Proton irradiation influence on gate-channel low-field carrier mobility of AlGaN/GaN HEMTs

AlGaN/GaN high electron mobility transistors (HEMTs) with different device sizes were prepared and exposed to 0.4 MeV proton irradiation. The low-field carrier transport characteristics of the gate channel are obtained from the capacitance-voltage curves and current-voltage curves. For the device with a longer gate-drain distance (30 μm), after 0.4 MeV proton irradiation, the gate-channel low-field carrier mobility increases by 14.3% on average. For the device with a shorter gate-drain distance (15 μm), the carrier mobility decreases by 13.4% on average after proton irradiation. This phenomenon is studied with regard to the polarization scattering effect. It is found that the polarization distribution in the AlGaN/GaN HEMTs changes after proton irradiation and different gate-drain distances correspond to different polarization distributions.


Introduction
The AlGaN/GaN high electron mobility transistor (HEMT) possesses a lot of advantages, such as a high breakdown voltage, high carrier mobility and high carrier saturation velocity [1][2][3]. Excellent performance makes the AlGaN/GaN * Author to whom any correspondence should be addressed.
Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
HEMT especially suitable for applications in aerospace technology [1][2][3][4][5][6]. Since outer space is filled with protons, the proton irradiation effect is a main focus in aerospace applications. Although the excellent irradiation resistance of GaN has been proven in theory, the growth process of epitaxial layers is accompanied by lattice mismatch and thermal mismatch between epitaxial layers, which make for a high density of defects. Moreover, the device preparation process, such as thermal annealing, will induce serious lattice damage [7,8]. This makes the irradiation resistance of AlGaN/GaN HEMT greatly challenging. Although many studies have focussed on the proton irradiation effect of the AlGaN/GaN HEMT, the micro-mechanism is still not thoroughly clarified [9][10][11][12][13][14]. Therefore, the radiation resistance of the AlGaN/GaN HEMT needs to be studied and clarified more systematically in order to take more reliable reinforcement measures.
Proton irradiation may induce displacement damage. Previous studies mainly focus on the defects, which are induced by displacement damage and are able to capture and scatter carriers [9][10][11][12][13][14]. This explanation is effective for Sibased semiconductors as well GaN-based semiconductors. However, what is different for AlGaN/GaN HEMTs, is that the displacement damage will inevitably change the strain distribution at the AlGaN/GaN interface, which may in turn affect the polarization charge distribution. The polarization charge distribution is closely related to carrier mobility in the view of polarization Coulomb field (PCF) scattering, which has been proven as an important and specific scattering mechanism in AlGaN/GaN HEMTs [15][16][17][18][19]. PCF scattering originates from the nonuniform polarization charges at the AlGaN/GaN interface and is directly related to the device size. The proton irradiation affecting the PCF scattering of AlGaN/GaN HEMTs with different sizes has not been considered in previous studies [9][10][11][12][13][14]. Therefore, a study of the proton influence on the carrier transport characteristics of AlGaN/GaN HEMT with regard to PCF scattering is significant toward the proposal of irradiation reinforcement measures.
In this paper, AlGaN/GaN HEMTs with four different gate-drain distances were prepared, and the capacitancevoltage (C-V) curves and current-voltage (I-V) curves were measured. The main scattering mechanisms are extracted separately, and the proton irradiation influence on the gate-channel low-field carrier transport characteristics is analyzed.

Experiment
The AlGaN/GaN HEMTs used in this paper were fabricated based on AlGaN/GaN heterostructure material. The AlGaN/GaN heterostructure material was fabricated by metalorganic chemical vapor deposition on a Si substrate. The epitaxial layers consist of a 3.9 µm GaN buffer layer, 0.26 µm i-GaN layer, 27.5 nm Al 0.25 Ga 0.75 N barrier layer, and a 3.3 nm GaN cap layer. The material structure is shown in figure 1.
AlGaN/GaN HEMTs were fabricated by the traditional process. As shown in figure 2, the source and drain are both Ohmic contacts and the gate is a Schottky contact. Ti/Al/Ni/Au as the Ohmic contacts were deposited on the material by ebeam evaporation and lift-off, which then were annealed in an ambient N 2 environment at 800 • C for 30 s in a rapid thermal annealing system. Using Ni/Au, the Schottky contacts were created by e-beam evaporation. The gate length L G is 4 µm. The gate width W G is 104 µm. The gate-source distance L GS is 4 µm. The gate-drain distances L GD are 30 µm, 25 µm, 20 µm and 15 µm, which are marked as Device 1, Device 2, Device 3 and Device 4, respectively. The proton irradiation experiment was performed at the GIC4117 accelerator with 0.4 MeV protons at a fluence of 2.16 × 10 12 cm −2 (beam current 7 pA)   and at room temperature. Protons irradiated the upper surface of the devices directly. During the whole irradiation process, the temperature was measured and the maximum temperature rise was only 0.6 K. The C-V curves (at 1 MHz) and I-V curves of the devices were measured at room temperature using a Keithley 4200 A semiconductor parameter analyzer, before and after proton irradiation. The Hall mobility of the AlGaN/GaN heterostructure material was measured at 300 K and 77 K from the Hall measurements before and after proton irradiation, and is shown in table 1. The Hall mobility both decreased at 300 K and 77 K after 0.4 MeV proton irradiation. Figure 3 shows the C-V curves of the fabricated devices before and after proton irradiation. Based on these curves, the electron density (n 2D ) of the two-dimensional electron gas (2DEG)  in the gate channel at different gate voltages can be obtained by following [20]:

