5G: The Market GaN Needs, Discussed by IDTechEx

Estimated reading time: 3 minutes

People could be forgiven for assuming that 5G is already here and a done thing. However, much of the 5G rollout to date has been in lower frequency bands more similar to existing 4G rather than taking full advantage of the higher frequencies. The current state of 5G has seen less technological innovation than might have initially been expected with the promise of high frequency, Gigabit download speeds, and millisecond latency yet to be realized in a major way. There is certainly more scope for technical development and hence opportunities for several technologies and materials, a critical one of which is the semiconductor technology with wide bandgap semiconductors. IDTechEx's latest research report on "Thermal Management for 5G 2022-2032" finds that GaN (gallium nitride) has a significant opportunity within the 5G market and this creates a downstream effect on other components such as die attach materials.

LDMOS (laterally-diffused metal-oxide semiconductor) devices have been the technology of choice for power amplifiers through the 4G era. These power amplifiers provide the crucial role of boosting the signal for transmission. The trouble is once we move above around 4 GHz, LDMOS starts to become inefficient. Efficiency is a critical factor for telecoms infrastructure as it directly impacts the energy consumption of the antenna. With much of 5G infrastructure being deployed alongside existing equipment, the energy consumption of telecoms towers is set to increase dramatically, adoption of wide bandgap semiconductors like GaN is one method to reduce this future impact. GaN provides greatly improved efficiencies at higher frequencies, depending on the specific use case, this can be in the region of 10% or more improvements for efficiency.

GaN started being deployed in 4G networks with Huawei equipment but has seen only moderate adoption globally so far due to the increased price, lower manufacturing availability, and difficulty in integration with other components. However, it is far from niche technology developed by startup companies, for example, Japanese technology giant Sumitomo provides RF components to companies like Huawei including their GaN devices. With the growth of 5G continuing, especially for higher frequencies, we're expecting a significant uptake in GaN over the next decade, especially for the higher end of the sub-6 GHz infrastructure where higher powers are being used and component integration is not quite as challenging as it is in the mmWave region. For this application, IDTechEx predicts a 4 fold increase in GaN demand per year by the end of the decade.

The adoption of wide bandgap semiconductors typically raises the junction temperature of devices and starts to bring more thermal management considerations. One critical failure point with thermal cycling is how the semiconductor device is connected, or the die attach material. Junction temperatures for GaN devices are often above 175 °C, at this point we start to limit the options for typical solder materials, especially when lead-free is a requirement in most markets. This is leading many players to consider sintering materials. Sintering involves the application of a (typically silver) paste that is heated causing densification. The upshot is a more reliable connection with improved thermal conductivity. This has already started to be adopted in a big way in the electric vehicle market thanks to the transition to SiC (silicon carbide) and 800 V platforms.

The key limitation historically has been the lack of commercial experience, long curing times, and the need for an inert atmosphere or higher pressures, but developments of these materials, greater market adoption, and the trend towards GaN could see sintering start to make a big impact in the 5G market too. IDTechEx is expecting a 10-fold increase in demand for sintering materials in 5G infrastructure by 2030. There is also a great interest in the development of copper sintering materials over silver due to the potentially reduced costs and improved performance but falls into the same issues as silver sintering had originally compared to solder.