The Role of Carbon Doping on Breakdown, Current Collapse and Dynamic On-Resistance Recovery in AlGaN/GaN High Electron Mobility Transistors on Semi‐Insulating SiC Substrates

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as

doi: 10.1002/pssa.201900762.

ON

AlGaN/GaN High Electron Mobility Transistors on Semi-Insulating SiC Substrates

Nicolò Zagni*, Alessandro Chini, Francesco Maria Puglisi, Paolo Pavan, Giovanni Verzellesi

N. Zagni, Prof. A. Chini, Prof. F. M. Puglisi, Prof. P. Pavan

Department of Engineering “Enzo Ferrari”, University of Modena and Reggio Emilia, via P. Vivarelli 10, 41125, Modena, Italy

E-mail: nicolo.zagni@unimore.it Prof. G. Verzellesi

Department of Sciences and Methods for Engineering (DISMI) and EN&TECH Center, University of Modena and Reggio Emilia, via G. Amendola, 2, 42122, Reggio Emilia, Italy

Keywords: GaN HEMTs, C-doping, impact ionization, breakdown, current collapse, dynamic 𝑅�� recovery

Abstract: In this work, the critical role of carbon doping in the electrical behavior of AlGaN/GaN High Electron Mobility Transistors (HEMTs) on semi-insulating SiC substrates is assessed by investigating the off-state three terminal breakdown, current collapse and dynamic on-resistance recovery at high drain-source voltages. Extensive device simulations of typical GaN HEMT structures are carried out and compared to experimental data from published, state-of-the-art technologies to: i) explain the slope of the breakdown voltage as a function of the gate-to-drain spacing lower than GaN critical electric field as a result of the non-uniform electrical field distribution in the gate-drain access region; ii) attribute the drain current collapse to trapping in deep acceptor states in the buffer associated with carbon doping; iii) interpret the partial dynamic on-resistance recovery after off-state stress at high drain-source voltages as a consequence of hole generation and trapping.

1. Introduction

AlGaN/GaN High Electron Mobility Transistors (HEMTs) are widely investigated and developed for power switching and power RF applications thanks to their potentially superior

Accepted Article

nicolo.zagni@unimore.it

Accepted Article

nicolo.zagni@unimore.it Prof. G. Verzellesi

Accepted Article

Prof. G. Verzellesi Department of

Accepted Article

Department of Sciences and Methods for

Accepted Article

Sciences and Methods for of Modena and Reggio Emilia, via

Accepted Article

of Modena and Reggio Emilia, via GaN HEMTs, C

Accepted Article

GaN HEMTs, C

In this work, the critical role of carbon doping

Accepted Article

In this work, the critical role of carbon doping High Electron Mobility Transistors (HEMTs)

Accepted Article

High Electron Mobility Transistors (HEMTs) investigating the off

Accepted Article

investigating the

off-Accepted Article

-investigating the off-investigating the off

Accepted Article

investigating the

off-investigating the off state three terminal breakdown

Accepted Article

state three terminal breakdown at high drain

Accepted Article

at high

drain-Accepted Article

-source voltages

Accepted Article

source voltages

Accepted Article

are carried out

Accepted Article

are carried out to:

Accepted Article

to: i)

Accepted Article

i) explain

Accepted Article

explain

lower than GaN critical electric field

Accepted Article

lower than GaN critical electric field distribution in the gate

Accepted Article

distribution in the

gate-Accepted Article

-drain access region

Accepted Article

drain access region

Accepted Article

deep acceptor states in the buffer associated with carbon doping;

Accepted Article

deep acceptor states in the buffer associated with carbon doping; resistance recovery after

Accepted Article

resistance recovery after generation and trapping

Accepted Article

generation and trapping.

Accepted Article

.

1. Introduction

Accepted Article

1. Introduction

AlGaN/GaN High Electron Mobility Transistors (HEMTs) are widely

Accepted Article

AlGaN/GaN High Electron Mobility Transistors (HEMTs) are widely

performance compared to Si and SiC based devices. For example, in terms of the Baliga Figure Of Merit (B-FOM) – that relates the breakdown voltage, 𝑉��, to the specific on-resistance 𝑅��× 𝐴 via material intrinsic properties – GaN has an outstanding value of 870 compared to 340 of SiC, with Si taken as reference (i.e., equal to 1) [1]. The reduction of the off-state punch-through currents to increase the high-voltage capability of AlGaN/GaN HEMTs is often achieved by introducing acceptor trap states in the buffer layer(s) underlying the device channel through carbon (C) dopant [2].

