Version 4.3 April 2021 - poweramericainstitute.org

i

Version 4.3

April 2021

ii

Table of Contents

Acknowledgements ............................................................................................................................ iii

About This Roadmap ............................................................................................................................ 1

Executive Summary ............................................................................................................................. 2

Background/Introduction ................................................................................................................. 3

Market Forecast ............................................................................................................................... 4

Current SiC and GaN Landscape ........................................................................................................... 5

Key Markets and Applications .......................................................................................................... 5

Applications with Highest Near-Term Priority ............................................................................... 6

Applications with Highest Longer-Term Priority ............................................................................ 9

Wide Bandgap Power Electronics Pricing Comparison .................................................................... 10

Summary .................................................................................................................................... 13

Device Bank ................................................................................................................................... 14

PowerAmerica’s 5-Year Roadmap Strategy ........................................................................................ 15

Thrust 1: Reducing Cost .................................................................................................................. 16

Key Challenges ........................................................................................................................... 16

Key Activities .............................................................................................................................. 17

Thrust 2: Improving Reliability and Quality ..................................................................................... 18

Key Challenges ........................................................................................................................... 18

Key Activities .............................................................................................................................. 20

Thrust 3: Enhancing Performance Capabilities ................................................................................ 20

Key Challenges ........................................................................................................................... 20

Key Activities .............................................................................................................................. 22

Thrust 4 Strengthening the Power Electronics Ecosystem ............................................................... 24

Key Challenges ........................................................................................................................... 24

Key Activities .............................................................................................................................. 25

Path Forward ..................................................................................................................................... 28

Appendix A. Acronyms ....................................................................................................................... 29

Appendix B. PowerAmerica Device Bank ............................................................................................ 31

Appendix C. Priority Concepts to Advance SiC and GaN Technology………………………………………………… 33

iii

Acknowledgements This roadmap was developed under the direction of PowerAmerica Executive Director and Chief Technology Officer (CTO) Victor Veliadis, Program Manager Rogelio Sullivan, and Membership and Industry Relations Director Jim LeMunyon. Roadmap development and subsequent updates were informed by a variety of stakeholders, including experts from the semiconductor and power electronics industries (especially from PowerAmerica member organizations), academia, national laboratories, and other research organizations. These stakeholders made vital contributions through workshop participation and subsequent working group support, interviews, online surveys, and roadmap reviews. PowerAmerica is indebted to those who contributed their time and expertise in developing and revising this roadmap.

1

About This Roadmap

In December 2014, the PowerAmerica Institute was established through Manufacturing USA—a network of public-private partnerships committed to increasing U.S. manufacturing competitiveness. Led by North Carolina State University in Raleigh, NC, PowerAmerica is a consortium of 50 companies, universities, and federal labs, which aims to accelerate the adoption of wide bandgap (WBG) semiconductor power electronics (PE). By improving technical capabilities, supporting domestic manufacturing, and strengthening the WBG semiconductor ecosystem, PowerAmerica efforts are expected to contribute to energy savings, new jobs creation, and a strengthened U.S. manufacturing sector.

PowerAmerica focuses specifically on advancing silicon carbide (SiC) and gallium nitride (GaN)—both WBG semiconductors—which offer improved performance across a range of applications. PowerAmerica’s member organizations help drive progress and facilitate collaboration across the PE community, including between end users and experts from prominent universities and government agencies. The institute also receives support from the U.S. Department of Energy’s Advanced Manufacturing Office (AMO) and the state of North Carolina, as well as investments from industry, academia, and other partners.

To ensure the Institute’s investments and activities best meet the industry’s current needs and anticipated challenges, PowerAmerica solicits input from industry experts, especially from PowerAmerica members. PowerAmerica’s roadmapping process began in earnest in mid-2016 with the convening of in-person and virtual workshops with members and outside experts. Nexight Group, a consulting company supporting PowerAmerica’s roadmapping efforts, also conducted phone interviews with key experts, distributed an online survey to gather additional input, and performed a literature review of relevant resources in this field. This resulted in Version 1.0 of the Roadmap. PowerAmerica members and outside experts also contributed to subsequent updates: Version 2.0 in February 2018, Version 3.0 (interim roadmap update) in July 2018, and Versions 4.0 in December 2018, 4.1 In February 2019 and this Version 4.2 in April 2020.

Version 4.0 incorporated updated feedback from the August 2018 roadmapping workshop, where stakeholders discussed the current state of WBG semiconductor technology, ongoing challenges, and key activities to address those challenges. It also included a new market forecast for WBG power electronics, an analysis of SiC and GaN device cost based on market information, and information on PowerAmerica’s Device Bank. Results from the mid-2018 survey of members regarding the challenges facing the adoption of WBG semiconductor technology were also incorporated into this version. Versions 4.1 and 4.2 are updates that include several suggestions made by PowerAmerica members during a review of the Roadmap at the PowerAmerica Annual Meeting in February 2019 and Summer Workshop in August 2019. These suggestions were refined and subsequently approved by the Member Advisory Committee.

