+44 (0)24 7671 8970
More publications     •     Advertise with us     •     Contact us
 
Technical Insight

SiC and GaN electronics: Where, when and how big?

Switching from silicon electronics to wide bandgap alternatives promises to deliver improvements in the performance of power supplies, wind turbines, solar systems, hybrid electric vehicles, trams, trains and industrial machinery. Richard Eden from IMS Research, which was recently acquired by IHS, takes a detailed looked at all of these opportunities for wide bandgap devices and identifies the barrier to their mass adoption.



Silicon’s vice-like grip of the power electronics market in coming under increasing pressure from a pair of wide bandgap semiconductors. One of these is SiC, a material that has been touted for use in power electronics for some 30 years. And the other is GaN, a semiconductor that has recently received tremendous investment and is most famous for its widespread use in blue and white LEDs in high-volume consumer applications. The SiC Schottky barrier diode, the first SiC power device brought to market, made its debut in 2001. Back then it was hoped that SiC would quickly become a suitable semiconductor material for high-efficiency power electronics; however, development has been delayed by difficulties associated with the manufacture of SiC substrates. Its rival, GaN, is now arguably the best semiconductor material, in terms of the combination of performance and cost. It is used today for making RF power amplifiers, thanks to its inherent advantages in voltage and temperature performance over GaAs. Strengths of this wide bandgap semiconductor include the promise of making devices with incredibly low loss, and the opportunity to deposit epilayers on standard silicon substrates. The latter virtue enables production costs to be significantly below those for SiC.

At IMS Research, a market research firm based in the UK, we have been looking at various applications for SiC and GaN power semiconductors, and figuring out how they will evolve over the next ten years. In the remainder of this article we’ll look at the impact of this pair of wide bandgap devices on a wide range of applications: Power factor correction (PFC) power supplies, uninterruptable power supplies (UPS), hybrid and electric vehicles, industrial motor drives, PV inverters, wind turbines, and traction.

For SiC and GaN devices, many of the barriers to mass adoption are similar. Neither of these devices will make a big impact until their prices approach those of silicon SuperJunction MOSFETs and high-voltage IGBTs. This affordability target will be pursued through moves to manufacture on larger substrates and improve material quality.

Other factors that will determine how quickly SiC and GaN devices can make an impact are: The time taken for SiC and GaN transistors and power modules to demonstrate long-term reliability in long-life industrial or health-and-safety-critical applications; how quickly the strength of competition increases between the qualified suppliers of SiC wafers and GaN-on-silicon epitaxial wafers; and whether there will be new legislation or government initiatives on the energy efficiency of power electronic systems. Greater competition from increasing numbers of established power semiconductor manufacturers offering SiC and/or GaN products will also help to drive greater adoption of wide bandgap devices, which can sell in higher volumes as their portfolios broaden. The SiC product range should grow to include higher-voltage devices operating at 1700 V, 2.5 kV, 3.3 kV, 4.5 kV and above, plus GaN diodes and transistors at 1200 V and above.

Power supplies

PFC power supplies were the first key application to use both forms of wide bandgap power device. In this sector, the long-term winner is likely to be GaN. 600 V and 650 V will be the typical device voltage ratings for power supplies in consumer and office equipment running off of the mains, and we forecast that GaN transistors and diodes operating in this regime will cost less than SiC-based equivalents – they could ultimately match silicon prices. Industrial three-phase mains power supplies sell in smaller quantities than PFC supplies, and they require 1200 V-rated devices, a requirement that is easier to satisfy with SiC technology.

SiC and GaN switches offer substantial benefits in PFC stages of hard-switching power supply units (PSUs); less so in soft-switching circuits. The increase in power conversion efficiency is actually quite small – perhaps tenths-of-a-percent to 1 percent. The far bigger gain resulting from the introduction of wide bandgap electronics in a substantial cut in the size of the PSU, with increased power density made possible by fewer, smaller secondary components required in the snubber circuit. The reason behind this is the opportunity to increase switching frequency, which is made possible with SiC and GaN diodes.

