Abstract
The effects of gamma irradiation on the electrical and trapping properties of AlGaN/GaN high-electron-mobility transistors (HEMTs) are investigated in detail. During the irradiation, the gate–source leakage current of the HEMT is monitored online when applying a reverse gate voltage. The variations of electrical properties of the device, including an increase in drain–source current, the negative threshold voltage shift, and a decrease of leakage current, are observed. In particular, three traps in the device are identified using the voltage-transient method and the variations of these traps after irradiation are also investigated. The results show that the absolute amplitudes of the three traps in the device decrease after irradiation, which indicates a reduction in the density of the traps. Furthermore, it is proposed that the time constants and energy levels of the three traps decrease after irradiation. The observed changes in the trapping behaviors are ascribed to the structural ordering of the defects, which is consistent with the improvement in the electrical characteristics of the device.
Export citation and abstract BibTeX RIS
1. Introduction
GaN-based high-electron-mobility transistors (HEMTs) have shown outstanding performance because of their high power densities, low on-resistance and high thermal conductivity [1–3]. In addition, because of their low thermal generation rates and high breakdown fields, these devices have also shown great promise for application in satellite and military systems in hard radiation environment [4, 5]. Therefore, it is crucial to study and understand the mechanism of interaction of gamma rays with the material and devices [6–9].
The gamma irradiation degradation as well as improvements in GaN devices properties have been discussed in literature. For example, Vitusevich et al observed the deterioration of the HEMTs at a total dose higher than 1 Mrad [5]. Additionally, O. Aktas et al reported that the drain–source current and transconductance of HEMTs increased at a dose up to 600 Mrad [10]. Specifically, a gate degradation was identified with a negative gate–source voltage (VGS) applied to the HEMT at a dose of 3.9 Mrad irradiation [11]. The effects of irradiation on the devices depend on the structures, technologies and experiment conditions [12, 13], which may induce the different experiment results [6, 14]. However, the change in device performance was normally characterized after irradiation, whereas the improvement or degradation of performance was not monitored during the irradiation process. Furthermore, it was concluded that gamma irradiation can introduce additional traps or reconfigure the pre-existing traps, affecting the electrical characteristics of AlGaN/GaN HEMTs [15]. Although a great deal of research has been conducted on the electrical properties in GaN devices after gamma irradiation, a relatively limited amount of insight exists into the trapping behaviors of irradiated GaN devices. In previous studies, the time-domain measurements of drain–source current (IDS) on HEMTs were performed before and after irradiation [5, 16], whereas the information of trapping behaviors included in the curves was not discussed in detail. An analysis of the change of trapping phenomenon and trap level after irradiation should also be carried out, to acquire a deep understanding of the gamma irradiation on the HEMTs.
In this paper, we examined the effects of 3.5 Mrad gamma irradiation on the GaN HEMT when applying a reverse gate voltage. During the irradiation, the negative transient variations of gate–source current (IGS) of the HEMT was monitored online. The electrical characteristics of the HEMTs were measured before and after gamma irradiation. In particular, we investigated the trapping effects in the HEMT before and after irradiation using the voltage-transient method [17] and the differential amplitude spectrum (DAS) [18]. With these methods, changes in the amplitudes, time constants, and energy levels of the HEMT can be determined clearly and the possible origin of the change was discussed.
2. Experimental details
Figure 1 shows the structure of the device used in this work. The schematic diagram of the cross section of the GaN HEMT is shown in figure 1(a). The device was grown on a 4 H-SiC substrate by metalorganic chemical vapor deposition. A Ti/Al/Pt/Au structure was used to form the source/drain ohmic electrodes and Ni/Au structure was used as the Schottky gate. The thicknesses of the AlGaN layer, the GaN layer, and the chip were 18 nm, 2 μm, and 80 μm, respectively. The Al composition was 22%. The device was composed of ten fingers with a gate length of 0.25 μm and gate width of 1.25 mm. Figure 1(b) shows the SEM image of the device using the tabletop microscope (HITACHI TM4000Plus).
3. Methodology
The device was irradiated with 60Co-γ rays for a total dose of 3.5 MRad for 100 min at a dose rate of 583 rad s−1. A reverse VGS of −5 V was applied to the device and the drain was floating during the irradiation, which was based on the previous study [11]. To observe the transient variation of IGS during the irradiation, the gate–source leakage current was monitored online during the entire irradiation process using the data acquisition unit (Agilent 34 970 A).
