Breakdown Voltage Enhancement of Al0.1Ga0.9 N Channel HEMT with Recessed Floating Field Plate

In this paper, electrical and microwave characteristics of Al0.1Ga0.9 N channel HEMTs was reported. The device performance were evaluated for conventional gate, field plate gate, and recessed floating field plate with Silicon nitride (SiN)/Hafnium oxide (HfO2) passivation. The recessed floating field plate HEMT with gate length LG = 0.8 μm, gate to drain distance LGD = 1 μm, and HfO2 (SiN) passivation HEMT reports peak drain current density (IDS) of 0.282(0.288) A/mm at VGS = 0 V, three terminal off-state breakdown voltage (VBR) of 677 (617) V, 6.38 Ω.mm of ON-resistance (RON), transconductance (gm,max) of 93(95) mS/mm, and FT/FMAX of 11.4/49 (12/22) GHz. The HfO2 (SiN) passivation device demonstrated the Johnson figure of merit (JFoM)) of 7.71 (7.404) THz.V and FMAX x VBR product of 33.173 (13.574) THz.V. The high JFoM along with high FMAX x VBR indicates the potential of the ultrawide bandgap AlGaN HEMTs for future power switching and high-power microwave applications. The breakdown voltage (VBR) of the floating field plate HEMT is improved 54 % from conventional HEMT and 31 % improvement from gate field plate HEMT.


Introduction
Group III-nitride based wide-bandgap semiconductors are used in high power microwave and switching applications. Owing the unique combination of high electron density, higher mobility, and high critical electic field of conventional GaN channel-based high electron mobility transistors enables high power and high frequency operation for the past two decades [1][2][3][4] and GaN based based power switches, RF amplifiers, and power diodes are commercially available in the market from multiple vendor [5]. Moreover, GaN-HEMTs are widely used in low-noise microwave applications due to its excellent noise performance [6][7][8]. On other hand, AlGaN ternary channel-based HEMTs are also an interesting devices alternate to GaN-channel HEMTs interms of high breakdown voltage due to ultrawide bandgap features of AlGaN material. AlGaN channel based HEMTs had proven its potential for fast-switching and low switching loss applications, particularly in high temperature and high radiation environments . Existence of large critical electrical field (> 3.3 MV/cm) and on par with GaN saturation velocity [30], the AlGaN channel based HEMTs are recognized as the most optimistics candidates future high-power switching as well as microwave applications in harsh environments [31].
After the first successful demonstartion of AlGaN channel based HEMT for high power and microwave application [10], several research groups reported the high performance of AlGaN channel HEMTs. T. Nanjo et.al. investigated the HEMT on AlN buffer and the device shown enrich V BR by suppressing the drain leakage current [11]. Takuma [15]. A high breakdown voltage of 2200 V was attained for 22 μm gate to drain distance device by adopting ohmic/Schottky-hybrid drain c o n t a c t s [ 1 6 ] . M i n g X i a o e t a l . r e p o r t e d G a N / Al 0.35 Ga 0.65 N/AlN/graded channel HEMT and the device shown enhanced I DS and high I on /I off ratio with improved breakdown voltage [17]. The first RF performance on AlGaN channel based HEMT was reported byAlbert G. Baca et.al. An 80 nm gate length HEMT yields I DS of 0.16 A/mm, and 24 mS/mm of g m,max , and having F T / F MAX of 28.4/18.5 GHz [18]. An RF simulation study on AlGaN HEMT channel based HEMTs reported its potential for large signal RF application as well as power switching applications [24]. Zhang et al. proposed AlGaN double channel HEMTs for improving carrier transport and 2DEG density [27], however the large negative threshold voltage of the device may result in off-state power loss.
The bandgap of the Al x Ga 1−x N tailoring by varying the Al composition (0 < x < 1).The high critical breakdown field and low on resistance (R ON ) are key parameters for power switching applications. Lateral Figure of Merit (LFOM)is used to estimate the potential of a material for power switching [32]; The LFOM of a material depends on the sheet charge density (n s ), critical electric field (E c ), and mobility of the channel (μ ch ). Since the critical electric field of AlGaN channel is higher E c ∼E n g ; 2 < n < 2:5 than the GaN channel, Al x Ga 1−x N channel offers significant improvement in V BR even at high temperature over GaN. The JFOM measures the ability of the materials for high power microwave applications; The JFOM of material systems is the product of F T and V BR . The critical electric field (E c ), and electron saturation velocity ( sat ) influences the JFOM of the HEMT. Since the low Al composition Al x Ga 1−x N channel saturation velocity is on par with GaN channel, along with enhanced critical field improves the JFOM of AlGaN channel HEMTs than GaNbased HEMTs. The high V BR is achieved for long channel L G , long L GD HEMTs, along with Al-richAlGaN channel . Whereas, the smaller L G, L GD , and low Al composition of AlGaN channel results in improved cut-off frequency with the suppressed V BR and hence, there is a trade-off between V BR and device speed (cut-off frequency) of HEMTs.
In this work, we proposed the recessed floating field plate, Al 0.1 Ga 0.9 N channel HEMT for improve the V BR of the device with satisfactory RF performance. L G = 0.8 μm, and L GD = 1 μm Al 0.31 Ga 0.69 N/Al 0.1 Ga 0.9 N HEMT on sapphire substrate is investigated using Silvaco ATLAS TCAD numerical simulation for SiN and HfO 2 passivation. The HfO 2 passivation device shown remarkable improvement in breakdown voltage than SiN passivation. The organization of this work as follows; Device structure description for conventional gate HEMT (Device A), Gate field plate HEMT (Device B), and recessed floating field plate HEMTs (Device C) is discussed in Sec. 2. The physics-based simulation models are described in Sec. 3. The DC and microwave characteristic of proposed HEMT with experimental validation is discussed in Sec. 4 with concluding remarks.