Few vendors of the other big selling class of power supplies, the UPS, are building products incorporating SiC or GaN power devices. However, the relatively small market for SiC Schottky diodes in UPS is no reflection of the improvement wrought with wide bandgap semiconductors. In our view, in UPS, the revenue potential for SiC or GaN Schottky diodes is relatively small. Sales should be greater for SiC power transistors and SiC hybrid and full-power modules. We predict that discrete SiC and GaN power devices will be used in just smaller single-phase UPS systems rated up to 5.1 kVA, while SiC power modules will be deployed in UPS systems rated between 5.1 kVA to 250 kVA.

Better electronics for better cars

Exciting opportunities for wide bandgap power devices also exist in the hybrid and electric vehicle market. Adoption of the devices could increase inverter efficiency, leading to a longer vehicle range between recharges.

There are three potential areas for the deployment of wide bandgap devices in hybrid and electric vehicles: The mains battery charger, found only on plug-in hybrids and battery electric vehicles; DC-DC voltage conversion systems; and the drivetrain. Increasing the power density in these electronic systems, which is possible with a switch from silicon to wide bandgap semiconductors, allows for reductions in size and weight. This is highly desirable because space in an electric car is tight, and using less of it for electronic systems gives designers more freedom. In addition, reducing weight increases the vehicle’s acceleration and range. 



Inverters are needed to convert the DC output from the cells into an AC form that can be fed into the grid. Increases in efficiency that lead to better returns are possible by replacing the silicon electronics in the inverter with SiC devices.

Today’s HEVs are fitted with silicon power electronics, and use two separate cooling systems. That’s because the inverter mounted on or near the internal combustion engine block must be maintained at a lower temperature than the engine. When SiC devices start to take the place of silicon IGBTs that have inferior thermal capabilities, it may allow the inverter to run at the same temperature as the engine casing and consequently use the same cooling system.

Unfortunately, none of the current SiC switch technologies are sufficiently mature to match the reliability of silicon devices or even of SiC Schottky diodes. This will delay adoption of SiC transistors in HEVs. SiC JFETs should meet the reliability requirements first, because they have the same basic structure and materials as diodes. But these transistors may not dominate the market, because there could be significant competition from SiC BJTs, which seem to offer good reliability in terms of life test, high-temperature operation and temperature cycling, plus robustness to shocks and vibrations.

In the long term, GaN transistors and diodes are of great interest for HEVs, because they could eventually undercut the price of equivalent products built from

SiC. Discrete GaN devices will probably only find deployment in the battery charger, but it is possible that they might also be used in drivetrain power modules.

As with SiC devices, adoption of GaN will not take place until devices have proven reliability to automotive specifications.

Industrial motor drives

The greatest revenue opportunity for SiC power devices, particularly for power modules, is in the industrial motor drive market. Here, SiC can be used in equipment for driving and controlling industrial motors within factory automation systems. This equipment can be single-phase or three-phase; drive AC, DC or servo motors; and operate at variable speed or variable frequency. The main benefit of this equipment is improved control and efficiency when driving motors.

The increase in power density resulting from the switch from silicon devices to those built with SiC trims the inverter’s size, its weight and its cooling requirements. However, this market is extremely cost-sensitive, with drives sold on price versus performance and size. To win sales, the prices of SiC devices must plummet to a level comparable with silicon equivalents.

Power conversion efficiency values will become increasingly important. The market for low-voltage AC and DC drives should benefit from European Union legislation that has established minimum motor efficiency requirements as a result of the Eco-design of Energy-related Products Directive. Beginning in 2015, a mandate that has led from this directive has dictated the use of either an IE3 efficiency motor (which operates at ‘premium efficiency’), or an IE2 efficiency (which  operates at ‘high efficiency) motor coupled with a drive. Most modern motors are IE2 efficient, and we believe that many customers will take the cheaper option of adding a drive to an existing IE2 motor, rather than replacing it with a new IE3 motor. This should lead to ramping sales in the drives market.

Designers of systems with long operational life, such as industrial motor drives, only want devices with excellent reliability. These are some unanswered questions concerning the reliability of SiC devices, and diodes and transistors will only be qualified after extensive load-cycling, temperature-cycling and life-test data has been produced and evaluated. However, we predict that this scrutiny will not reveal any major weaknesses, and that there will come a time when discrete SiC and GaN power devices are used in single-phase drives, and SiC power modules in three-phase, industrial motor drives.

