-based devices to electron and proton irradiation was examined in a number of studies (see the corresponding references in [5, 6]). It was found that one of the most important parameters (electron removal rate ηe) under electron irradiation may vary by more than 2 orders of magnitude (from 0.015 [7] to 1.67 cm–1 [8]) depending on the electron energy, the material fabrication method, and the doping nature and level. In the case of proton irradiation of SiC, carrier removal rate ηp falls within the range from ~10 [9] to ~110 cm–1 [10].

The electron removal rate in the course of both electron and proton irradiation of n-type GaN also depends on the irradiation energy and dose, the fabrication method and the initial carrier concentration of GaN, and on the dislocation density in the irradiated material [11]. The values of ηe vary with these parameters, falling within the range from ~10–1 to 10 cm–1 [12]. An increase in the carrier concentration in GaN SBDs subjected to proton irradiation was observed in [13]. As was noted in reviews [11, 14], this effect may be indicative of the formation of shallow donor levels under irradiation and is possibly attributable to an insufficient purity of the initial epitaxial layers. In all the other cases, proton irradiation led to removal of electrons from the conduction band. The determined values of ηp range widely from 40 [15] to 104 cm–1 [16].

In the present study, the effect of irradiation with electrons with an energy of 0.9 MeV and protons with an energy of 15 MeV on the parameters of SBDs based on test GaN structures grown by metalorganic vapor-phase epitaxy is examined. Rates of carrier removal ηe and ηp from the base layers of the studied structures are determined.

2 EXPERIMENTAL CONDITIONS

The studied structures were grown by metalorganic vapor-phase epitaxy (MOVPE) on (0001) sapphire substrates 2 inches in diameter with the use of standard compounds in a Dragon 125 setup with a horizontal reactor with induction heating. A 2.4-μm-thick buffer layer of undoped GaN was grown first on a substrate, and layers doped heavily and weakly with silicon, each with a thickness of ~1 μm, were grown after that. The concentration of electrons in these layers determined from capacitance-voltage measurements was 6 × 1018 and 8 × 1016 cm–3, respectively. The end stage of growth was in situ deposition of a thin passivating Si3N4 dielectric layer that suppressed leakage currents [17]. Nickel contacts 600 μm in diameter, which formed Schottky barriers, were fabricated by thermal deposition of Ni through a shadow mask.

Irradiation with protons with an energy of 15 MeV was performed in the pulsed mode at an MGTs-20 cyclotron. The repetition rate and the duration of pulses were 100 Hz and 2.5 ms, respectively. Irradiation with electrons with an energy of 0.9 MeV was performed in the pulsed mode with the repetition rate and the duration of pulses set to 490 Hz and 330 μs, respectively. Proton and electron irradiation was performed at room temperature. The temperature in these experiments was maintained with an accuracy of ±5°C.

Isothermal current–voltage curves of diodes were measured at room temperature in the single-pulse mode. The pulse duration was 5 μs, and the repetition rate was 100 Hz.

3 RESULTS AND DISCUSSION

Figure 1 shows the forward current-voltage curves of the initial diode (curve 1) and diodes irradiated with four doses of protons with an energy of 15 MeV at room temperature.

Fig. 1.
figure 1

Forward current–voltage curves of diodes after irradiation with protons with an energy of 15 MeV at different fluences Ф, cm−2: 1—0, 2—2 × 1014, 3—4 × 1014, and 4—5 × 1014. The dependence of electron concentration in the diode base on fluence Ф in shown in the inset. (A color version of the figure is provided in the online version of the paper.)

Full size image

At all fluences Ф, current-voltage curves were measured within the following range of current densities: 5 × 10–6 ≤ j ≤ 1 A/cm2. As in the case of SiC Schottky diodes, irradiation has almost no effect on the current-voltage curves under biases U lower than cut-off voltage Uc (when almost the entire applied voltage falls on the Schottky barrier and dependence I(U) is exponential; see, e.g., [6]).

At U > Uc, the differential base resistance of diodes increases monotonically with increasing fluence Ф. The variation of mobility under irradiation may be neglected at relatively low values of Ф [18]. The electron concentration is then proportional to the differential base resistance, and rate ηρ of electron removal from the base under irradiation may be calculated as ηp = (n0 – n)/Ф, where n0 is the electron concentration in the base in the initial sample and n is the concentration after irradiation with fluence Ф.

The inset in Fig. 1 demonstrates that the carrier concentration decreases linearly with increasing fluence at relatively small values of Ф. The slope of dependence n(Ф) within this section corresponds to carrier removal rate ηp ≈ 140 cm–1. With this n(Ф) dependence, condition n = 0 should be satisfied at Ф ≈ 5.5 × 1014 cm–2 (dashed line in the inset in Fig. 1). However, the value of n at Ф = 5 × 1014 cm–2 is significantly higher than the one corresponding to a linear n(Ф) dependence. According to the analysis reported in [19], this result may indicate that GaN differs from SiC in supporting the following compensation mechanism under proton irradiation: a radiation-induced defect (vacancy) interacts with a shallow impurity atom, forming an electrically neutral or acceptor center. This compensation mechanism is typical, for example, in the case of electron irradiation of Si.

Figure 2 shows the forward current–voltage curves of the initial diode (curve 1) and diodes irradiated with three doses of electrons with an energy of 0.9 MeV at room temperature.

Fig. 2.
figure 2

Forward current–voltage curves of diodes after irradiation with electrons with an energy of 0.9 MeV at different fluences Ф, cm−2: 1—0, 2—2 × 1016, 3—4 × 1016, and 4—6 × 1016. The dependence of carrier concentration in the diode base on fluence Ф in shown in the inset.

Full size image

Current–voltage curves were measured within the 1 × 10–8 ≤ j ≤ 1 A/cm2 current density range. As in the case of proton irradiation, irradiation with electrons has almost no effect on current–voltage curves under biases U < Uc.

In the 0 ≤ Ф ≤ 6 × 1016 cm–2 region, the differential base resistance increases monotonically with increasing Ф. If one follows the same approach as the one used for proton irradiation and neglects the change in mobility, it becomes easy to calculate the variation of electron concentration with fluence Ф (see the inset in Fig. 2) based on the data from Fig. 2. It is evident that the electron concentration decreases linearly with increasing fluence. The slope of dependence n(Ф) corresponds to electron removal rate ηe ≈ 0.47 cm–1.

The determined value of ηp ≈ 140 cm–1 is close to the lower boundary of the range of carrier removal rates under proton irradiation and indicates a sufficiently high level of radiation resistance. The ηe ≈ 0.47 cm–1 value corresponds roughly to the center of the literature range of carrier removal rates under electron irradiation.

4 CONCLUSION

In conclusion, it should be noted that the value of ηp ≈ 140 cm–1 determined for n-type GaN is close to the lower boundary of the range of carrier removal rates under proton irradiation and indicates a sufficiently high level of radiation resistance (specifically, a level comparable to the radiation resistance of n-type silicon carbide). The electron removal rate of ηe ≈ 0.47 cm–1 determined in the study corresponds roughly to the center of the literature range of ηe values for n-type GaN. This value is also comparable to ηe levels typical of n-type SiC.