In this work, we discuss the critical role of C-doping to achieve lower drain-source leakage current and higher breakdown voltage as well as its impact on the dynamic on-resistance, 𝑅��. The analysis is carried out by means of numerical device simulations. Experimental data from state-of-art technologies reported in the open literature are adopted in order to calibrate device simulations as well as to show their suitability to explain actual device behavior. More specifically three experimentally observed phenomena are analyzed: i) 𝑉�� dependence on gate-to-drain length, 𝐿��, with a slope lower than GaN critical field; ii) current-collapse in pulsed measurements of output characteristics; iii) dynamic 𝑅�� recovery with increasing off-state stress bias. The rest of the paper is organized as follows. Section 2 describes the simulated device cross-section and the models employed to reproduce experimental data. Section 3 shows the results of the calibration process and the comparison between simulations and experimental data on 𝑉�� vs. 𝐿�� and 𝐼�− 𝑉�� current collapse. In Section 4, we exploit the device model to interpret the results on the dynamic 𝑅�� recovery found in several device technologies reported in the literature [3]–[6]. Conclusions are drawn in Section 5.

2. Simulated Device Structure and Models

The simulated device cross-section is shown schematically in Figure 1. Device dimensions resemble the ones of the fabricated device in [2]. Gate-source spacing (𝐿��), gate length (𝐿�) and gate field plate overhang (𝐿��) are 1, 0.7, 0.6 μm, respectively. The substrate is semi-insulating SiC,

Accepted Article

the high

Accepted Article

the

high-Accepted Article

-voltage capability of

Accepted Article

voltage capability of acceptor trap states in the buffer

Accepted Article

acceptor trap states in the buffer

In this

Accepted Article

In this work

Accepted Article

work, we discuss

Accepted Article

, we discuss current and higher breakdown voltage

Accepted Article

current and higher breakdown voltage analysis is carried out by means of

Accepted Article

analysis is carried out by means of art technologies

Accepted Article

art technologies reported in the open literature are adopted in order to calibrate device simulations

Accepted Article

reported in the open literature are adopted in order to calibrate device simulations as well as to show their suitability to explain actual device behavior. More specifically

Accepted Article

as well as to show their suitability to explain actual device behavior. More specifically experimentally observed phenomena

Accepted Article

experimentally observed phenomena lower than GaN critical field

Accepted Article

lower than GaN critical field characteristic

Accepted Article

characteristics

Accepted Article

s

Accepted Article

;

Accepted Article

; iii)

Accepted Article

iii) dynamic

Accepted Article

dynamic is organized as follows.

Accepted Article

is organized as follows.

employed to reproduce experimental data

Accepted Article

employed to reproduce experimental data

the comparison between simulations and experimental data on

Accepted Article

the comparison between simulations and experimental data on collapse. In Section 4

Accepted Article

collapse. In Section 4, we exploit the

Accepted Article

, we exploit the found in

Accepted Article

found in several device

Accepted Article

several device Section 5

Accepted Article

left floating during both measurements and simulations (except for the dynamic 𝑅�� recovery analysis, as explained in Section 4). In general, the epitaxy of GaN HEMTs is more complex than that of [2], especially in technologies for power switching applications, in which the substrate is typically p-type Si and complex strain-relief layers are adopted. The conducting Si substrate is typically grounded, resulting in high vertical fields that influence charge storage and potential distribution in the device. This work focuses on GaN HEMTs on semi-insulating SiC which is the typical choice for RF applications. In particular, the simulated structure does not include strain relief layers, it includes a 100 nm thick AlN nucleation layer and a 100 um thick undoped SiC substrate. The main feature of the device is the C-doped GaN back-barrier underneath the channel that renders the buffer semi-insulating thereby increasing the maximum breakdown voltage and improving the B-FOM, as discussed more in detail in [2]. The simulation software adopted for the device analysis is the commercial suite SDeviceTM _{(Synopsys Inc.). Charge transport was accounted} for by means of the drift-diffusion model. Deep-level traps were accounted for by including, for each distinct trap level, one trap-balance equation, describing, within the framework of the Shockley-Read-Hall theory, the dynamics of trap occupation without any quasi-static approximation. Piezoelectric polarization was included by using the default strain model of SDeviceTM_{. Both the gate and source/drain contacts were modeled as Schottky contacts with proper} energy barriers. Electron tunneling was activated at the contacts to reproduce the gate leakage current and to mimic ohmic behavior at the source and drain contacts. Impact-ionization coefficients for both electrons and holes were set in agreement with recent Monte-Carlo calculations [7]. C-doping in the GaN back-barrier was modeled by considering a dominant deep acceptor level (at 0.9 eV from the valence-band edge, 𝐸�), partially compensated by a shallow donor level (at 0.11 eV from the conduction-band edge, 𝐸�). Specifically, the adopted C-related acceptor and donor concentrations were 8×1017 _cm-3 _{and 4×10}17 _cm-3_{, respectively. For more} accurate determination of the effective trap concentration, one should rely on C-V measurements

Accepted Article

typically grounded,

Accepted Article

typically grounded, result

Accepted Article

result distribution in the device.