This roadmap outlines key markets and application areas as well as the performance targets GaN and SiC technologies are expected to meet over time, technical barriers to achieving those targets, and activities needed to overcome those barriers. Roadmapping is an ongoing process that guides PowerAmerica’s strategic decisions and provides a common vision of the future for the WBG community to work toward.

Information on important acronyms and the PowerAmerica Device Bank is provided in the Appendices.

2

guides PowerAmerica’s strategic decisions and provides a common vision of the future for the WBG community to work toward.

Information on contributors, important acronyms, survey results, and the PowerAmerica Device Bank is provided in the Appendices.

Executive Summary

This roadmap outlines key markets and application areas for SiC and GaN PE, performance targets for competitive SiC and GaN technologies, technical barriers to achieving those targets, and the PowerAmerica activities needed to overcome those barriers. PowerAmerica’s role will be to facilitate coordination across industry, academia, and national labs to implement the priority activities identified in this roadmap and to make strategic investments in technology development, workforce training, and WBG manufacturing. The following high-level recommendations are a summary of the actions found in Section 5: PowerAmerica’s 5-Year Roadmap Strategy.

Reducing Cost • Lower the $/mOhm of WBG devices and power modules. • Support vertically integrated fabrication. • Support and promote early adopter, high-volume WBG applications. • Establish SiC and GaN open foundries to scale to high-volume manufacturing.

Improving Reliability and Quality • Establish WBG PE reliability at system-level and investigate degradation/failure mechanisms of

devices, modules, or systems. • Develop open databases for reliability data. • Develop capability to perform AECQ or JEDEC standard tests for WBG power devices. • Set dedicated standards for WBG PE.

Enhancing Performance Capabilities • Focus on near-term applications to demonstrate the system-level advantages of WBG power

devices. • Support pathways to commercialization for industry-led projects. • Promote reference designs, advanced gate drives and modules, and work in advanced peripherals.

Strengthening the Power Electronics Ecosystem • Continue to offer the Device Bank for quick access to SiC and GaN devices. • Continue to provide communication mechanisms for different levels of stakeholders, from vendors

to end users. • Train a WBG PE workforce. • Monitor basic core technologies, state-of-the-art complementary technologies, and long-term

applications to identify promising opportunities.

3

The Need for Advanced Wide Bandgap Power Electronics

Background/Introduction Our society has become increasingly dependent on complex devices, machines, and systems—from handheld electronic devices like smartphones and laptop computers to electric vehicles (EVs) and grid-scale renewable energy systems. None of these technologies would be possible without cross-functional semiconductor power electronics (PE) capable of converting power and controlling electrical energy (i.e., tuning voltage, current, and frequency) from the point of energy generation to distribution.

WBG semiconductors hold great promise to significantly outperform and eventually replace traditional Si-based PE technology. While there are research and development (R&D) efforts in various WBG semiconductors—including diamond, aluminum nitride, and gallium oxide—that could be used in advanced PE, SiC and GaN have currently reached a level of maturity that allows use in PE applications. SiC and GaN have enabled the development of compact (i.e., high power density), cost-effective, energy-efficient, and robust power components that operate at higher temperature, voltage, and frequency conditions.

Both SiC- and GaN-based power devices have distinct benefits for specific applications: SiC is generally a stronger candidate for PE above 1.2kV, while GaN is highly competitive for PE below 1.2kV. The device voltage range between 650V and 1.2kV is a competitive space that can be supported by either SiC or GaN technologies. Compared to Si, SiC-based power devices can operate at higher temperatures with higher thermal conductivity, higher breakdown voltage at lower on-stage resistance, faster switching speed, lower conduction and switching on-state loss, and exceptional radiation hardness. Advantages of GaN-based power devices include higher electron mobility and lower losses at higher frequencies, which can enable smaller devices with increased power density.

While WBG technologies offer significant capabilities that can advance PE, industry must overcome numerous challenges including high material and manufacturing costs, reliability perceptions, packaging and performance requirements, and difficulty coordinating efforts across the entire WBG PE ecosystem. Recent progress against these challenges in automotive applications, PV inverters, and power supplies is encouraging; however, SiC and GaN have not taken off as rapidly in traction applications, industrial motor drives, and wind turbines. Further strides are needed to begin manufacturing these devices at high volumes and competitive costs across the full range of useful applications.

Semiconductor Materials and their Bandgap Energies (Eg)

WBG Semiconductors: • Silicon carbide (SiC): 3.3eV • Gallium nitride (GaN): 3.4eV • Zinc oxide (ZnO): 3.4eV • Gallium oxide (β-Ga2O3): 4.8–4.9eV • Diamond (C): 5.5eV • Aluminum Nitride (AlN): 6.0eV

Conventional Semiconductors: • Silicon (Si): 1.1eV • Germanium (Ge): 0.7eV • Gallium Arsenide (GaAs): 1.4eV

4

Market Forecast Prices for SiC and GaN devices have been falling rapidly in the last few years, helping fuel recent market growth. SiC metal-oxide-semiconductor field-effect transistor (MOSFET) prices, for example, dropped 50% between 2012 and 2015 (according to IHS Markit). Though prices rose in 2017 due to wafer supply shortages, a growing number of wafer suppliers and improved wafer performance should allow prices to stabilize. This increasing cost competitiveness has already helped SiC begin to dislodge Si in some applications (e.g., hybrid vehicles) and has enabled mass production of GaN-based end products (mainly in server and telecom rectifier power supplies). In addition, leading manufacturers now have trillions of hours of field device experience to assuage any reliability concerns that might dampen growth.

