Gallium Nitride (GaN) has seen a rapid development in applications in recent years, starting from the consumer electronics sector, such as fast-charging adapters for mobile phones, and has expanded into data centers, photovoltaics, energy storage, and even electric vehicles and charging piles. According to forecasts by TrendForce, the global power GaN market is expected to grow from $180 million in 2022 to $1.33 billion by 2026, with a compound annual growth rate (CAGR) of 65%.

In terms of gate technology for power GaN devices, there are currently two mainstream directions, which are the enhancement-mode (E-Mode) and the depletion-mode (D-Mode). What are the differences between these two technological paths?

Depletion-mode (D-Mode) GaN Devices

D-Mode GaN is a normally-on device, which means it is in a conductive state without the application of a gate voltage. A negative gate voltage is required to reduce the electron concentration in the channel, depleting the channel electrons, thereby reducing or shutting off the current.

Advertisement

In GaN devices, the two-dimensional electron gas is spontaneously formed at the interface between the GaN and AlGaN layers. This two-dimensional electron gas layer has a very high electron density and mobility, enabling GaN devices to achieve fast electron transport and low resistance, leading to high efficiency and high-speed switching performance. D-Mode GaN retains these advantages, thus offering extremely fast switching speeds.

Due to its high-frequency switching and low on-resistance characteristics, D-Mode GaN is suitable for high-efficiency and high-frequency applications, with significant potential in high power density and high-power applications, such as photovoltaic microinverters, On-Board Chargers (OBCs), charging piles, and data center power supplies.

In practical applications, the normally-on D-Mode GaN devices cannot be used directly and require additional peripheral components to be converted into a normally-off configuration. For example, in fast-charging applications, D-Mode GaN devices typically need to be used in series with a low-voltage silicon MOSFET, using the switching of the low-voltage silicon MOSFET to drive the overall switch, converting the normally-on device into a normally-off device.

D-Mode GaN mainly includes two technological architectures, which are the cascode and direct drive. The previously mentioned series connection of low-voltage silicon MOSFET to drive the overall switch is the cascode D-Mode GaN architecture, which is also the current mainstream architecture. It offers higher reliability and is more suitable for high-power, high-voltage, and high-current applications, such as in electric vehicles and industrial applications.

The driving compatibility of D-Mode GaN is also better. For instance, the cascode D-Mode GaN can use the same driving circuit as traditional silicon MOSFETs. However, due to the switching frequency and speed being much higher than those of silicon MOSFETs, the driving integrated circuit must be able to operate stably under high dv/dt conditions.

The main power GaN manufacturers following the D-Mode path include Transphorm, Power Integrations (PI), Texas Instruments (TI), Nexperia, Gafuture, CR Micro (Runxin Micro), Nenghua Micro, and Xingguo, among others.Enhanced Mode (E-Mode) GaN Devices

Enhanced mode GaN is a normally-off device, which means that without the application of a gate voltage, the device is in an off state and does not conduct electricity. A positive gate voltage must be applied to form a conductive channel, allowing the device to conduct.

As previously mentioned, in GaN devices, the two-dimensional electron gas (2DEG) is spontaneously formed at the interface between the GaN and AlGaN layers. To achieve normally-off operation, it is necessary to introduce a p-type doped GaN layer (p-GaN) beneath the gate. This doping creates an inherent negative bias, akin to a small built-in battery, which helps to deplete the 2DEG channel, thereby achieving normally-off operation.

Enhanced mode GaN can directly achieve 0V off and positive voltage conduction without compromising the conduction and switching characteristics of GaN. In hard-switching applications, it can effectively reduce switching losses and EMI noise, making it more suitable for low to medium power applications with strict fault safety requirements, such as DC-DC converters, LED drivers, chargers, etc.

In applications, since E-Mode GaN is a normally-off device, its usage can be similar to that of traditional silicon MOSFETs, but it requires a more complex peripheral circuit design. For direct driving, it is necessary to match it with a dedicated GaN driver IC. Examples include the E-mode GaN dedicated high-voltage half-bridge gate driver NSD2621 from Naxinwei and the GaN EiceDRIVER series high-side gate driver from Infineon.

However, E-Mode GaN devices also have disadvantages. First, because they need to control the two-dimensional electron gas to achieve the normally-off characteristic, this may affect the device's dynamic performance, and it may not switch as fast as D-Mode devices. Additionally, in terms of thermal management, the on-resistance of E-Mode GaN increases with temperature, so more reliable heat dissipation design is needed for high-power applications. The gate of E-Mode GaN devices may also have stability and leakage current issues, which could affect the reliability and performance of the devices.

The main power GaN manufacturers that follow the E-Mode approach include Infineon (GaN Systems), EPC, Innoscience, Navitas Semiconductor, Panasonic, and Quanta-Bolt.

How to choose between E-Mode and D-Mode?

In practical applications, how to choose between E-Mode or D-Mode GaN devices? Based on the analysis of the two forms of GaN devices mentioned earlier, the choice can be made from several aspects according to the actual application requirements.

Firstly, consider the system's safety requirements. If the application requires fault-safe operation, the normally-off E-MODE GaN is a better choice. On the other hand, the normally-on D-Mode GaN is more suitable when the system can achieve safe operation through external control, as D-Mode GaN can provide faster switching speeds, higher frequencies, and lower losses.In high-frequency applications, D-Mode has a relative advantage, and its characteristics of low on-resistance and high-speed switching will perform better in scenarios with limited heat dissipation or where high-efficiency conversion is required.

In terms of driving circuits, since E-Mode GaN is a normally-closed device, the driving circuitry is relatively simpler, while D-Mode GaN requires a more complex circuit design. Similarly, if upgrading and integrating into existing systems, the usage and characteristics of E-Mode that are similar to traditional silicon MOSFETs can also be an advantage.

In harsh environments such as industrial or automotive applications, D-Mode exhibits higher reliability and longevity in high-temperature and high-voltage conditions, making it more suitable for these application scenarios.