Results and discussion
where C is the measured capacitance value, V T is the threshold voltage, which can be obtained by differentiating the C-V curve, V GS is the gate-source voltage, M is the gate Schottky contact area, and e is the electron charge. The obtained n 2D before and after proton irradiation is shown in figure 4. Compared to n 2D before proton irradiation, n 2D is lower after proton irradiation. Figure 5 shows the I-V curves before and after proton irradiation of Devices 1-4. As L GD decreases, the drain-source current I DS after irradiation tends to first increase and then decrease compared to before irradiation. According to the I-V curves and the C-V curves, the gate-channel 2DEG electron mobility (µ n ) can be calculated by the following equations [20,21]: where R d is the gate-drain access resistance, R s is the gatesource access resistance, R c = 3.54 Ω•mm is the Ohmic contact resistance, which is obtained by the transmission line method. n 2D0 and µ n0 are the 2DEG electron density and electron mobility of the gate channel at V GS = 0 V. In this study, the drain-source voltage V DS = 0.1 V is chosen because the electron density in the channel at low drain-source fields can be approximated as not varying with the channel position [21][22][23]. Moreover, according to a previous study [24], the gradient of the polarization charge density at high field is larger than that at low field, inducing the effect where PCF scattering becomes more significant with V DS increasing. Thus, the analysis of low-field mobility offers guidelines for high-field situations.
The obtained gate-channel 2DEG electron mobility is shown in figure 6. The mobility curves of the four devices before irradiation are almost the same. However, the mobility curves after irradiation are obviously separated and decrease as L GD decreases. For Device 1, the 2DEG electron mobility is increased by 14.3% on average after irradiation. For Device 4, the 2DEG electron mobility decreases by 13.4% on average after irradiation. This phenomenon is analyzed as follows.
Mobility is determined by the scattering mechanisms. According to previous studies, the main scattering mechanisms in AlGaN/GaN HEMT are polar optical phonon (POP) scattering, acoustic phonon (AP) scattering, interface roughness (IFR) scattering, dislocation (DIS) scattering and PCF scattering [23,25,26]. POP scattering, AP scattering, IFR scattering and DIS scattering are all determined by the inherent properties of the AlGaN/GaN heterostructure material and are not related to device size. Under the influence of these scattering mechanisms, the mobility variation of all four devices should be consistent after irradiation, which is obviously inconsistent with the result in figure 6. PCF scattering has been proven to be closely related to device size. Therefore, the behavior of the electron mobility (figure 6) has to be attributed to PCF scattering.
In order to provide a clearer quantitative analysis, the mobility corresponding to different scattering mechanisms  was calculated following the previous studies [22,26]. Figures 7 and 8 show that the electron mobility before and after proton irradiation corresponds to all the scatterings (µ PCF , µ POP , µ AP , µ DIS and µ IFR ) and the total electron mobility (µ TOTAL ). Figure 9 compares the µ PCF of the fabricated four AlGaN/GaN HEMTs before and after proton irradiation. Being similar to figure 6, it is found from figure 9 that the µ PCF curves of the four devices largely overlap before proton irradiation. After irradiation, the µ PCF curves are separated obviously and decrease as L GD decreases.
For a clear comparison, the µ PCF before and after proton irradiation of each device is shown by figure 10. For Device 1, µ PCF gets larger after irradiation. For Device 4, µ PCF gets smaller after irradiation.  Based on the results above, it is quantitatively demonstrated that the different mobility behaviors result from PCF scattering. PCF scattering originates from the nonuniform distribution of polarization charges at the AlGaN/GaN interface. The difference between the nonuniform polarization charges and the original (heterostructure material before device fabrication) uniform polarization charges is called the additional polarization charges ∆σ. In this study, the additional polarization charge influence near the Ohmic contact is negligible, which is proved by figure 9(a) [23]. The additional polarization charges at the AlGaN/GaN interface consist of two parts, as shown in figure 11. The applied gate bias changes the strain via the converse piezoelectric effect, generating additional polarization charges ∆σ 1 in region I. Since the lattices are interconnected, and the stress (induced by gate bias) in the region I will apply an opposite stress in region II through the interconnected lattice, generating ∆σ 2 [1,22]. In addition, the additional polarization charge distribution in region II is expected to be linear, since the degree of lattice variation decreases with increasing distance. The PCF scattering potential V(x,y,z) generated by the additional polarization charges is expressed as follows [23]: where ε s is the GaN dielectric constant. It is obvious that PCF scattering is closely related to the device size and additional polarization charges. According to figure 9(a) and equation (5), before irradiation, the influence of ∆σ 2 is negligible, and V(x,y,z) is dominated by ∆σ 1 . During proton irradiation, region II lacks the coverage of the gate metal. The protons entering region II enhance the tensile strain of the barrier layer and ∆σ 2 turns into ∆σ 2 + ∆σ 2 ′ . Since the protons uniformly irradiate the top of the devices, ∆σ 2 ′ is consistent in each device. Due to the lattice continuity, the enhanced tensile strain in region II induces compressive strain or weakens the original tensile strain in region I, with ∆σ 1 turning into ∆σ 1 + ∆σ 1 ′ . According to equation (5), there is an offsetting effect between positive ∆σ 2 + ∆σ 2 ′ and negative ∆σ 1 + ∆σ 1 ′ . A larger L GD will correspond to a stronger offsetting effect. For Device 1, L GD is larger, and the influence of ∆σ 2 + ∆σ 2 ′ is stronger. Following equation (5), the PCF scattering potential gets smaller, and µ PCF becomes larger after irradiation. For Device 4, L GD is smaller, so the integration interval of the last term in equation (5) is smaller and the influence of ∆σ 2 + ∆σ 2 ′ is weaker. As a result, ∆σ 1 + ∆σ 1 ′ plays a dominant role, and the PCF scattering potential gets stronger, inducing a smaller µ PCF after irradiation. This is the reason for the mobility behavior before and after proton irradiation, as shown by figures 6 and 9.
To further verify the above analysis, the n 2D variation level ∆n 2D after proton irradiation at V GS = 0 V is compared, as shown by figure 12. Since the 2DEG originates from the strain and polarization charges, the change of n 2D can indirectly reflect the strain of the lattice under the gate [1]. It can be found from figure 12 that ∆n 2D is approximately linear variation as a function of L GD , and ∆n 2D decreases with L GD increasing. This indicates that, in region I, the compressive strain increases or tensile strain weakens. The larger L GD is, the stronger the effect is. This phenomenon exactly confirms the conclusion of this study.