After irradiation, the device was kept with no bias at room temperature for at least 4 h to ensure that the device performance was stable [11, 19–21]. The electrical properties were characterized using a semiconductor parameter analyzer (Agilent B1500A). In addition, the drain–source transient voltages were also measured using the voltage-transient method [17] and the DAS [18] to investigate the trapping behavior occurring before and after the irradiation. To ensure that the device was in the same initial condition after each detrapping transient measurement, it was recovered completely by shining a 365 nm ultraviolet light of 1 W for 30 s [18, 22].
4. Results and discussion
4.1. Leakage current characteristics
Figure 2(a) presents the transient variation of IGS at VGS = −5 V during the gamma irradiation. The inset shows that IGS fluctuated slightly at −1.36 μA during the irradiation. Figure 2(b) shows the IGS before and after irradiation, which was characterized with the drain floating as well. In contrast, the leakage current decreased after irradiation, whereas the variation of the forward gate–source current was negligible. The point marked in figure 2(b) corresponds to the bias conditions of online monitoring in figure 2(a). It can be seen that the variation of IGS is minor under the condition of VGS = −5 V, and thus the transient curve in figure 2(a) may not display the slight decrease of leakage current during gamma irradiation completely.
Download figure:
Standard image High-resolution imageThe result that the leakage current of the HEMT decreased after gamma irradiation has been reported in previous studies [4, 6, 23, 24]. Chandan et al suggested that the reduction in leakage current after gamma irradiation corroborates the redistribution of defects [4]. Reference [24] concluded that the gate leakage of HEMTs significantly decreased after irradiation in the low bias region where surface generation recombination is dominant and also at higher voltage, due to an increase in channel resistance. The analysis and verification of this result in our experiment will be given in subsequent sections.
4.2. Transfer and output curves characteristics
The transfer and output curves of the device before and after irradiation are shown in figures 3(a) and (b), respectively. Specifically, figure 3(a) shows the transfer curves at drain–source voltage VDS = 1 V [5, 11, 19, 20]. Under the conditions of VGS = 0 V and VDS = 1 V, the value of the IDS increases from 375 to 439 mA, indicating the increase in the IDS in the linear region. The maximum value of transconductance (gm) also increases from 0.271 to 0.315 S at VDS = 1 V after irradiation. In addition, the threshold voltage (Vth) shifts toward the reverse direction slightly. Figure 3(b) clearly shows the increase in the IDS in the saturation region. The value of the saturated IDS increases from 465 to 579 mA at VGS = −1 V.
Download figure:
Standard image High-resolution imageThese results suggest that the electrical properties of the device were improved after gamma irradiation, which is consistent with the decrease of the leakage current. Several investigations have shown that there is a significant improvement for dc device performance characteristics after gamma irradiation [4, 5, 10, 24, 25]. Aktas et al suggested that the the threshold voltage shift can be explained by the creation of mutually compensating acceptor and donor defects [10]. Reference [5] concluded that the improvement of saturation current and transconductance can be attributed to the structural ordering of native defects. In addition, it is indicated that the threshold voltage shift and the increase in the drain–source current can be ascribed to the decrease in surface roughness and redistribution of traps at the AlGaN/GaN interface [4]. In our experiments, this improvement can be attributed to the reduction in the density of traps, along with the ordering of native defects. The details and analysis will be discussed in subsequent sections.
4.3. Comparison of the detrapping transients
To further analyze the effects of gamma irradiation on the trapping behaviors, the drain–source transient voltages before and after gamma irradiation were measured under the same bias conditions. Detailed information about the measurement methods can be found in [17] and [18]. The biasing sequence used in the experiments is illustrated in figure 4(a). During the filling process, carrier trapping was induced by applying the filling conditions of (VGF; VDF) = (−7 V; 10 V) for 30 s [17, 26]. After that, we monitored the recovery voltage transient with bias conditions of (VGM; IDM) = (0 V; 200 mA) from 20 μs to 120 s. The VDS under these measurement conditions was approximately 0.5 V and the thermal resistance value (Rth) extracted using the structure function method was 9.7 °C W−1 [27]. Therefore, the power in the experiments was 0.1 W (P= VDS × IDS) and the highest temperature rise of 0.97 °C (ΔT = P × Rth) can hardly affect the measurement results.
