April 03, 2025 
Professional magazine and portal dedicated to electronics 
The maturation of gallium nitride (GaN) is well underway. GaN has scaled up with the charger market which needed the power and efficiency benefits GaN offers to deliver on the promise of USB-PD.
Now the industry is shipping millions of parts into this lucrative market, and the GaN charger market size [1], valued at $1.10B in 2023, is predicted to reach $4.22B by 2030, with a CAGR of 19.9% from 2024 to 2030, according to market research and management consulting firm Next Move Strategy Consulting. But if GaN can be said to have ‘grown up’ with the charger market by delivering huge volumes of reliable product, it is now set for its next chapter. The platform of available GaN devices is burgeoning from a few key products into a wide variety of GaN devices – both high and low voltage, and with integrated solutions as well as discrete HEMTs – to suit a much wider range of applications.
As a quick reminder, GaN is revolutionizing the power electronics market because it delivers more power, at a higher power density and greater efficiency with almost zero switching losses. GaN has a Figure of Merit, Ron Qg, which is 10 times better than silicon. Therefore, systems can switch at a higher frequency without incurring any loss in efficiency, giving designers the freedom to use smaller passives and reduce the overall BOM. High efficiency also means that there is less heat to dissipate, so the thermal management challenge is less onerous and heatsinks and fans can be eliminated or reduced in size.
GaN is also a compelling solution because simple power conversion topologies can be used. GaN has no body diode, so the reverse recovery current is zero. Therefore, for example, a totem pole bridgeless PFC design can be employed, rather than the more complex, greater component count, dual boost bridgeless PFC circuit which would be necessary when using even the best silicon MOSFETS to keep efficiency high.
If GaN is to proliferate in a broad range of application fields as it has done in the USB-PD charger market, a much broader selection of GaN devices will be required. Industry is responding; as an example, Innoscience, the world’s largest 8in GaN-on-silicon integrated device manufacturer (IDM), now offers a very wide portfolio of GaN power devices that addresses low, medium and high voltage applications. The company offers low voltage parts spanning 30 - 150V with ON resistances as low as 1.2mΩ, and high voltage devices up to 700V with RDS(on) values between 30 - 600mΩ. These GaN FETs are also available in different packages including wafer level CSP, flip chip QFN with a top- or bottom-side side cooling, DFN and TO-xxx styles. Therefore, designers that are used to working with those packages from their experience of silicon devices can easily switch to GaN. To support these devices, Innoscience recently introduced a single- channel gate-driver optimized for GaN HEMTs.
Switching to GaN can deliver significant performance increases as we have already discussed. By combining these new GaN devices, even greater progress can be made. Consider the 4.2kW power supply unit (PSU) in Figure 1 which uses a 700V high voltage GaN part on the primary side together with a 150V GaN device on the secondary side. The totem pole PFC plus LLC design uses 30mΩ RDS(on) and 70mΩ RDS(on) high voltage GaN power switches in the TOLL package for the primary side, and for the secondary, four pairs of 3.2mΩ RDS(on) low voltage parts in a flip chip QFN package. The resulting unit measures just 185 x 69 x 37mm and achieves an astonishing power density of 130 watts per cubic inch, peaking at 97.5% efficiency, and easily exceeding the bar set by the 80 Plus titanium standard. To put it into context, an equivalent silicon MOSFET PSU would be twice the size.
Figure 1: By specifying GaN in both the primary and secondary PSU circuits, astonishing power density specs can be achieved.
Other examples of the use of multiple GaN ICs include: LED drivers where the increased power density of GaN driver can fit inside the track light fixture; and a GaN-based 1000W DC/DC converter that is 70% smaller and more efficient than a 600W silicon unit. Inverters, motor drivers and photo-voltaic systems can benefit similarly.
Integrated GaN solutions
Initially, GaN was largely only available as discrete GaN transistors. Whilst discretes offer great design flexibility, increasingly integrated solutions are also being offered. These simplify the design process and can reduce component count and design size. Figure 2 contrasts the two different approaches.