A presentation by IHS Markit at PowerAmerica’s 2019 Summer Workshop projected annual, global SiC revenues will reach more than $5 billion by 2027, with hybrid and electric vehicles making up most sales. Annual GaN revenues are projected to top $1.4 billion over the same timeframe, with power supplies, hybrid and electric vehicles, and military and aerospace applications holding the largest shares. In comparison, revenues for SiC and GaN combined were only $210 million1 in 2015. Across these applications, discrete power devices would account for most of the growth as they are expected to take off faster than power modules and integrated circuits.

A separate forecast from Cree Inc.2 also predicts EVs will present a tremendous growth opportunity for WBG materials, particularly SiC. To date, automakers have announced plans to spend $150 billion in the EV market.3 Cree estimates that even modest EV adoption—approximately 10% of total vehicles sales by 2027—could result in SiC revenues of $6 billion. The same forecast places the total SiC PE market at over $5 billion by 2022, largely driven by EV adoption but also industrial and telecom applications. For GaN, telecommunications stand out as an opportunity for strong growth. GaN devices support 10 times faster download speeds and better cellular coverage, which can enable the transition to 5G internet service.

The future is bright for WBG PE technologies. SiC and GaN have already proven their technical advantages over Si, and now decreasing prices are also driving adoption. SiC turned a corner in 2016 and GaN growth should shortly follow. For devices within certain voltage ranges, SiC and GaN will be viable options within the next 10 years and should continue to displace Si in the market.

1Richard Eden, “Market for GaN and SiC power semiconductors to top $10 billion in 2027,” https://technology.ihs. com/602187/market-for-gan-and-sic-power-semiconductors-to-top-10-billion-in-2027 (April 24, 2018) 2 Cengiz Balkas, “Wolfspeed,” https://www.sec.gov/Archives/edgar/data/895419/000089541918000019/ analystdayfebruary262018.htm (February 26, 2018) 3 Nic Lutsey et al., “Power Play: How Governments Are Spurring the Electric Vehicle Industry,” https://www. theicct.org/sites/default/files/publications/EV_Government_WhitePaper_20180514.pdf (May 2018)

https://technology.ihs.com/602187/market-for-gan-and-sic-power-semiconductors-to-top-10-billion-in-2027

https://technology.ihs.com/602187/market-for-gan-and-sic-power-semiconductors-to-top-10-billion-in-2027

https://www.sec.gov/Archives/edgar/data/895419/000089541918000019/analystdayfebruary262018.htm

https://www.sec.gov/Archives/edgar/data/895419/000089541918000019/analystdayfebruary262018.htm

https://www.theicct.org/sites/default/files/publications/EV_Government_WhitePaper_20180514.pdf

https://www.theicct.org/sites/default/files/publications/EV_Government_WhitePaper_20180514.pdf

5

Current SiC and GaN Landscape

Key Markets and Applications To accelerate widespread adoption of WBG semiconductor components, it is critical to support early commercialization success for applications in high-value markets. PowerAmerica’s commercialization strategy starts by identifying a market and deployment timeline, as well as assessing the competition, barriers to market penetration, and the impact on U.S. competitiveness. This section describes and prioritizes the main PE applications for SiC and GaN technologies. Tables of performance targets, as identified by PowerAmerica roadmapping workshop participants and working group members, accompany the application descriptions.

Figures 1 and 2 show various WBG device applications and map their priority level (vertical axis), voltage range (horizontal axis), and timeframe for commercial viability (color). Prioritization is based on two factors: system impact (i.e., potential to improve efficiency, power density, and cost) and broader economic impact (i.e., expected market size and likelihood of adoption). Therefore, an application with a high value is one that PowerAmerica should focus on immediately.

Figure 1. Primary Markets and Applications for SiC-based PE Devices. Colors distinguish timeframes for commercial viability.

6

Figure 2. Primary Markets and Applications for GaN-based PE Devices

Figure 2. Primary Markets and Applications for GaN-based PE Devices. Colors distinguish timeframes for commercial viability

Applications with Highest Near-Term Priority To accelerate adoption of WBG semiconductors, the PE community must focus on applications for both SiC and GaN technologies that are most likely to deliver immediate (within 5 years) improvements in efficiency, reliability, and total cost of ownership. These in turn will encourage further market growth and industry demand. As listed below, the key near-term applications with high priority values are EV inverters (SiC), PV inverters (SiC), enterprise equipment (GaN), data centers (SiC/GaN), and industrial AC/DC power supplies (GaN). Table 1 shows performance targets for some of these near-term priority applications.