There are two significant, far from obvious benefits that result from the use of SiC or GaN devices in industrial motor drives: Higher switching frequencies are possible, leading to quieter motors; and there is an opportunity to modify the inverter circuit topology to incorporate a transformer. Go down the latter route and it is possible to generate sine-wave voltage and current outputs, which can drastically diminish RF electromagnetic compatibility noise.

Boosting PV efficiencies

Owners of all forms of photovoltaic system want to generate as much energy as possible. Increasing the efficiency of the inverter helps, which is possible by replacing silicon devices with those made from SiC or GaN. Even an improvement of just 0.5 percent is viewed as highly worthwhile, because it allows the end-user to make more money selling the electricity to power companies.

Consequently, it is of no surprise that the PV inverter market is the second largest one for SiC Schottky diodes and it will be the first one to start using SiC transistors in volume. An inverter built with SiC diodes and transistors can operate at a switching frequency that is two-to-three times that of a silicon-based equivalent, and potentially be half the size.

Despite their relatively high price, 600 V SiC Schottky diodes should win deployment in micro-inverters and single-phase string inverters. However, we believe that those devices will subsequently be displaced by lower-priced GaN Schottky diodes, which will make inroads once their reliability is proven.

Moving on to transistors, our view is that those made from GaN could eventually replace high-voltage MOSFETs in 600 V systems for string inverters and micro-inverters operating below 5 kW. SiC transistors will also enjoy success, and are likely to replace silicon IGBTs in 900 V or 1200 V single-phase string inverters operating above 5 kW. In 1200 V three-phase central inverters, we expect to see the introduction of SiC power modules. 



Industrial motor drives, such as those found on large wood-working machines, offer the greatest revenue opportunity for SiC power devices.

Making best use of the wind

Wind turbines are another form of renewables that can benefit from wide bandgap power electronics. Experiments reveal that replacing silicon IGBT-based power modules with SiC equivalents delivers an improvement in overall system efficiency of approximately 1 percent. This efficiency boost translates to an operation cost benefit of $1.50 per hour per MW for the owners of a turbine generating electricity that can be sold for $150 per MW-hour. The switch to SiC also enables a hike in power density, leading to the design of smaller power converter systems.

There are several significant barriers preventing the adoption of SiC modules: High cost, reliability concerns, and the lack of availability of products with high current ratings. Today, SiC power devices are limited to a maximum current rating of 30 A, far less than the 100-150 A rating for silicon IGBTs deployed in wind turbines. 

Note that it is also unlikely that discrete SiC power semiconductors will be used in wind turbines, because one would need dozens of discretes to switch the required power. The preference will be for power modules, which integrate six or more devices in a housing for heat-sinking, possibly with protection and control circuitry included.

In virtually all cases, wind turbines are built with 1700 V power modules, which are designed to work with 690 V, three-phase mains voltages. These turbines are being built with longer and longer blades, driving up output power. This is putting pressure on the makers of power modules to increase the handling current of their products, or reduce prices. We believe that this trend is hampering the uptake of SiC modules and depressing their penetration in this application.

With wind turbines, like many other applications, device reliability is crucial. The location of the wind turbine determines the consequences of the failure of an individual module, which stops a wind turbine power converter from operating. If the turbine is on land, an electrician can make the necessary repairs in a day.

But if the turbine is off-shore, specialist engineers must get on a service boat, which can only reach the required destination in good weather and never during the winter months.

Off-shore turbines tend to be larger, causing failures that lead to losses in average annual generated-electricity revenues roughly a thousand times higher than those of on-shore installations.

Aiding transportation

In traction applications three-phase 690 V mains is standard, so most power modules are typically rated at 1700 V. However, some are rated higher – the maximum is 6.5 kV. Two types of inverter tend to be used in electric traction vehicles; a main inverter for the propulsion motor drive and an auxiliary inverter for battery charging. In a two or three car tram, six power modules are needed in an inverter; ten times this are required in a locomotive. Current ratings for power modules used in motor propulsion are between 1200 A and 2400 A, which significantly exceeds the capabilities of any of today’s SiC power modules.