Download figure:
Standard image High-resolution imageFigure 4(b) shows that the total variation of the drain–source voltage transients decreased, which indicated that the charge trapping was reduced after irradiation and is consistent with the results reported in previous studies [4, 16, 28, 29]. Three detrapping behaviors, designated DP1, DP2, and DP3, were identified on the time constant spectrum shown in figure 4(c), indicating a reduction of time constants after irradiation. The information of time constants and trap locations will be discussed in detail in the following section. Figure 4(d) presents the absolute amplitudes of the three detrapping behaviors. The values of amplitudes that were read from figure 4(d) are shown in figure 5, the sum of which was consistent with the total decline in the voltage transients illustrated in figure 4(b). In addition, the amplitudes of three detrapping behaviors all decreased after irradiation, indicating a reduction in the trapping densities of the three traps [18].
Download figure:
Standard image High-resolution image4.4. Identification of the energy level
Detrapping experiments were performed at various temperatures to identify the energy levels of the traps, as shown in figure 6(a). The time constant spectra are presented in figure 6(b) and the three traps were identified based on the Arrhenius plots with the energy levels (Ea) shown in figure 6(c), which respresented the electron trap energy levels difference from the conduction band [2, 30]. In addition, figure 6(d) presents the DASs of the three trapping behaviors at various temperatures.
Download figure:
Standard image High-resolution imageDuring the filling process at off-state, the traps located at AlGaN surface, AlGaN barrier and GaN layer can be filled and the details are presented in [26]. Specifically, DP1 was identified at an energy level of 0.477 eV. The trapping behavior in an AlGaN layer under the gate with similar energy level were observed in [18, 31, 32]. Furthermore, the decrease of the leakage current is related to the reduction of traps in the AlGaN layer under the gate [15, 19]. DP2 was identified at an energy level of 0.385 eV, and a trap at this energy level in the AlGaN/GaN interface can be found in [18, 33]. In addition, trapping in the interface with a similar time constant was also reported in [18]. In particular, the time constant and amplitude of DP2 decreased obviously after irradiation shown in figures 4 and 5, indicating that DP2 may dominate the change of device performance of gamma irradiation. It was consistent with the the increase in IDS and the negative shift of Vth in our study, which can be ascribed to the redistribution of traps at the interface after gamma irradiation [4, 34, 35]. The energy level of DP3 was 0.289 eV and traps in the AlGaN surface with similar energy level were widely reported in [2, 36, 37]. In addition, the amplitude of DP3 is observed to decrease after irradiation (figure 5), which is consistent with the results of previous studies that the leakage current decreased after irradiation because of the surface defect re-structuring [4, 24].
4.5. Effect of gamma irradiation on detrapping behaviors
Figure 6(c) presented the difference in trap energy levels after gamma irradiation in our study. The change in Ea of the specific traps of GaN HEMTs under a certain stress has been previously reported. For example, the uniaxial tensile strains were performed on the GaN HEMTs in [38], and the difference in trap energy level of the specific traps under the stress reached 0.035 eV [38]. Furthermore, the change in Ea of GaN HEMTs after gamma irradiation has been widely investigated to characterize the variation of trapping behaviors [6, 12, 14, 39, 40]. Several studies have reported the increase in Ea after irradiation, which was attributed to the induced additional traps and resulted in the degradation of device performance [6, 14]. However, the electrical properties of the device in our experiment were improved after irradiation and no additional traps were observed with the voltage-transient method. In contrast, the decrease in Ea is also reported in numerous studies, which was concluded that Compton scattering of electrons leads to occupation of the trap levels, contributing to increased values of gm and IDS [6, 12, 39, 40]. The results in figure 6(c) show the similar trend that the fitted Ea decreased after gamma irradiation in our experiment. In addition, the electrical performance of the device was improved after irradiation in our study, which was also consistent with the previous studies [6, 12, 39, 40]. Therefore, we believed that the decrease in Ea in our experiment may be related to the occupation of the trap levels, which resulted in the increased values of gm and IDS as reported in the previous studies.
The results also indicated that the time constants of the three traps decreased after irradiation, as shown in figure 4(c), with the energy levels of the three traps decreased in figure 6(c). To provide further understanding of the mechanisms of the detrapping behaviors after gamma irradiation, we must consider the detrapping process in detail. The escape frequency of the trapped carriers when escaping from the trap sites is given by [41]:
where is the attempt-to-escape frequency, T is the temperature, kB is the Boltzmann constant, and ΔS is the change in entropy required to escape from trap level Ea [38, 41]. The measured time constant should be the time at which the maximum number of charge carriers is released from the trap sites, which is inversely proportional to the escape frequency of the trapped carriers [38]. In our experiments, the time constants of the three traps decreased after irradiation, thus indicating a reduction of the number of the traps.