EV Inverters (SiC): According to the Electric Vehicles Initiative (EVI), over 20 million electric cars may be on the road in 2020, climbing to over 200 million by 2030 in the most ambitious scenario.4 Motor drive inverters with high power density and efficiency at elevated temperatures are essential to EVs. WBG semiconductors are ideal for such applications because of their advantages over Si in high temperature and frequency applications. Lower system costs and vehicle design simplification will improve PE devices integration with vehicles. The cost competitiveness of WBG semiconductors is the main hurdle for EV/HEV inverter applications.

4 Clean Energy Ministerial, “EV30@30 Campaign,” http://www.cleanenergyministerial.org/sites/default/files/2018-07/EV30%4030%20fact%20sheet%20%28May%202018%29.pdf (July 2018)

http://www.cleanenergyministerial.org/sites/default/files/2018-07/EV30%4030%20fact%20sheet%20%28May%202018%29.pdf

http://www.cleanenergyministerial.org/sites/default/files/2018-07/EV30%4030%20fact%20sheet%20%28May%202018%29.pdf

7

PV Inverters (SiC): PE are essential components of renewable energy power conversion, particularly the high-efficiency DC-to-AC conversion needed for photovoltaic (PV) energy generated at utility-scale solar farms. WBG-based PV inverters and PV systems convert power more efficiently, yielding significant energy savings. Along with PV systems, the PE community is also looking into WBG to increase the efficiency, power density, reliability, and other requirements (e.g., portability and EMI specification) of power inverters for wind, geothermal, and other renewable energy systems. Ease of installation and low maintenance costs across the entire grid are also critical considerations that will impact adoption.

Enterprise Equipment (GaN): GaN-based enterprise equipment has evolved rapidly in recent years. A wide variety of applications for wireless power transfer systems—ideal for fast charging laptops or smart phones using GaN power amplifiers and E-HEMT transistors—have emerged using GaN semiconductors. Other examples include GaN-on-Si for switches, routers, servers, and data center power converters. The benefits of these GaN applications include improved power quality (or reduced power loss) in electric transmission and distribution that are closely related to smart grids and renewable energy applications (e.g., wind and solar power systems).

Data Centers (SiC and GaN): WBG semiconductors play a major role in mitigating the growing energy consumption (and environmental impact) of data centers, which consume approximately one percent of all electric power produced in the United States.5 Combined with system-level improvements to power architecture, WBG power devices will dramatically increase power delivery efficiency and simplify the design of data center power systems. The simplified design of DC-powered data centers compared to traditional AC-powered centers provides lower energy conversion losses, higher reliability, and smaller equipment footprints for power conversion and cooling equipment. Key enablers for success in this near-term priority include high power density, low switching losses, and high-temperature operation, which ultimately increase system efficiency.

Industrial AC/DC Power Supplies (SiC and GaN): Power supplies are ubiquitous in electronics because all require converting power inputs to the required outputs. Power supplies using SiC and GaN are now widespread in the PE market due to several key factors. These include high power efficiency (e.g., >96%) related to fast switching speed, high power density/smaller size, good reliability, and potential for reduced costs.

Heavy-duty Vehicle (SiC): Integrating WBG PE in heavy-duty vehicle systems will be challenging because integration must be achieved with reducing system-level cost and complexity and improved reliability. The reduction in passive components, enabled by WBG devices, is particularly important in these applications.

Additional near-term applications were identified by PowerAmerica members in 2019, and several “longer-term” applciations were moved to “near-term.” The related performance characteristics needed for some of these applications will be provided in the next update to this Technlology Roadmap. These applications are: aerospace and defense (GaN and SiC), vehicle on-board charger/DC and off-board fast charging (SiC), electric and hybrid electic vehicles (SiC) (in addition to inverters), motor drives (SiC), uniterruptable power supplies SiC), and heavy-duty vehicles (SiC), laptops, mobile charging (GaN), and low voltage converters (GaN).

5 https://www.networkworld.com/article/3531316/data-center-power-consumption-holds-steady.html

https://www.networkworld.com/article/3531316/data-center-power-consumption-holds-steady.html

8

Table 1. Performance targets for SiC and GaN applications with highest near-term priority

Performance Targets Year 1-5 Year 6-10

Electric Vehicle Inverters (SiC)

Device Type MOSFET MOSFET Rated Current for

MOSFET [A] >60 >150

Rated Current for MOSFET Body Diode

[A] >60 >150

Cost $*mOhm @ 150C <0.3 packaged Cost parity to Si IGBT

Enterprise Equipment (GaN)

Device Type Lateral normally-off

Lateral normally-off

Rated Current for HEMT [A]

Primary: 600–650V 50A Secondary: 80–100V 100A

Primary: 600–650V 100A Secondary: 80–100V 150A

Cost $*mOhm @ 150C <0.5 packaged assumes no integrated gate driver <0.15 packaged

PV Inverters (1200-1700V SiC)