Device Structure Description
The Al 0.1 Ga 0.9 N channel geometry of conventional gate HEMT (Device A), Gate field plate HEMT (Device B), and recessed floating field plate HEMTs (Device C) are displayed in Fig. 1

Simulation Models
The proposed HEMTis analyse during several device physics models in TCAD simulation including mobility model, recombination models, carrier transport, and polarization models. Polarization charge at the heterostructure interface is as follows [33]; The P total at the top/bottom heterointerface depends on the spontaneous (P SP ) and piezoelectric polarization (P PE ) of the materials. The temperature dependent mobility model μ 0 T ; N ð Þ describes as following form [30]; The Selberherr's impact ionization model considered for device breakdown simulation [35] and the impact ionization carrier generation rate described as follows: Where, E is the electric field, electrons ionization rate n ¼ A n exp ÀB n ð Þ= E j j and holes ionization rate p ¼ A p exp ÀB p À Á = E j j. The fitting parameters A n , A p , B n , and B p values are taken from [34] for the simulation.

Results and Discussions
The proposed recessed floating gate HEMT (Device C) transfer characteristic is depicted in Fig. 3 for V DS = 10 V and V GS swept from − 6 to 0 V. The conventional and gate field plate HEMTs also exhibited similar response. Al 0.1 Ga 0.9 N HEMT with L G = 0.8 µm has reached the maximum output current (I DS ) of 0.28 A/mm and g m,max of 95 mS/mm. The threshold voltage (V th ) of the HEMT is extracted as -3 V. The V th of the     Fig. 4 for V GS = -3 to 0 V and V DS swept from 0 V to 10 V. Device A and Device B also exhibited similar output characteristics. The extracted ONresistance (R ON ) of the device from the V-I characteristics (V GS = 0 V) is 6.38 Ω.mm. The R on resistance of the device extracted from the output characteristics at V GS = 0 V By taking the slope (1/Ron = I DS / V DS ) corresponding to ¼ th of maximum drain current (I DS, max ). The device is perfectly pinched-off at V GS = -3 V. The proposed device R ON resistance is comparatively lower than the reported works [10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28]. The transfer and output characteristics of simulation result of conventional Al 0.1 Ga 0.9 N channel HEMT is validated with experiment result of [27] and it is shown in Figs. 5 and 6 respectively. The simulated results were well correlated with the reported experimental work.
The HEMT breakdown simulation was carried out at offstate condition (V GS = -8 V). The electric field (E-filed) distributions for SiN and HfO 2 passivation of (a) Device A, (b) Device B and (c) Device C are depicted in Figs. 7 and 8 respectively. The permittivity (Ɛ r ) and thickness (t) of the insulator (passivation) modulates the E-field. High permittivity HfO 2 (Ɛ r =25) passivation smoothening the E-filed at gate and drain edge [35][36][37][38].
From Fig. 7(a) (SiN passivation) and Fig. 8(a) (HfO 2 passivation), it is observed that a peak E-field exist at the gate edge, which lower the V B R of the device. Therefore, sinking the field distribution becomes the alternate solution to enhance the V BR . In general, the gate field plate (FPs), source field plate (SFPs), and drain connected field plate techniques are used to suppress the electric field [39]. The gate field plate techniques demonstarted the improved breakdown voltage by alleviating high E-field as shown in Fig. 7(b) (SiN passivation) and Fig. 8(b) (HfO 2 passivation) and it reshaping the field distribution.
In this work, a recessed floating field plate structure is considered for further, to enhance the breakdown voltage of the HEMTs. The electic field distribution of proposed HEMTs are shown in Fig. 7(c) (SiN passivation) and Fig. 8(c) (HfO 2  passivation).
The introduction of floating field plate, reshaped the electric field and suppressed the E-field near the gate edge effectively and a peak electric field found at the drain side edge of the floating field plate because of much closer with drain electrode and also the recessed field plate is very closer with 2DEG channel than gate field plate results in better surface field distribution, which enhances the breakdown voltage of the HEMTs. The E-field engineering soley depends on field plate length, passivation permittivity, thickness of the passivation, and recess depth.
