GaAs has been with us for some time and has provided high efficiency, high power devices in the low GHz regime. Depending on the particular application, it might not be obvious whether your power electronics or RF application needs GaN vs. GaAs components. So which material and components are best for your next application?
GaN is becoming more commercialized and extensively used thanks to its higher efficiency at mmWave frequencies compared to GaAs. If you need a high-efficiency power amplifier, you’ll need to consider the material properties of these two important materials. Let’s take a brief dive into the important material properties of GaN vs. GaAs components for power electronics and RF applications.
The important material properties of these two semiconductor materials are electron mobility, thermal conductivity, energy bandgap, and bandwidth. For power amplifiers, a higher energy bandgap allows the device to be run at higher voltage as the breakdown field in the device will be higher for a given device size. The thermal conductivity is also important from a device lifetime perspective. When the thermal conductivity is larger, the device will run cooler, which will provide longer lifetime. In addition, since a device with higher thermal conductivity runs cooler, physically smaller heat sinks can be used with the device. Although cost of producing GaN on 4H-SiC is greater than the cost to produce a GaAs wafer, the overall amplifier cost is lower (both for integrated circuits and larger amplifier units) due to lower packaging/housing and heat sink costs.
The bandwidth is a function of crystal structure and the construction of the device itself. Bandwidths can reach about a decade in both devices, although the center frequencies are different for GaN vs. GaAs. Finally, the electron mobility is related to the conductivity of the device in the ON state. A device with higher electron mobility will have higher conductivity in the ON state, thus a power amplifier or RF amplifier will be more efficient. In the right frequency ranges, each type of device will have different efficiencies (more on this later).
The table below shows a comparison of these important material properties for GaN vs. GaAs. The material properties for Si are shown for comparison.
Here, we see that GaN has a bandgap in the UV, making it an ideal material for integrated UV photonic/electronic circuits. This makes the breakdown field an order of magnitude larger than that of Si and GaAs. We see that GaN and Si have similar thermal conductivity values, which are much higher than that of GaAs.
For power amplifiers, both for DC and at RF frequencies, one should note that GaN has higher electron mobility in the inversion layer than in the bulk crystal, while the opposite is true in Si. Charge carriers in the inversion layer will move easier through the active region in a GaN device in the ON state as the ON state resistance is lower. This means GaN has higher efficiency and available power output than Si and GaAs in the ON state. In general, this extends up to higher frequencies than Si amplifiers.
Amplifier power output vs. frequency regions for different semiconductor materials. Source: Analog Devices.
Choosing a GaN vs. GaAs power amplifier for RF applications and power electronics applications is all about balancing the relevant frequency range against efficiency and cost. As GaN is normally deposited on SiC, it will have much higher efficiency at high frequencies while also having longer lifetime. This takes advantage of the high thermal conductivity of 4H-SiC (4H-SiC is 490 W/m•K), which easily dissipates heat down to a die-attached paddle.
The primary high frequency application in automotive, defense, and aerospace is mmWave W-band radar (for automotive) and M-band radar (NATO’s band). GaN devices can support these higher frequencies thanks to their flat dispersion. W-band radars are moving away from Si and GaAs in favor of GaN devices as the larger current output equates to higher total power output. This, in turn, provides longer range for an mmWave radar module.
There is another aspect of GaN that is not often considered, which is the high power input rolloff in power amplifiers. Compared to GaAs, GaN has a more gentle rolloff as the input power increases and eventually drives the device into saturation (see the corresponding graph). This sets the driving power corresponding to the 3OIP (third-order intermodulation point, read more here) to a higher value.
This, in turn, lowers the 1 dB compression point. Although a 1 dB compression point is viewed as desirable from the standpoint of distortion, the real important value is the linearity in the input-output curve. This will limit the input power at which intermodulation products in FM signals become noticeable in the sidebands. This will then put pressure on filter design to suppress these intermodulation products on the amplifier output.
Comparison of GaN vs. GaAs amplifier saturation. [Image source]
During amplifier design, both for ICs and PCBs, amplifier behavior can be determined using circuit simulations with the right GaN SPICE models. Circuit models have been difficult to build in the past, although the current industry-standard circuit model for GaN power transistors is the Angelov model. Without a circuit model, analytical equations will need to be used with standard values for material properties to analyze device behavior in a simulation.
If you’re looking for a PCB design and layout firm to help you build and deploy switching power systems or RF devices, look no further than NWES. We know how to help you decide between GaN vs. GaAs devices and custom PCBs for your new system. We’re also a digital marketing firm, and we’ll develop an SEO-driven content strategy to market your new product and engage with your target customers. Contact NWES today for a consultation.