MOSFET [A] >60 >150


[A]

>60

>150

Cost $*mOhm @ 150C <0.3 packaged <0.2 packaged

SiC 650V: Data center


MOSFET [A] >50

>100

Rated Current for

MOSFET Body Diode [A]

>50 >100

Cost $*mOhm @ 150C <0.3 packaged

<0.2 packaged

Heavy-duty vehicle (SiC 1.7-3.3kV)


MOSFET [A] < 20 < 60


[A] <20 <60

9

Cost $*mOhm @ 150C < 3 packaged < 1 packaged Onboard and offboard EV and HEV charging (GaN 600-900V)

Device Type GaN HEMPT GaN HEMT Rated Current for GaN

HEMT > 50 > 100

Cost $*mOhm @ 150C Si parity Si parity

Applications with Highest Longer-Term Priority The applications listed in Figure 2 are also high-priority focus areas but have a longer timetable (6-10 years) to reach commercial competitiveness. Note that many of these applications have higher voltage ratings than their near-term counterparts because fabricating high-voltage devices at large volumes will remain a challenge for at least the next few years. Table 2 shows performance targets for some of these longer-term priority applications.

Grid-tied energy storage (SiC): Energy storage systems—which consist of facilities, energy storage devices, and power electronics for power conversion—have long been a focus for government agencies such as DOE Office of Electricity. PE plays a key role at the bidirectional interface of DC batteries and the AC power supply for grid-tied systems. The ability to transfer power between the battery and the grid more efficiently and reliably is crucial for grid-tied energy storage systems. WBG semiconductors offer reduced size, increased simplicity (e.g., reduced use of PE switching devices), minimal round-trip energy losses, reliability under harsh conditions, and reduced overall system costs for electric utilities.

Residential PV (SiC): PV technology continues to improve and is quickly becoming an important source of energy for residential and industrial use. Coupled with reductions in PV module costs and government tax credits, residential solar power systems will eventually turn to WBG PE to enable smaller device size and increase efficiency, yielding home energy savings. For example, transformer-less WBG PV inverters would be a good choice for relatively low-power applications in residential areas. Although many technical challenges remain, R&D activities are ongoing.

Solid-state circuit breaker (SiC): PE technology is not just about converting power and controlling electrical energy; in electrical systems, it is important to detect issues such as overloads and short-circuit currents extremely quickly. SiC-based solid-state circuit breakers (SSCB) can increase safety by enabling fast fault isolation. WBG-based SSCB technology is currently expanding into areas where fast-switching without noise or losses at higher frequencies is needed (e.g., switching for grid applications).

Additional longer-term applications were identified by PowerAmerica members in 2019. Defining the related performance characteristics needed for these applications will appear in subsequent revisions to this Technlology Roadmap. Aadditional longer-term applications are: rail transportation (SiC), high speed electric drives for oil and gas industries (SiC), more electric aircraft (SiC and GaN), wireless power transfer (GaN), power correction factor circuits (SiC), and solid state transformers (SiC).

10

Table 2. Performance targets for SiC and GaN applications with highest longer-term priority

Performance Targets Year 1-5 Year 6-10

Grid-tied energy storage (SiC 3.3-10kV))

Device Type MOSFET

MOSFET

Rated Current for MOSFET [A]

>20 >60


[A]

>20

>60

Cost $*mOhm @ 150C <3 packaged <1 packaged

Wide Bandgap Power Electronics Pricing Comparison Figures 3a–3f below compare the cost of various Si, SiC, and GaN power devices. The data were obtained in December of 2017 from the Digikey distribution price list. SiC diode and MOSFET devices listed in PowerAmerica’s Device Bank within the last year were also included in the comparison, although the number of comparable devices was limited.

Figure 3a. SiC Schottky Diodes. At approximately 600V, SiC Schottky diodes are similar in price (US cents/Amp) to Si diodes. However, the prices quickly diverge as voltage increases. At 1200V, the price of the SiC diodes is 2.5-3x higher than that of their Si counterparts. At 10kV, SiC diodes can cost as much as $20/Amp.

11

Figure 3b. SiC MOSFETs vs. Si CoolMOS and Si IGBTs (<=1200V). Compared to Si super junction MOSFETs, SiC MOSFET prices are similar in the 500-800V range. Si IGBTs remain less expensive than SiC MOSFETs for all voltage ratings.

Figure 3c. Low-Voltage GaN HEMTs vs. Si MOSFETs (Cut-Tape, EPC/Infineon). The current price of GaN eHEMTs is overall higher than that of Si MOSFET. Price scalability with chip area is different for GaN and Si devices.

RDS (ON) (mΩ)

Unit P

rice (

US $) GaN e-HEMTs

Si MOSFETs80-100V

30-40V

12

Figure 3d. 30–60V GaN HEMTs vs Si MOSFETs. In Figures 3d and 3e, a higher voltage rating for a given device generally corresponds with a higher unit price and on-resistance value.