For the proposed Device C dimensions, the off-state breakdown voltage characteristics of HEMTs are plotted in Fig. 9. The breakdown voltage of the HEMT extracted from I D -V D curves at the intersection of the extrapolated saturation segment (an V D saturated and a sudden increase in I D ). The recessed floating field plate HEMT shown remarkable V BR of 677 (617) V for HfO 2 (SiN) passivation. However, the conventional HEMT shown V BR of 307(297) V and gate filed plate shown V BR of 462 (428) V. From this analysis of off state breakdown voltage simulation, the floating field plate demonstared 54 % improvement in V BR than conventional HEMT, and 31 % higher than the gate field plate HEMT.
The Figs. 10 and 11 shows the simulation results of electron concentration distribution of HEMTs at breakdown condition. The larger depletion region of floating field plate shown in Figs. 10(c) and 11(c), suppressed E-field near the drain side of gate edge, and high critical field of Al 0.1 Ga 0.9 N channel are the major reason for improving the breakdown voltage of the proposed HEMTs.
The microwave performance of the SiN and HfO 2 passivated Device C are depicted in Figs. 12 and 13 respectively. The F T /F MAX of the device extracted from current gain and power gain when it reaches 0 dB respectively. The conventional HEMTs are showing better F T than field plated HEMTs. Due to the introduction of additional field plate in the device structure,increases the parasitic capacitance (C GS and C GD ) which limits the speed of the HEMTs as given in the Eqs. (9); The I DS and V BR of a transistor expected to be very high for delivering high RF output power density as shown in Eq. (8). The proposed recessed floating field plate HEMT shown a F T /F MAX of 11.4/49 GHz for HfO2 passivation and 12/22 GHz for SiN passivation. The improved power performance along with the satisfactory cut-off frequencyof the proposed HEMTs demonstrated its capability for future RF and power switching applications. Further the small signal RF charcateristics can be enhanced by scale down the device dimensions.  The comparison of state of art AlGaN channel based HEMTs parameter and performance along with proposed Device C are tabulated in Table 1. The proposed Device C shows improved drain current, breakdown volatge, and RF perfomance among the reported work for L GD (1 μm) and L G (0.8 μm).

Conclusions
Breakdown voltage performance analysis of Al 0.1 Ga 0.9 N channel HEMT is reported with floating field plate and AlGaN buffer. A Physics-based Technology Computer Aided Design(TCAD) simulation results shows that the peak electric field near the drain-side of gate edge is majorly reduced by recessed floating filed plate technique, which further elevated the breakdown performance of the AlGaN channel HEMTs. The insertion of floating filed plate reshaped the Efield distribution in the access region effectively and reduced the E-field near the drain side of gate edge and thus enhanced the breakdown voltage. The proposed recessed floating field plate HEMT yields V BR of 677 Vand F T /F max 11.4/49 GHz for a high-k HfO 2 passivation. The breakdown voltage (V BR ) of the HfO 2 passivation floating field plate HEMT is improved 54 % from conventional HEMT and 31 % improvement from gate field plate HEMT. The excellent breakdown power performances along with RF characteristics of the ultrawide bandgap AlGaN channel based HEMTs are attractive alternate devices for next geneartion power electronics and RF applications.
Data Availability Not Applicable.
Code Availability Not Applicable.

Declarations
Manuscripts is prepared strictly adhering to the norm of the journal, contains 5 Sections and no sub sections.
Ethics Approval Not Applicable.

Consent to Participate Not Applicable.
Conflicts of Interest/Competing Interests The authors declare that there is no conflict of interest reported in this paper.
Consent for Publication Not applicable as the manuscript does not contain any data from individual.

Disclosure of Potential Conflicts of Interest Not Applicable.
Research Involving Human Participants and/or Animals Not Applicable.
Informed Consent Not Applicable.