From Principle to Practice: Why is GaN so promising?

Power semiconductors are the core of power conversion and circuit control in electronic devices. They mainly refer to discrete semiconductor devices that can withstand high voltage or high current. They are mainly used to change the voltage and frequency, DC to AC conversion, etc. in electronic devices. In the development path of power semiconductors, power semiconductors have been comprehensively improved in terms of structure, process, technology, process, integration, materials, etc., and the main direction of its evolution is higher power density, smaller size, lower cost and loss. In particular, in terms of material iteration, the upgrade from silicon Si materials to wide bandgap materials such as gallium nitride (GaN) has significantly improved the size and performance of power devices.

    So what is the third-generation semiconductor GaN? It is a semiconductor material composed of nitrogen and gallium. Because its bandgap width is greater than 2.2eV, it is also called a wide bandgap semiconductor material.

Table 1 Comparison of key properties of GaN and Si

    Table 1 compares several physical parameters of GaN and Si. It is undeniable that GaN shows better performance advantages, which can be divided into the following four points:

    1. Wide bandgap: Wide bandgap enables the material to withstand higher temperatures and greater electric field strengths. When the operating temperature of the device rises, the concentration of intrinsically excited carriers will not be very high, so it can be applied to special environments with higher temperatures.

    2. High breakdown electric field: The breakdown field strength of GaN itself is 3.3E+06, which is about 11 times that of Si. Under the same withstand voltage conditions, the width of the GaN depletion region can be reduced to 0.1 times that of Si, which greatly reduces the resistivity of the drift region to obtain lower Ron and higher power performance.

    3. High electron saturation drift rate: In the working process of semiconductor devices, most of them use electrons as carriers to realize current transmission. High electron saturation drift rate can ensure that semiconductor devices can still maintain high mobility when working in high electric field materials, and thus have a large current density, which is the key to the device to obtain a large power output density. This is also the most obvious advantage of GaN materials.

    As you can see in the table, the electron mobility of GaN is not high. Why is it called a high electron mobility transistor? The reason is that GaN & AlGaN form a two-dimensional electron gas (2DEG) at the interface due to the material characteristics. 2DEG exists in a thin layer of 2-4nm and is confined to a very small range. This confinement increases the electron mobility to 1500~2000cm²/(V·s). Current technology has made the electron mobility reach 2200 cm²/(V·s).

    4. Good temperature resistance: It can be seen that the thermal conductivity of GaN and Si is basically the same, but GaN can have a higher junction temperature than Si. Therefore, good thermal conductivity and higher thermal tolerance can jointly improve the service life and reliability of the device.

    The superior performance of GaN devices is also closely related to their device structure. At present, the two routes of industrialized GaN devices are P-GaN enhanced devices and common source and common gate structures. The two structures have different opinions in the market, and everyone has their own opinions.

Figure 1 Two mainstream GaN structures

    Since GaN devices are extremely sensitive to parasitic parameters, the driving requirements of GaN are more stringent than those of traditional Si-based semiconductor devices, so it is very meaningful to study its driving circuit. In the actual application of high-voltage power GaN devices, we compared the switching characteristics and dynamic characteristics of GaN devices and the current mainstream SJ MOSFET to understand the differences in more detail.

Table 2 GaN device DC parameters

    From the DC parameters of GaN transistors in the figure above, we can see that there is no reverse diode (0 Reverse Recovery) in DC parameters, mainly because GaN transistors do not have the parasitic PN junction of SJ MOSFET. In addition, there are also significant differences in DC parameters and Vth between the two. Under the same specifications, GaN transistors have smaller saturation current and higher BV value than SJ MOS, which is also limited by its special characteristics of chip area and no avalanche capability; at the same time, lower driving voltage and gate charge Qg create its excellent switching characteristics of high frequency and low loss.

Figure 2 Comparison of GaN and Si capacitor characteristics

    From the capacitance of the device, we can see that the capacitance of SJ MOSFET has obvious nonlinear characteristics within 50V, and the overall capacitance value is much larger than that of GaN devices (the junction capacitance is 3 times that of GaN). This is because, although the two-dimensional electrically coupled SJ device has a smaller device area than the planar MOS, it relies on the lateral depletion of the PN junction to achieve voltage resistance, so the contact area of ​​the PN junction is much larger. When the voltage between the device DS is low, the contact surface formed by the built-in electric field of the PN junction causes its initial Coss & Crss and other parameters to be several orders of magnitude larger than the DS high voltage state; at the same time, the device goes from incomplete depletion to full depletion state, and the device space charge region widens, resulting in mutation points in the CGD and CDS capacitance curves. This sudden change of the electric field in a very narrow voltage range also affects the EMI problem that engineers are concerned about. How to optimize it to slow down the curve has become a special process for many design companies.

Figure 3 Cgd mutation of silicon device 

    However, the emergence of GaN has easily solved this problem. The capacitance curve of GaN changes in a relatively small range and there is no sudden change. Therefore, in the EMI debugging process of power supply applications, the effect is better than SJ MOSFET. The nearly linear Coss makes the dv/dt waveform of the application switching process closer to a slanted line without arc, making it elegant.

 Figure 4 GaN flyback Vds switch rising edge

    The lower junction capacitance also makes the device's energy equivalent capacitance (Coer) and Eoss much smaller than SJ MOS devices of the same specification, which greatly reduces the capacitive loss of the power supply during hard switching and can significantly reduce device heating. At the same time, less junction capacitance charge is extracted to achieve ZVS during soft switching of the power supply, allowing the system to have a higher switching frequency and a smaller dead time, further reducing the system size.

Figure 5 Differences in Eoss and Coer between GaN and Si devices

    As the high efficiency of GaN is verified in practice, the market's confidence in GaN is gradually increasing, its advantages are becoming more and more significant, and its usage is growing. In the future, power GaN technology will become the new standard for high-efficiency power conversion. The following is the E-Mode GaN device newly launched by Wei'an. You are welcome to request samples and discuss its characteristics with Wei'an's experts.