The batteries that are fitted to trains, power the control systems, the brakes, lighting, heating, ventilation and all the electrical sockets in the carriages. The battery charger circuit normally uses fewer modules than the propulsion system, because current and power ratings are lower. For this reason, battery-charging inverters will lead the way with the uptake of SiC power modules.

Switching from silicon to SiC produces a 1 percent increase in power-conversion efficiency. This gain can lead to a trimming of energy consumption or an increase in speed. Thanks to the wide operating temperature capability of SiC, it is also possible to reduce the size of the cooling systems. And on top of this benefit, SiC enables the introduction of higher switching frequencies, a step that can lead to the use of smaller inductive components. The upshot of all these strengths is a lighter train that requires less power to drive, particularly when moving off from stationary. This benefit is more significant than the trimming of inverter size or volume, because space is not a big issue in traction locomotives. Arguably, audible motor noise is more of a concern, and this could fall through an increase is switching frequencies.

If SiC devices are going to make an impact, prices must fall, concerns regarding reliability must be banished, and modules with high current ratings, particular those based fully on SiC, must become available. High prices are the biggest barriers, and it will take several years before they fall to the level necessary for widespread adoption.

Other opportunities

There are many other opportunities for wide bandgap technologies.  High-voltage SiC devices could find deployment in the military and aerospace industry; in drilling and mining; in medical equipment; and other industrial applications. Meanwhile, for low-voltage GaN devices, applications include: The secondary side of high-end switch-mode power supplies for telecom and datacom networks and servers; DC-DC conversion, including point-of-load; power over Ethernet; and many emerging technologies that will drive significant growth in the future, such as wireless charging.

Our view is that the greatest potential for SiC power devices is in military and aerospace applications. In aviation, the SiC devices can be used for power distribution within the airframe; the low-voltage side is handled by MOSFETs, but 1200 V SiC MOSFETs are attractive candidates for replacing contactors in the high-voltage stage. Key benefits are increased power density – there is a five-fold-to-ten-fold increase in switching frequency, so size and weight diminish – and the opportunity to operate devices at higher temperatures. The latter benefit reduces cooling/heatsink requirements, and also saves weight.

GaN transistors will make their initial inroads in applications where they can replace low-voltage silicon MOSFETs and yield an increase in power conversion efficiency. Such an opportunity exists in a wide range of power supplies: DC-DC power supply units, such as power-over-Ethernet equipment, network power supplies, point-of-load converters, 48 V telecom power supply units, server-farm 12 V or 48 V power rails, medical equipment, and various forms of lighting, including that based on LEDs.

Other opportunities are also there for GaN. When low-voltage devices made from this wide bandgap semiconductor match those of silicon MOSFETs, GaN transistors will be adopted in the secondary side of switch-mode power supplies in high volume applications like consumer/domestic PCs, notebooks and tablets. In addition, 600 V GaN transistors will be adopted in other applications, such as electric bicycles and domestic appliance drives and inverters. What’s more, GaN is being investigated for space and military applications, because the radiation hardness of this material is inherently superior to that of silicon.

And it is anticipated that GaN power devices will be a disruptive technology that will allow commercialisation of inventions struggling to succeed with silicon devices.

Examples include wireless charging, digital power conversion and RF envelope tracking.

Grabbing market share

Due to the great potential of GaN and SiC power devices in a vast number of applications, sales of these products are outpacing the growth of the total market for power devices. Although wide bandgap sales made up less than 1 percent of total power semiconductor revenues last year, if we assume that materials maintain their trajectory of falling cost and devices demonstrate long-term reliability, we can expect a market worth over $3 billion by 2021 – that’s approaching 9 percent of the total market for power semiconductors.

In other words, in the next ten years, penetration of these devices in the power market will increase ten-fold.
×
Search the news archive

To close this popup you can press escape or click the close icon.
×
Logo
×
Register - Step 1

You may choose to subscribe to the Compound Semiconductor Magazine, the Compound Semiconductor Newsletter, or both. You may also request additional information if required, before submitting your application.


Please subscribe me to:

 

You chose the industry type of "Other"

Please enter the industry that you work in:
Please enter the industry that you work in: