2.2. The Physical Models and Other Simulation Information for the TCAD
In this paper, the physical models used for simulation include the impurity incomplete ionization model, the mobility model, the impact ionization model, the drift diffusion model, the carrier recombination model (SRH and Auger), the bandgap narrowing model, et al. In addition, the heavy ion irradiation model will be used when studying SEE [
18].
The diagram of the heavy ion irradiation model is shown in
Figure 2.
The generation rate of electron hole pairs caused by heavy ions can be calculated using the following equation [
18].
In Equation (1), the density of the generated electron hole pairs is determined GLET(l), which can be controlled by defining the Linear Energy Transfer (LET). The variable corresponding to LET in the simulation software is LET_f; R(ω,l) and T(t) describe the range and time of the generated electron hole pairs respectively.
The occurrence of SEE in devices is influenced by the LET value; LET is used to describe the energy loss of particles, that is, the average energy loss per unit path length. When the incident particles impact with the SiC atomic lattice, the LET value remains unchanged over a short distance, and the energy loss of the incident particle can generally be characterized by the surface LET value of the device. The different LET energy values at the time of the particle incident correspond to different particles.
R(
ω,
l) can be defined as either an exponential function or a Gaussian function, with the corresponding equation as follows [
18]:
In Equations (2) and (3), ω is the radius of the ion trajectory, the characteristic distance is ωt, and the corresponding variable in the simulation software is Wt_hi.
T(
t) is a Gaussian function; the corresponding equation is as follows [
18].
In Equation (4), t0 is the time of ion incidence and Shi is the characteristic value of the Gaussian function.
When adding a heavy ion model to the physical process of the numerical analysis module named Sentaurus Device in the software, the main parameter variables to be set for the heavy ion irradiation model are energy value (Let_f), incident position (Location), incident direction (Direction), incident depth (Length), characteristic distance (W
t_hi), time, et al. In SEB simulation, the selection of basic parameters for the generation of incident particles is based on the typical values of SiC MOSFETs. Specifically, the particles are incident vertically into the device, the characteristic distance is 0.05 μm, the depth is that the particles are incident throughout the entire device, the LET is 75 MeV·cm
2/mg, and the time is set to 10
−11 s [
19]. These parameters are initial values set based on typical values in SiC devices, and can be validated and optimized based on the results of the SEE test in the future.
2.3. Principle of Single-Event Effect
The process of SEE is as follows: in a blocked state, a large number of electron-hole pairs are generated by impact ionization after heavy ions are incident into SiC MOSFET. In the electric field formed by the bias voltage of the drain-source, electrons move towards the drain electrode, holes move towards the source electrode, enter the P-well, and then move to the P+ body contact region, which changes the electric field distribution in the N-drift, and a high electric field peak appears at the homojunction formed by the N-drift and substrate. In the action of the high electric field, intense impact ionization occurs after acceleration, further producing extremely high concentrations of electron-hole pairs. When the carrier avalanche doubling effect occurs, it causes a transient large current. The source and drain electrodes form a conductive path.
SEB is mainly affected by the electric field. The electric field provides energy for the impact ionization process between electron-hole and silicon carbide atoms. The electron-hole pairs accelerate the impact ionization process by obtaining kinetic energy through the energy provided by the electric field. The relative velocity of generation and recombination of electron-hole pairs is affected by the magnitude of the electric field. When the electric field is weak and the generation speed of the electron-hole pairs is smaller than the recombination speed, the magnitude of the single-event transient current will gradually decrease to zero over time; in the case of a strong electric field, the generation speed of electron-hole pairs is greater than the recombination speed, the transient current remains at a stable level over time, SEB effect occurs, and when the bias voltage applied to the device causes the SEB effect to occur, the critical voltage is defined as the SEB threshold voltage, at which the device that experiences SEB will permanently fail.
2.4. Analysis of Single-Event Burnout Effect at Different Incident Positions
The sensitivity of different positions of the device to incident particles varies.
Figure 3 shows three SEE incident positions selected from three different regions of the device, namely position A above the midpoint of the JFET region, position B above the edge of the split gate, and position C above the PN junction formed by the N+source and P-well in the source region.
Figure 4 shows the transient current curves at different incident positions in different regions. According to the experimental and the irradiation requirements in the actual space environment, the LET is not less than 75 MeV·cm
2/mg, the unit of energy is represented by pC/μm in the simulation, and the conversion formula for LET is 1 pC/μm = 151 MeV/mg/cm
2; therefore, the LET is generally greater than 0.5 pC/μm [
20,
21]. During the simulation research, the bias conditions are as follows: LET = 0.5 pC/μm, V
GS = 0 V, V
DS = 400 V.
From
Figure 4, it can be seen that the transient current variation trend is consistent at different incident positions. After heavy ions are incident into the device from different positions, the current increases sharply with time, reaching a transient peak current at around 10 ps. Since the bias voltage does not reach the SEB threshold voltage of the device, the transient current will recover to 0 after a certain period of time. Among them, position B has the greatest variation and is most sensitive to the SEB effect, and the transient current changes most dramatically with time.
Figure 5 shows the transient current curves over time at different bias voltages when incident from position B, which is the most sensitive to SEE. From
Figure 5, it can be seen that when the bias voltage reaches 430 V, the drain current reaches its peak at 10 ps and then remains at a stable value, and does not return to the initial state after 10
−7 s, resulting in device burnout. Therefore, the analysis suggests that the threshold voltage for SEB at position B is approximately 430 V. Subsequently, the SEB threshold voltages at positions A and C were analyzed using the same method.
Table 1 lists the SEB threshold voltages at different incident positions.
From
Table 1, it can be seen that, although position B is more sensitive, the SEB thresholds’ voltage of position A and position B are approximately the same because the current peak of position A and position B are very close. Overall, when heavy ions are incident from the JFET region near the concentration of the electric field, the SEE of the device is more sensitive and the SEB threshold voltage is lower than when they are incident from the source region near the presence of parasitic BJT.
2.5. The Regularity of Electric Field Peak Distribution Transfer
SEB is closely related to the internal electric field of the device.
Figure 6 shows the electric field distribution transfer when heavy ions are incident from position B on the condition of the drain-source bias voltage of 450 V and SEB occurring.
From
Figure 6, it can be seen that before the heavy ion incident, the electric field intensity at the PN junction formed in the P-well and N-drift is the highest. After the heavy ion irradiation, the electrons and holes generated by impact ionization move in the opposite direction at the drain voltage, the holes gradually accumulate in the JFET and channel regions below the gate, causing an increase in the electric field on the channel surface of the device, and the electric field peak gradually transfers to the channel surface. When the accumulated holes cannot be removed in time, the electric field will couple to the gate oxide layer. At 10~100 ps, the maximum electric field is located in the gate oxide layer, as shown in
Figure 6e,f. As the high current inside the device changes, a large amount of charge will be concentrated on the substrate surface, which causes the peak electric field to also transfer to there. After 10
−7 s, the maximum electric field will transfer to the homogeneous junction formed between the N-drift and the substrate surface.