Figure 3e. 80–100V GaN HEMTs vs. Si MOSFETs. In Figures 3d and 3e, a higher voltage rating for a given device generally corresponds with a higher unit price and on-resistance value.

13

Figure 3f. 600–650V GaN HEMTs vs. Si MOSFETs. At 650V, the price of GaN cascode devices is 2-3x higher than their Si counterparts.

Summary In general, SiC power devices become increasingly more expensive than their Si counterparts as voltage rating increases (Figures 3a and 3b). At 600V, there is approximate cost parity; above 900V, SiC device prices are 2-4x higher than Si IGBT prices. The price differential is primarily attributed to the higher material cost of the SiC substrate/epitaxy and the larger chip area. Devices with larger chip areas and thicker epitaxy experience a non-linear increase in material defects which reduce yield.

At lower voltages, GaN power devices are the main competitor to Si super junction MOSFETs (SiC MOSFETs are only rated for >600V). While Si MOSFETs and IGBTs are still the least expensive devices available, the price differential between GaN and Si devices generally diminishes as on-resistance increases (Figures 3c-3e). At low chip resistances, only a few GaN products have prices similar to Si MOSFETs. This suggest that the costs of materials (GaN-on Si) and device fabrication/packaging are still higher for GaN devices than for Si devices. Reducing the unit price of GaN devices with low on-resistance is key to market adoption.

14

Device Bank PowerAmerica developed the Device Bank as a resource to help its members quickly obtain SiC and GaN devices and modules for research, testing, and product development. The long lead times and significant expense of acquiring these types of advanced engineering samples present a significant barrier to advancing and commercializing WBG technologies. The Device Bank fills this gap by offering an inventory of readily available WBG devices and modules that individuals and organizations can use in their projects. In exchange for access to these devices, Device Bank customers provide performance information back to PowerAmerica that manufacturers can use to improve their products.

The inventory in the Device Bank is open to any organization or individual interested in purchasing WBG devices and is offered on a priority basis to PowerAmerica members, particularly those receiving Institute funding for technical or educational products. Members have also taken advantage of this valuable resource for WBG projects funded by other federal agencies and DoD. All device bank users agree to the terms of a material transfer agreement and end use agreement that stipulates the conditions of their use.

The current Device Bank inventory is available at https://poweramericainstitute.org/devicebank/ engineering-samples/. Appendix B also contains a sample of the types of devices maintained in the Bank.

https://poweramericainstitute.org/devicebank/engineering-samples/


15

PowerAmerica’s 5-Year Roadmap Strategy

WBG technologies have the potential to reduce energy consumption and emissions in a variety of industries, while also creating manufacturing jobs across the United States. To capitalize on this potential, this roadmap offers a strategy for making WBG semiconductor technologies cost competitive with Si-based PE and for accelerating the adoption of SiC- and GaN-based components in new markets and applications. The roadmap thrusts—Reducing Cost, Improving Reliability and Quality, Enhancing Performance Capabilities, and Strengthening the Power Electronics Ecosystem—are intimately connected to form an integrated, collaborative strategy for advancing SiC and GaN technologies for PE.

PowerAmerica recognizes that accelerating large-scale adoption and high-volume manufacturing of WBG semiconductor devices requires a coordinated, interdisciplinary approach involving stakeholders from throughout the PE industry, including large, small, and start-up companies; universities; and national laboratories. With the goals of large-scale adoption and high-volume manufacturing in mind, the four roadmap thrusts discussed in the section are not only strategies to pursue, but also the challenges that industry must solve. The following section outlines the challenges facing large-scale commercialization of SiC and GaN technologies followed by key activities for PowerAmerica to pursue to address those challenges. Tables 3-10 categorize these challenges and activities as either application-specific (e.g., 650V data center) or crosscutting (applying to multiple applications) based on member inputs.

Figure 4. Overview of PowerAmerica’s Roadmap Strategy

16

Thrust 1: Reducing Cost Reducing the cost of WBG semiconductors is critical to accelerating their adoption. A key driver for successfully reducing costs to levels competitive with Si-based PE is to facilitate high-volume manufacturing and coordinate efforts across the PE community to:

1) Ensure availability of PE components, from materials and wafers to devices and modules; 2) Streamline design, manufacturing, packaging, and system integration; and 3) Engage stakeholders, from materials providers to manufacturers to end users.

This thrust focuses on identifying early adopters, high-volume markets, and applications in which there is the greatest need for improved performance and cost-effective manufacturing strategies. Demonstrating the competitive total cost of ownership of WBG power electronics in key applications can help advance this thrust. Coupled with Thrust 4, this thrust is critical to achieving the mission of PowerAmerica.

Key Challenges Although recent advances in WBG-based technologies are rapidly reducing the cost gap between SiC/GaN power devices and Si devices (e.g., 100A), the higher cost of WBG PE devices still makes their market penetration slow and difficult for some applications. PowerAmerica’s mid-2018 survey of its members confirmed this point, with most members responding that cost is a significant, even “severe” challenge in the adoption of WBG technology. The cost of WBG devices is mainly linked to the high cost of raw materials, such as SiC substrates, which account for up to 40% of the overall cost of a device. In particular, the GaN community suffers from a lack of low-cost epi-wafers with low defect density, as well as substrates suitable for specific applications.