Conclusion
The proton irradiation influence on gate-channel low-field carrier mobility of AlGaN/GaN HEMTs is analyzed with regard to PCF scattering, which has not been examined in previous studies. It is found that proton irradiation affects the intensity of the PCF scattering by changing the strain and polarization distribution, which in turn influences the carrier mobility of devices. For the device with a 30 µm gate-drain distance, carrier mobility increases due to the decreased intensity of PCF scattering after 0.4 MeV proton irradiation. The carrier mobility of the device with a 15 µm gate-drain distance decreases due to the increasing in PCF scattering intensity after 0.4 MeV proton irradiation. We analyse the variation of PCF scattering with regard to lattice strain, and the lattice strain corresponds to a microcosmic understanding of the influence of proton irradiation on AlGaN/GaN HEMTs. Moreover, the results of the experiment show that the I DS of the device with a larger gate-drain distance gets larger after 0.4 MeV proton irradiation. This means that the performance of AlGaN/GaN HEMTs can be improved by proton irradiation with appropriate energy and fluence. Different gate-drain distances correspond to different trends of current variation after 0.4 MeV proton irradiation, suggesting that adjusting gate-drain distance properly can also provide a new dimension for enhancing the irradiation resistance of devices, which will of benefit for the further optimization of applications in the aerospace field.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).