In addition, the difference in trap energy ΔEa of the device before and after gamma irradiation can be derived from (1) as following:
From the experimental results in figures 4(c) and 6(c), we can obtain the time constants and energy levels of three detrapping behaviors before and after irradiation, designated τ(0), τ(γ), Ea(0) and Ea(γ), respectively. Therefore, the variation of ΔS can be derived from (2), as shown in table 1. It suggests that the negative value of ΔS should increase after irradiation (|ΔS(γ)| > |ΔS(0)|), thus indicating that the system structure tends toward order [41]. In addition, the previous investigations have shown that gamma irradiation in the HEMT structures can result in structural ordering of native defects [28, 42]. Furthermore, the decrease of trap density from structural reordering of native defects may lead to an increase in the drain–source current [6, 16], which is consistent with our results. Therefore, we assumed that the negative ΔS was related to the redistribution of native defects and structural ordering caused by the gamma irradiation.
Table 1. The variations of ΔS derived from the experimental results. is difference in trap energy of the device before and after irradiation. τ(0) and τ(γ) are the time constants of traps before and after irradiation, respectively. With these parameters, the negative value of the change in entropy after irradiation ΔS(γ) is determined to be greater than that before irradiation ΔS(0).
Detrapping behavior | ΔEa | kBTln(τ(γ)/τ(0)) | T(ΔS(γ) − ΔS(0)) |
---|---|---|---|
DP1 | −0.045 | −0.009 | −0.036 |
DP2 | −0.050 | −0.014 | −0.036 |
DP3 | −0.029 | −0.011 | −0.018 |
Based on the discussions above, the possible locations of the traps and the effect of gamma irradiation can be concluded as follows. During the filling process, the high reverse gate voltage depleted the 2DEG under the gate and also also depleted the bulk vertically and laterally [26]. The electrons were injected from the gate under the high reverse gate voltage and get trapped into the surface, AlGaN layer and GaN layer [22, 26, 43]. During the measuring process, the electrons get detrapped and the 2DEG density in the channel was recovered. The detailed mechanisms of the trapping behaviors in the off-state can be found in [26], and the band schematic and possible locations of the detrapping behaviors before irradiation were presented in figures 7(a) and (b), respectively. After gamma irradiation, the system structure tends toward order, which was indicated with the increase in the negative value of ΔS [41] and was consistent with the previous studies [28, 42]. The structure ordering resulted in the redistribution of native defects, leading to the decrease in traps' density after irradiation [4, 13, 24, 34], which was consistent with our results that the amplitudes of the traps decreased after irradiation in figure 5. The detrapping behaviors with less traps after gamma irradiation were presented in figures 7(c) and (d). Furthermore, the structural reordering of native defects may lead to the improvement of device performance [6, 16], which was consistent with our results that the electrical properties of the device were improved after gamma irradiation.
Download figure:
Standard image High-resolution image5. Conclusion
In this paper, we investigated the effects of gamma irradiation on the electrical characteristics and trapping behaviors of GaN-based HEMTs in detail. To determine the variation of electrical properties of the device, the gate–source leakage current was monitored online during the irradiation. In addition, the current–voltage measurements were performed before and after irradiation. The results showed a slight negative threshold voltage shift, along with an increase in IDS and a reduction of gate leakage current, which indicated the improvement of electrical properties. In particular, we identified three traps in the HEMT using the voltage-transient method and the DAS and also investigated the variations in these traps after irradiation. Analysis of the transient variations revealed that the amplitudes of the three traps decreased, indicating a reduction in the trap densities after irradiation. In addition, the time constants and energy levels of the three traps decreased after irradiation, and the negative value of ΔS was determined to increase, thus indicating that the structure of the defects tends toward order. This result was consistent with the observed improvement in the electrical properties. This article provides a good understanding of how gamma irradiation affects the electrical and physical properties of AlGaN/GaN HEMTs. The results also suggested that the improvement of device performance may be achieved with the gamma irradiation in fabrication.
Acknowledgments
This work was supported in part by the Beijing Municipal Education Commission under Grant No. KZ202110005001, in part by the National Natural Science Foundation of China under Grant No. 61974007.
Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).