The obvious challenge for the industry is to cost-effectively manufacture WBG devices at large scale (see Table 3).

For SiC power devices, modules, and applications:

• Focus on funding manufacturing activities that continuously lower the unit price of WBG devices. This especially applies to SiC devices with 600V to 1.7kV rating and >100A current capability, which represent the dominant and emerging markets in the next 5 years.

• Higher current density operation, which will produce a high-current device with a smaller chip area, is key to reducing SiC device prices.

• Higher current density operation requires development of new packaging technologies with low thermal resistance, advanced cooling techniques, and more compact chip designs.

For GaN power devices, modules, and applications:

• Focus on funding manufacturing activities that lead to continuous improvement in the device/module reliability, and high-frequency power applications with compact packaging configurations.

17

Table 3. Reducing Cost: Application-Specific Challenges

Application Thrust 1 Challenges

SiC

650V-1.2kV: Data center (750W power supply) EV traction, EV charging • High device cost compared to Si, particularly for high-current devices

1.2-1.7kV: PV string inverter, traction (e.g., EV/PHEV and rail), grid-tied

energy storage, heavy-duty vehicles, electric aircraft, industrial motor drive,

circuit protection

• High cost due to large chip size

10kV, 15kV: MV drives, MV DC naval platform, wind, advanced distribution

system, solid-state circuit breaker • Market exists, but drivers are unknown

Crosscutting Challenges • High costs across the board • Need more substrates and epitaxy from more vendors • Challenge controlling/limiting gate drive dv/dt

GaN

Enterprise equipment (e.g., DC/DC converters, data center, HV DC/DC)

• High GaN manufacturing cost

Residential PV systems (100/650/900V) • Need for improved cost and quality of epiwafers; Packaging cost for PV reliability grade devices

Low/mid-voltage non-traction automotive electronics

• Lack of low-cost epiwafer (<$800 for 150 mm, <$1,400 for 200mm), low defect density 106/cm2 GaN-on-Si

• Need for gate drive design Mobile chargers (wired & wireless

laptops, tablets, mobile devices) and LED driver

• Need for drive Integration, Lower epi cost, reduced cell pitch, improved 2DEG conductivity

EV & HEV charging (wired & wireless) • Reducing hetero-epitaxy cost and improving yields for large area devices; lower cost reliable packaging

Motor drives for fractional to integral horse-power motors (ind. motion

control & robotics, white goods, HVAC)

• Need creative cost engineering (e.g., epi, device, and monolithic integration)

• Need to lower GaN/Si epitaxy cost with good epi uniformity; thicker epi for higher voltage switches

Military/Aviation • System cost of manufacturing small quantities is too high

Consumer AC/DC • Need for hetero-epitaxy cost reduction and lowest cost foundry processes

Crosscutting Challenges • Risk that SiC will be cheaper, more applicable, or more promising

than GaN • Loss of PCB due to device failure

Key Activities To increase the availability and accessibility of state-of-the-art WBG power components, PowerAmerica must work to make these technologies cost-competitive with Si technologies. These cost savings can be achieved by improving the efficiency and precision of manufacturing processes to create higher-quality, higher-value products that are more reliable, increasing their marketability for demanding PE applications. Increasing manufacturing capacity can also reduce direct manufacturing costs by minimizing the need for outsourcing. To avoid the high initial investment costs of building new WBG foundries, PowerAmerica has leveraged X-FAB to fabricate WBG devices (e.g., SiC MOSFET). This open foundry model significantly reduces the equipment costs and overhead of WBG device manufacturing. The activities listed in Table 4 are actions that PowerAmerica can reasonably address within its authorized scope of activities.

18

Table 4. Reducing Cost: Application-Specific Activities

Application Technical Activities

SiC

1.2-1.7kV: PV string inverter, traction (e.g., EV/PHEV and rail), grid-tied energy

storage, heavy-duty vehicles, electric aircraft, industrial motor drive, circuit

protection 1.7-3.3kV: UPS, rail traction, power

quality, wind: conventional wind machine with WBG, rail auxiliary power

supplies with 1.5kV bus HVDC

• Bring down bus costs

4.5kV, 6.5kV: Rail traction, grid-tied charging, UPS


system, solid-state circuit breaker

• Bring down bus costs

Crosscutting Activities

• Reduce cost by multiple methods: reduce material/fab costs, shrink die size, etc.

• Conduct application demos (open BoM) comparing Si, SiC, and GaN devices with value proposition to educate user community

GaN Crosscutting Activities

• Work to reduce the cost and increase the quantity of vertical GaN epiwafers

• Address GaN epi cost, quality, and capacity • Define a cost-model roadmap similar to the GaN LED cost model

roadmap made by DOE • Identify fabs and conduct cost analysis (overseas vs. US) • Lateral GaN epiwafer cost and quality

Thrust 2: Improving Reliability and Quality While SiC and GaN devices have demonstrated higher efficiency than Si-based devices in PE applications, some reliability concerns still limit the market acceptance of WBG technologies. However, stakeholders at the August 2018 roadmapping workshop reported that the focus on improving reliability is beginning to shift to quality as more reliability data is collected. As confidence in the reliability of WBG technologies improves, it frees industry to dedicate more attention to the quality of WBG materials and devices to make them competitive with traditional Si-based technologies.