Figure 2: Discrete v Integrated GaN methodology
Innoscience, for example, is introducing a range of SolidGaN™ integrated solutions that combine power transistor, driver, current sense and other functions within a single, industry-standard QFN 6x8mm package. The 700V ISG610x devices cover the range from 140 to 450mohm, and save PCB space and BOM count, while increasing efficiency and simplifying design for applications including USB-PD chargers, LED lighting, AC/DC and DC/DC power supplies and PFC, QR flyback, ACF, half bridge, and full bridge circuits. The integrated devices feature a wide 9-80V VCC range which is beneficial in USB-PD applications that require up to 28V output.
Figure 3: Integrated GaN ICs from Innoscience are easy to use.
For low power operation, ISG610x family ICs also feature a low, 115µA quiescent current, thanks to an innovative automatic standby mode which is activated when the PWM signal voltage remains below VPWM_LO for a certain time period. During this time, most of the internal circuitry is turned off, dramatically reducing energy wastage, enabling devices to meet the No-Load, Low-Load specifications of regulatory bodies such as ENERGYSTAR.
The loss-less current sensing with 7% accuracy of the new SolidGaN devices offers several benefits. Firstly, because the current sensing resistor loss is eliminated, a larger RDS(on) can be accommodated with no loss in performance, leading to cost reduction. Secondly component count is reduced and PCB footprint is minimized. Devices also feature a programmable switch turn-on slew rate to enable EMI to be minimised. An internal linear voltage regulator is included to assure a 6.5V supply, maximizing power efficiency while ensuring the reliability of the GaN HEMT. Finally, built-In under-voltage lock-out (UVLO), over-current (OCP) and over-temperature (OTP) protection are incorporated within the IC.
Dispelling the price and reliability myths
As we can see, integrated solutions offer many benefits, yet discrete GaN solutions still provide the ultimate in design flexibility and may be the only available option at higher powers until further paralleling of integrated devices is achieved – which is in the roadmap. But whatever your design requirements are, GaN – in some format – is probably the answer. Why then is there still a reluctance amongst certain sectors of the electronics industry to make the move to GaN?
Unfortunately, some decision-makers are working with out-dated information concerning price and reliability. Let’s address those.
Figure 4: Comparing the price of silicon super-junction devices (from various vendors) and Innoscience’s InnoGaN for similar current and voltage rating. Public catalogue prices sourced from www.digikey.com [2] and www.richardsonrfpd.com [3].
Initially, of course, GaN devices did carry a price premium. Partly, that was because they were being produced in relatively low numbers (compared to silicon MOSFETs), but also, companies need to recover their investment costs, and some traditional silicon companies have standard silicon device business to protect. However, intrinsically, transistors made using a GaN-on-silicon construction should be no more expensive to produce than standard silicon parts.
Figure 4 shows that, based on publicly available data, Innoscience’s InnoGaN GaN HEMTs are now price competitive with silicon super junction devices. Furthermore, GaN offers the possibility to save further costs at a system level because the increased efficiency will lead to a reduction in size – and therefore price – of the passives and magnetic components that will be required.
Innoscience has achieved price parity by building the world’s largest 8in GaN-on-Si fabs. By controlling all key manufacturing processes, including epitaxy, in house, the company achieves consistently high yields. It can also leverage massive economies of scale.
Turning to the second question, it is easy to point to the millions of years of operational life data that silicon has amassed and ask questions about the reliability of GaN. But while GaN cannot yet provide 50+ years of accumulated test results, GaN has been around for over 20 years and is well studied and understood. Devices are now tested to the international JEDEC standard and to specific JEDEC guidelines drafted for wideband gap devices (JEP 180) where devices are stressed under switching stress to mimic real application usage. Innoscience performs further extrapolated lifetime tests including HTGB (beyond max gate specs) and HTRB (beyond max off-state drain voltage specs).
Conclusion
GaN has proven its performance benefits in simple, low power USB-PD charger applications, where it is set to become the dominant technology. As GaN is able to compete successfully in such price-sensitive markets, issues relating to its commercial viability in other higher value applications should be easily answered. Therefore GaN is poised to transition into other power electronics markets and companies are launching a plethora of integrated GaN-powered solutions targeted at an ever-increasing range of applications. Therefore, power conversions and power management solutions can be made smaller, lighter, more efficient and simpler (less BOM). Devices are proven to be reliable, and price does not need to be a barrier.
References:
[1] www.nextmsc.com/report/gan-charger-market
[2] www.digikey.com