Accepted Article

distribution in the device. typical choice

Accepted Article

typical choice for RF applications.

Accepted Article

for RF applications.

rs, it includes a 100 nm thick AlN nucleation layer and a 100 um thick undoped SiC

Accepted Article

rs, it includes a 100 nm thick AlN nucleation layer and a 100 um thick undoped SiC The main feature of the device is the C

Accepted Article

The main feature of the device is the C that renders the buffer semi

Accepted Article

that renders the buffer semi improving the B

Accepted Article

improving the

B-Accepted Article

-FOM, as discussed more in detail in

Accepted Article

FOM, as discussed more in detail in device analysis is the commercial suite SDevice

Accepted Article

device analysis is the commercial suite SDevice for by means of

Accepted Article

for by means of the drift

Accepted Article

the drift

each distinct trap level, one trap

Accepted Article

each distinct trap level, one trap Read

Accepted Article

Read-Accepted Article

-Hall theory, the dynamics of trap occupation without any quasi

Accepted Article

Hall theory, the dynamics of trap occupation without any quasi approximation.

Accepted Article

approximation. Piezoelectric polarization was included by using

Accepted Article

Piezoelectric polarization was included by using . Both

Accepted Article

. Both the

Accepted Article

the gate and source/drain contacts were modeled as Schottky contacts with proper

Accepted Article

gate and source/drain contacts were modeled as Schottky contacts with proper energy barriers

Accepted Article

energy barriers. Electron tunneling was activated at the contacts to reproduce

Accepted Article

. Electron tunneling was activated at the contacts to reproduce and to mimic ohmic

Accepted Article

and to mimic ohmic

coefficients for both electrons and holes were set in agreement

Accepted Article

coefficients for both electrons and holes were set in agreement [7]

Accepted Article

[8] . In this work, we calibrated the carbon-related trap concentration in order to match results in terms of DC characteristics and off-state breakdown, as C-V measurements were absent for the fabricated devices [2]. The choice of the C-doping model adopted in this work is justified by the fact that by adjusting the donor-acceptor auto-compensation ratio it is possible to accurately reproduce dynamic effects in different AlGaN/GaN power HEMTs [9]–[13]. Moreover, a higher donor concentration in C-doped GaN compared to the donor density measured in unintentionally doped samples was experimentally confirmed in [14], and attributed to the auto-compensation between C related donors and acceptors, as assumed here. Finally, a similar C-doping model with high donor-acceptor auto-compensation was found also by other authors to be instrumental to reproduce breakdown effects in C-doped AlGaN/GaN HEMTs by means of device simulations [15]. 3. Breakdown and Current Collapse

The calibration of the simulation set-up against experimental results from [2] is illustrated by Figure 2. Figure 2a) shows the comparison between simulated and measured transfer characteristics of the GaN HEMTs under study. Figure 2b) shows the comparison between measurements and simulations of breakdown voltage for devices with different 𝐿�� . The breakdown condition is defined as the 𝑉�� bias necessary to reach 1 mA/mm of off-state 𝐼� [2]. Detailed comparison between measured and simulated off-state 𝐼�− 𝑉�� curves were reported elsewhere [16]. The achieved agreement shows that the model employed for the C-doping in the back-barrier is indeed effective in the prediction of the lateral breakdown limits of the fabricated devices. As shown in Figure 2b), the slope of the 𝑉�� vs 𝐿�� curve is lower than the critical field of GaN (of about 3.9 MV/cm [17]) despite breakdown being induced, according to our simulations, by avalanche generation. This is attributed to the highly non-uniform electric field distribution in the depletion region between gate and drain [16], allowing the critical field to be reached in a localized spot at the drain end of the gate-drain access region even if the average field is still much smaller than the critical field.

Accepted Article

dynamic effects in different

Accepted Article

dynamic effects in different donor concentration in

Accepted Article

donor concentration in C

Accepted Article

C doped samples

Accepted Article

doped samples was

Accepted Article

was experimentally

Accepted Article

experimentally related donors and acceptors

Accepted Article

related donors and acceptors acceptor

Accepted Article

acceptor auto

Accepted Article

auto breakdown effects in C

Accepted Article

breakdown effects in C Breakdown and

Accepted Article

Breakdown and Current Collapse

Accepted Article

Current Collapse The calibration of the simulation set

Accepted Article

The calibration of the simulation set Figure 2

Accepted Article

Figure 2.

Accepted Article

. Figure 2

Accepted Article

Figure 2 characteristics

Accepted Article

characteristics of the GaN HEMTs under study

Accepted Article

of the GaN HEMTs under study

measurements and simulations of breakdown voltage for devices with different

Accepted Article

measurements and simulations of breakdown voltage for devices with different breakdown condition is def

Accepted Article

breakdown condition is def

comparison between measured and simulated

Accepted Article

comparison between measured and simulated 6]

Accepted Article

6].

Accepted Article

. The

Accepted Article

The achieved

Accepted Article

achieved

barrier is indeed effective in the prediction of the lateral breakdown limits of the fabricated

Accepted Article

barrier is indeed effective in the prediction of the lateral breakdown limits of the fabricated As shown in Figure 2b), t

Accepted Article

As shown in Figure 2b), t GaN (of about

Accepted Article

One shortcoming of introducing traps in the GaN buffer is the DC-to-dynamic drain current dispersion, also referred to as current collapse, that limits the maximum achievable output power. This effect is commonly estimated through pulsed 𝐼�− 𝑉�� measurements with different baseline conditions to either induce or suppress trapping. This is shown in Figure 3, where two different baseline biases are applied during 𝐼�− 𝑉�� simulations and measurements (taken at 𝑉�� of 1 V and 2 V), namely �𝑉��,���, 𝑉��,���� = (0, 0) V and �𝑉��,���, 𝑉��,���� = (−0.8, 30) V . Measurements are still taken from Ref. 2. The first baseline for gate and drain is benign with respect to charge trapping. On the other hand, the second baseline biases the device in semi-on conditions with a high drain voltage (corresponding to class-AB operating point [2]), causing the acceptor states in the buffer to accumulate negative charges (by emitting holes) and giving rise to current dispersion as the device is pulsed to on-state conditions. The model adopted for the C-doping in the simulations, see Figure 3a), is able to reproduce the experimental data, shown in Figure 3b), and also the drain current dispersion due to the acceptor traps associated with C-doping.

4. Dynamic 𝑹𝑶𝑵 recovery

In this section, we exploit the device model employed in Section 3 to give an interpretation of the partial recovery of the dynamic 𝑅��. This behavior was observed in several different state-of-art power GaN HEMT technologies on Silicon substrates for 𝑉�� stress voltage beyond some critical value in the 100-300 V range [3]–[6]. A similar behavior is exhibited also by devices simulated in this work. The phenomenon is illustrated by the simulation results shown in Figure 4a). The simulations were carried out on a structure similar to the one shown in Figure 1, with grounded substrate and with a gate-to-drain spacing 𝐿�� of 10 m. All other parameters are the same as for simulations allowing to reproduce DC, breakdown, and current collapse curves described in Sections 2 and 3. All simulations were carried out with the same off-state gate bias 𝑉��,��� = −5 V, while the drain stress bias was increased from 0 to 600 V, activating/deactivating

Accepted Article

baseline biases are applied during

Accepted Article

baseline biases are applied during , namely

Accepted Article

, namely �

Accepted Article

�𝑉

Accepted Article

𝑉��

Accepted Article

�� 𝑉�� 𝑉

Accepted Article

𝑉�� 𝑉 ,

Accepted Article

,��

Accepted Article

���

Accepted Article

�,

Accepted Article

, 𝑉

Accepted Article

𝑉 are still taken from

Accepted Article

are still taken from Ref.

Accepted Article

Ref.

On the other hand, the second

Accepted Article

On the other hand, the second drain voltage

Accepted Article

drain voltage (corresponding to class

Accepted Article

(corresponding to class

accumulate negative charges (by emitting holes)

Accepted Article

accumulate negative charges (by emitting holes)

Accepted Article

pulsed to on

Accepted Article

pulsed to on

a), is able to reproduce the experimental data, s

Accepted Article

a), is able to reproduce the experimental data, s current dispersion due to the

Accepted Article

current dispersion due to the

𝑹

Accepted Article

𝑹𝑶𝑵

Accepted Article

𝑶𝑵 r

Accepted Article

recovery

Accepted Article

ecovery

In this section, we exploit the device model employed in Section 3 to give an interpretation

Accepted Article

In this section, we exploit the device model employed in Section 3 to give an interpretation

Accepted Article

partial recovery of

Accepted Article

partial recovery of

power GaN HEMT technologies

Accepted Article

power GaN HEMT technologies critical value in the 100

Accepted Article

critical value in the 100 simulated in this work

Accepted Article

simulated in this work.

Accepted Article

. The phenomenon

Accepted Article

The phenomenon

Accepted Article

. The simulations were carried out on a structure

Accepted Article

. The simulations were carried out on a structure

impact ionization. The results shown in Figure 4a) indicate that the 𝑅�� recovery at high 𝑉��,��� can be reproduced only when impact ionization is included (red squares). In fact, if impact ionization is deactivated (yellow triangles), 𝑅�� saturates and does not invert the degradation trend at high drain stress voltages. This can be appreciated also with the aid of Figure 5a), showing the simulated 𝐼�− 𝑉�� curves taken at 𝑉�� = 2 V after the off-state stress, for low and high drain pre-stress voltage ( 𝑉��,��� = 100 V and 600 V , blue and red curves, respectively), activating/deactivating impact ionization (solid and dashed curves, respectively). Indeed, Figure 5a) shows that the drain current in the linear regime, where the drain access resistance has an impact, is increased after the stress at 𝑉��,��� = 600 V with respect to the case at 𝑉��,��� = 100 V only if impact ionization is included in the device simulation. This result is consistent with the plot in Figure 4a) and with measured 𝐼�− 𝑉�� taken from [3] under similar conditions, shown in Figure 5b). The degradation followed by (partial) recovery of 𝑅�� was observed in several different devices as documented in the literature [3]–[6], see Figure 4b). We interpret this recovery behavior as a result of hole trapping into C-doping related acceptor traps in the buffer. This is in turn due to the generation of holes by impact ionization. This process is illustrated by Figure 6, showing contour plots of the net ionized acceptor trap concentration (i.e., 𝑁�� − 𝑁��) under the same conditions, with and without impact ionization. In the case with impact ionization included, compare Figure 6a), b), it is possible to observe a decrease in the concentration of the ionized acceptors with increasing drain stress voltage due to avalanche generated holes in the buffer and resulting hole trapping into C-related acceptor traps. Conversely, in the case without impact ionization included, compare Figure 6c), d), the ionized acceptor concentration increases (up to the effective acceptor concentration) with increasing stress, leading to the saturation of 𝑅��.

Other effects might arise which could explain the results discussed in this section. For example, the model in [4] relies on the vertical leakage paths between 2DEG and the carbon-doped GaN layer to explain the partial recovery of 𝑅��. In this work, we consider a simplified scenario

Accepted Article

at high drain stress voltage

Accepted Article

at high drain stress voltage −

Accepted Article

− 𝑉

Accepted Article

𝑉��

Accepted Article

�� 𝑉�� 𝑉

Accepted Article

𝑉�� 𝑉 curves

Accepted Article

curves stress voltage (

Accepted Article

stress voltage ( 𝑉

Accepted Article

𝑉��

Accepted Article

�� 𝑉�� 𝑉

Accepted Article

𝑉�� 𝑉 ,

Accepted Article

,���

Accepted Article

���

activating/deactivating impact ionization

Accepted Article

activating/deactivating impact ionization shows that the d

Accepted Article

shows that the drain current

Accepted Article

rain current after the stress at

Accepted Article

after the stress at

impact ionization is included in the

Accepted Article

impact ionization is included in the and with measured

Accepted Article

and with measured degradation

Accepted Article

degradation followed by (partial) recovery of

Accepted Article

followed by (partial) recovery of devices as documented in the literature

Accepted Article

devices as documented in the literature as a result of hole trapping into

Accepted Article

as a result of hole trapping into generation o

Accepted Article

generation of holes by impact ionization.

Accepted Article

f holes by impact ionization. contour plots of the

Accepted Article

contour plots of the net

Accepted Article

net

with and without impact ionization

Accepted Article

with and without impact ionization Figure

Accepted Article

Figure 6

Accepted Article

6a

Accepted Article

a),

Accepted Article

), b

Accepted Article

b),

Accepted Article

),

with increasing drain stress voltage

Accepted Article

with increasing drain stress voltage

Accepted Article

resulting hole trapping into C

Accepted Article

resulting hole trapping into C ionization included

Accepted Article

with no leakage paths and focus only on the role of C-doping and show that impact ionization might be at the origin of (or might contribute to) the recovery of 𝑅�� for high stress bias.

5. Conclusion

We conducted systematic device simulations of AlGaN/GaN HEMTs to assess the critical role of Carbon doping in the buffer. Through the simulation set-up calibrated through comparison with measurement data, we analyzed the breakdown voltage dependence on the gate-to-drain spacing, the current collapse under pulsed operation and the dynamic on-resistance recovery at increasing drain stress biases. We concluded that the observed breakdown voltage vs gate-to-drain spacing slope lower than the GaN critical field is due to the non-uniform electric field causing avalanche breakdown to occur in a localized spot of the drain access region. Current collapse is attributed to hole detrapping from the deep acceptor states associated with C-doping in the buffer during the off-state stress phase of pulsed measurements. Finally, it is suggested that generation of holes by impact ionization and subsequent hole trapping into the C-related acceptors can contribute to the dynamic on-resistance recovery at high drain stress voltages.

Received: ((will be filled in by the editorial staff)) Revised: ((will be filled in by the editorial staff)) Published online: ((will be filled in by the editorial staff)) References

[1] M. Higashiwaki, K. Sasaki, H. Murakami, Y. Kumagai, A. Koukitu, A. Kuramata, T. Masui, S. Yamakoshi, Semicond. Sci. Technol. 2016, vol. 31, p. 34001.

[2] E. Bahat-Treidel, F. Brunner, O. Hilt, E. Cho, J. Würfl, G. Trankle, IEEE Trans. Electron

Devices 2010, vol. 57, pp. 3050–3058.

[3] M. Meneghini, A. Tajalli, P. Moens, A. Banerjee, E. Zanoni, G. Meneghesso, Mater. Sci.

Semicond. Process. 2018, vol. 78, pp. 118–126.

[4] M. J. Uren, M. Caesar, S. Karboyan, P. Moens, P. Vanmeerbeek, M. Kuball, IEEE Electron

Device Lett. 2015, vol. 36, pp. 826–828.

Accepted Article

We conducted systematic de

Accepted Article

We conducted systematic de role of Carbon doping in the buffer

Accepted Article

role of Carbon doping in the buffer measurement

Accepted Article

measurement data,

Accepted Article

data, current collapse

Accepted Article

current collapse increasing drain stress biases

Accepted Article

increasing drain stress biases

spacing slope lower than the GaN critical field is due to the non

Accepted Article

spacing slope lower than the GaN critical field is due to the non avalanche breakdown

Accepted Article

avalanche breakdown to occur

Accepted Article

to occur

hole detrapping from the

Accepted Article

hole detrapping from the during the off

Accepted Article

during the

off-Accepted Article

-during the off-during the off

Accepted Article

during the

off-during the off state stress phase of pulsed measurements

Accepted Article

state stress phase of pulsed measurements

holes by impact ionization and subsequent hole trapping into the C

Accepted Article

holes by impact ionization and subsequent hole trapping into the C dynamic on

Accepted Article

dynamic

on-Accepted Article

-Accepted Article

resistance recovery at high drain stress

Accepted Article

resistance recovery at high drain stress

. Higashiwaki, K. Sasaki, H. Murakami, Y. Kumagai, A. Koukitu, A. Kuramata, T. Masui, S.

Accepted Article

. Higashiwaki, K. Sasaki, H. Murakami, Y. Kumagai, A. Koukitu, A. Kuramata, T. Masui, S. Yamakoshi,

Accepted Article

Yamakoshi, Semicond. Sci. Technol.

Accepted Article

Semicond. Sci. Technol.

E. Bahat

Accepted Article

[5] P. Moens, A. Banerjee, M. J. Uren, M. Meneghini, S. Karboyan, I. Chatterjee, P. Vanmeerbeek, M. Cäsar, C. Liu, A. Salih, E. Zanoni, G. Meneghesso, M. Kuball, M. Tack, Proc. of the IEEE

International Electron Devices Meeting (IEDM) 2015. DOI: 10.1109/IEDM.2015.749831.

[6] S. Karboyan, M. J. Uren, Manikant, J. W. Pomeroy, M. Kuball, Microelectron. Reliab. 2018,

vol. 81, pp. 306–311.

[7] E. Bellotti, F. Bertazzi, J. Appl. Phys. 2012, vol. 111, p. 103711.

[8] C. Koller, G. Pobegen, C. Ostermaier, D. Pogany, IEEE Trans. Electron Devices 2018, vol. 65, pp. 5314–5321.

[9] B. Rackauskas, M. J. Uren, S. Stoffels, M. Zhao, S. Decoutere, M. Kuball, IEEE Trans.

Electron Devices 2018, vol. 65, pp. 1838–1842.

[10] G. Meneghesso, R. Silvestri, M. Meneghini, A. Cester, E. Zanoni, G. Verzellesi, G. Pozzovivo, S. Lavanga, T. Detzel, O. Haberlen, G. Curatola, IEEE Int. Reliab. Phys. Symp. Proc. 2014. DOI: 10.1109/IRPS.2014.6861109.

[11] G. Verzellesi, L. Morassi, G. Meneghesso, M. Meneghini, E. Zanoni, G. Pozzovivo, S. Lavanga, T. Detzel, O. Häberlen, G. Curatola, IEEE Electron Device Lett. 2014, vol. 35, pp. 443–445.

[12] A. Chini, Meneghesso, M. Meneghini, F. Fantini, G. Verzellesi, A. Patti, F. Iucolano, IEEE

Trans. Electron Devices 2016, vol. 63, pp. 3473–3478.

[13] M. Singh, M. J. Uren, T. Martin, S. Karboyan, H. Chandrasekar, M. Kuball, IEEE Trans.

Electron Devices 2018, vol. 65, pp. 3746–3753.

[14] A. Fariza, A. Lesnik, J. Bläsing, M. P. Hoffmann, F. Hörich, P. Veit, H. Witte, A. Dadgar, A. Strittmatter, Appl. Phys. Lett. 2016, vol. 109, p. 212102.

[15] V. Joshi, S. P. Tiwari, M. Shrivastava, IEEE Trans. Electron Devices 2019, vol. 66, pp. 561– 569.

[16] N. Zagni, F. M. Puglisi, P. Pavan, A. Chini, G. Verzellesi, Microelectron. Reliab. 2019, vol.

Accepted Article

, pp. 306

Accepted Article

, pp. 306–

Accepted Article

–311.

Accepted Article

311. E. Bellotti

Accepted Article

E. Bellotti,

Accepted Article

, F. Bertazzi,

Accepted Article

F. Bertazzi,

C. Koller, G. Pobegen, C. Ostermaier

Accepted Article

C. Koller, G. Pobegen, C. Ostermaier pp. 5314

Accepted Article

pp. 5314–

Accepted Article

–5321.

Accepted Article

5321.

B. Rackauskas, M. J. Uren, S. Stoffels, M. Zhao, S. Decoutere, M. Kuball,

Accepted Article

B. Rackauskas, M. J. Uren, S. Stoffels, M. Zhao, S. Decoutere, M. Kuball,

Electron Devices

Accepted Article

Electron Devices 2018

Accepted Article

2018

G. Meneghesso, R. Silvestri, M. Meneghini, A. Cester, E. Zanoni, G. Verzellesi, G. Pozzovivo,

Accepted Article

G. Meneghesso, R. Silvestri, M. Meneghini, A. Cester, E. Zanoni, G. Verzellesi, G. Pozzovivo, S. Lavanga, T. Detzel, O. Haberlen, G. Curatola,

Accepted Article

S. Lavanga, T. Detzel, O. Haberlen, G. Curatola, DOI: 10.1109/IRPS.2014.6861109.

Accepted Article

DOI: 10.1109/IRPS.2014.6861109.

Accepted Article

G. Verzellesi, L. Morassi, G. Meneghess

Accepted Article

G. Verzellesi, L. Morassi, G. Meneghess Lavanga, T. Detzel, O. Häberlen, G. Curatola,

Accepted Article

Lavanga, T. Detzel, O. Häberlen, G. Curatola, 445.

Accepted Article

445.

A. Chini, Meneghesso, M. Meneghini, F. Fantini, G. Verzellesi, A. Patti, F. Iucolano,

Accepted Article

A. Chini, Meneghesso, M. Meneghini, F. Fantini, G. Verzellesi, A. Patti, F. Iucolano,

Trans. Electron Dev

Accepted Article

Trans. Electron Devices

Accepted Article

ices

M. Singh, M. J. Uren, T. Martin, S. Karboyan, H. Chandrasekar, M. Kuball,

Accepted Article

M. Singh, M. J. Uren, T. Martin, S. Karboyan, H. Chandrasekar, M. Kuball,

Electron Devices

Accepted Article

Electron Devices

Accepted Article

2018

Accepted Article

2018

A. Fariza, A. Lesnik, J. Bläsing, M. P. Hoffmann, F. Hörich, P. Veit, H. Witte, A. Dadga

Accepted Article

A. Fariza, A. Lesnik, J. Bläsing, M. P. Hoffmann, F. Hörich, P. Veit, H. Witte, A. Dadga

100–101, p. 113374.

[17] A. J. Fischer, J. R. Dickerson, A. M. Armstrong, M. W. Moseley, M. H. Crawford, M. P. King, A. A. Allerman, R. J. Kaplar, M. S. van Heukelom, J. J. Wierer, Electron. Lett. 2016, vol. 52, pp. 1170–1171.

Figure 1. Simulated AlGaN/GaN HEMT cross-section to reproduce results from [2].

Figure 2. a) 𝐼�− 𝑉�� curves, showing the agreement between measurements from [2] (red squares) and simulations (black solid curves). b) 𝑉�� vs. 𝐿�� curve showing that the simulation set-up reproduces the experimental breakdown data.

Accepted Article

Simulated AlGaN/GaN

Accepted Article

Simulated AlGaN/GaN

Accepted Article

Accepted Article

𝐼

Accepted Article

𝐼�

Accepted Article

� 𝐼� 𝐼

Accepted Article

𝐼� 𝐼 −

Accepted Article

− 𝑉

Accepted Article

𝑉��

Accepted Article

�� 𝑉�� 𝑉

Accepted Article

𝑉��

𝑉 curves, showing the agreement between measurements from

Accepted Article

curves, showing the agreement between measurements from and simulations (black solid curves). b)

Accepted Article

and simulations (black solid curves). b) reproduces the experimental breakdown data

Accepted Article

reproduces the experimental breakdown data

Accepted Article

Accepted Article

Figure 3. Current-Collapse (CC) in the a) simulated (lines) and b) measured (symbols) pulsed 𝐼�− 𝑉�� curves. Curves have been taken for above-threshold 𝑉�� of 1 V and 2 V. CC is evaluated as the drain current decrease 𝛥𝐼� occurring due to stress in the condition determined by the baseline voltages 𝑉��,�� and 𝑉��,��.

Figure 4. Dynamic 𝑅�� degradation and recovery in a) simulated and b) measured devices. In a) simulations are carried out with (red squares) and without (yellow triangles) impact ionization, to show that only with hole generation it is possible to qualitatively reproduce the 𝑅�� recovery behavior found in experimental data shown in b). A typical stress time of 100 s was adopted for simulations.

Accepted Article

Current

Accepted Article

Current-Accepted Article

-Collapse (CC) in the a) simulated (lines) and b) measured (symbols) pulsed

Accepted Article

Collapse (CC) in the a) simulated (lines) and b) measured (symbols) pulsed curves. Curves have been taken for above

Accepted Article

curves. Curves have been taken for above as the drain current decrease

Accepted Article

as the drain current decrease ��

Accepted Article

��and

Accepted Article

and 𝑉

Accepted Article

𝑉��

Accepted Article

�� 𝑉�� 𝑉

Accepted Article

𝑉�� 𝑉 ,

Accepted Article

,��

Accepted Article

��

Accepted Article

Dynamic

Accepted Article

Dynamic 𝑅

Accepted Article

𝑅��

Accepted Article

��

simulations are carried out with (

Accepted Article

simulations are carried out with ( show that only with

Accepted Article

show that only with

Accepted Article

hole

Accepted Article

hole

behavior found in experimental data shown in b).

Accepted Article

behavior found in experimental data shown in b).

Figure 5. a) Simulated 𝐼� − 𝑉�� curves taken at 𝑉�� = 2 𝑉 for low (𝑉��,��� = 100 𝑉, blue curves) and high (𝑉��,��� = 600 𝑉, red curves) drain pre-stress in off-state (𝑉��,��� = −5 𝑉) held for 100 s. Solid (dashed) lines are simulations with (without) impact ionization activated. The drain current recovers for high stress only if the impact ionization is active, indicating less electron charge is trapped in the gate-to-drain access region. b) Comparison with measurement data taken from [3] shows a similar trend with respect to the simulation results.

Figure 6. Contour plots of the net ionized acceptor trap density for the cases with (a, b) and without (c, d) impact ionization activated after 100 s of applied off-state stress with 𝑉��,��� = −5 𝑉 and (a,c) 𝑉��,��� = 100 𝑉, (b,d) 𝑉��,��� = 600 𝑉. The scale of the legend is chosen to emphasize the difference in the buffer of the net ionized acceptor density (the maximum value actually exceeds this range).

Table of Content

The Role of Carbon Doping on Breakdown, Current Collapse and Dynamic RON Recovery in

AlGaN/GaN High Electron Mobility Transistors on Semi-Insulating SiC Substrates

Nicolò Zagni*, Alessandro Chini, Francesco Maria Puglisi, Paolo Pavan, Giovanni Verzellesi

Accepted Article

Simulated

Accepted Article

Simulated 𝐼

Accepted Article

𝐼�

Accepted Article

� 𝐼� 𝐼

Accepted Article

𝐼� 𝐼 ��

Accepted Article

��,

Accepted Article

,���

Accepted Article

��� =

Accepted Article

= 600

Accepted Article

600

Solid (dashed) lines are simulations with (without) impact ionization activated. The drain current

Accepted Article

Solid (dashed) lines are simulations with (without) impact ionization activated. The drain current recovers for high stress only if the impact ionization is

Accepted Article

recovers for high stress only if the impact ionization is n the

Accepted Article

n the gate

Accepted Article

gate-Accepted Article

-to

Accepted Article

to-Accepted Article

-Accepted Article

drain access region

Accepted Article

drain access region

shows a similar trend with respect to the simulation results.

Accepted Article

shows a similar trend with respect to the simulation results.

Accepted Article

Contour plots of the

Accepted Article

Contour plots of the

(c, d) impact ionization activated after 100 s of applied off

Accepted Article

(c, d) impact ionization activated after 100 s of applied off =

Accepted Article

= 100

Accepted Article

100 𝑉

Accepted Article

𝑉, (b,d)

Accepted Article

, (b,d)

Accepted Article

difference in the buffer of the

Accepted Article

difference in the buffer of the

Table of Content

Accepted Article

Table of Content

Accepted ArticleAccepted Article

In this work, a carbon-related buffer-trap model for AlGaN/GaN HEMTs is developed by calibrating device simulations on static current-voltage characteristics and breakdown

measurements. With this model, it is possible to explain the experimentally observed dynamic on-resistance recovery after off-state stress with hole generation and trapping in carbon-related deep acceptor states in the buffer.

Accepted Article

In this work, a carbon

Accepted Article

In this work, a

carbon-Accepted Article

-related buffer

Accepted Article

related buffer calibrating device simulations on static

Accepted Article

calibrating device simulations on static measurements.

Accepted Article

measurements. With this model, it is possible to explain

Accepted Article

With this model, it is possible to explain resistance recovery after off

Accepted Article

resistance recovery after off acceptor states in the buffer.

Accepted Article

acceptor states in the buffer.

Accepted Article