As WBG technologies mature, the PE community cannot gain confidence in their use without strategies to establish their reliability and quality. A key first step is gaining a better understanding of degradation/failure mechanisms under harsh conditions (i.e., high voltages and/or high temperatures) as well as sources of product quality issues. It also requires generating high-quality data using advanced testing methods, developing standards, and effectively communicating reliability best practices and quality information to end users.

Key Challenges The success of WBG PE technologies in key markets and applications requires that industry first overcome several challenges (Table 5) that face the reliable production and operation of SiC/GaN devices. Responses to the mid-2018 survey found that reliability and cost were the most important factors when members considered WBG PE deployment. The reliable performance of power devices is highly influenced by factors such as structure, level of defects (e.g., basal plane dislocation [BPD]), processing/manufacturing

19

conditions, packaging, and device degradation from harsh operating conditions (e.g., high-temperature, high-voltage, high-frequency operation).

Table 5. Improving Reliability and Quality: Application-Specific Challenges


SiC

650V-1.2kV: Data center (750W power supply) EV traction, EV charging

• Lack of packaging that is field proven for thermal cycling • Redefine AEC standards for SiC • AQG-324 • Insufficient assurance of quality (parts per billion) for PE in

automotive applications/lack of data demonstrating automotive quality

1.7-3.3kV: UPS, rail traction, power quality, wind: conventional wind

machine with WBG, rail auxiliary power supplies with 1.5kV bus HVDC

• Reduce the negative gate bias during turn-off from-4V to -7V • Not yet on par with Si IGBT • Still not enough reliability data


• Not yet on par with Si IGBT • Still not enough reliability data


system, solid-state circuit breaker • Need qualification test • Need for optimized IGBT buffer-layer and lifetime

Crosscutting Challenges • Lack of JEDEC standards • Reliability data is not published enough • Lack of long-term reliability testing

GaN


• Insufficient standards for GaN reliability

Residential PV systems (100/650/900V) • Firming up device reliability specifications for PV inverter applications and validating GaN switch reliability

Low/mid-voltage non-traction automotive electronics • Need for reliability standards and benchmarking

Mobile chargers (wired & wireless laptops, tablets, mobile devices) and

LED driver

• Need for reduced epi layer defects and improved control of interface states

• Need for increased system-level testing/reliability (high temperature operating life)

EV & HEV charging (wired & wireless) • Establishing reliability testing equivalent to AEQ 101, but appropriate for GaN switches, and validating


control & robotics, white goods, HVAC) • Inadequate consumer and industrial grade reliability for respective

segments

Military/Aviation • Complying with military/FAA certifications • Device simulation models are not standardized • Lack of reliability testing capability 100V-200A/ 540-600V (1700V) • Lack of system-level testing/reliability

Consumer AC/DC • Consumer-grade reliability is currently inadequate; very low rates of early failure are acceptable—less than a few ppm

Crosscutting Challenges

• Questions surrounding what is needed to demonstrate reliability, (e.g., demonstration of use in the field?)

• Lack of reliability standards for high-probability area for GaN • Will the JEDEC WBG committee solve lingering reliability questions? • Root causes of failure mechanisms

20

Key Activities To continue to increase confidence in WBG devices, the PE community must qualify, validate, and establish reliability and quality standards for WBG PE components. To do so, PowerAmerica should continue facilitating the increased quantity and accessibility of data, establish qualification standards, and develop accelerated testing methodologies for consistent use throughout the industry. Reliability testing and correlation studies are critical to improving the understanding of degradation and failure mechanisms and their impact on the performance of WBG modules, packaging, and integrated systems. Table 6 outlines activities to improve the reliability and quality of WBG PE components for specific applications.

Table 6. Improving Reliability and Quality: Application-Specific Activities


SiC Crosscutting Activities

• Conduct large-scale, third-party reliability testing (per JEDEC standard) of SiC devices and create a field quality database

• Machine learning algorithm development • In-situ physics of failure mechanisms • Root cause mechanisms


• Create a lab to provide trusted, third-party testing • Conduct large-scale, third-party reliability testing (per JEDEC

standard), create a field quality database • Develop a roadmap of gaps in reliability requirements for emerging

applications (military, aero, space, automotive) and projects to meet the gaps

• Create mission profiles for key applications • Study reliability failure mechanisms and publicize openly • Develop well-publicized reliability benchmarks for applications • Conduct reliability benchmarking • Define reliability standards • Provide independent reliability testing labs and service • Support Fab device and test • High frequency magnetics • Machine learning algorithm development • In-situ physics of failure mechanisms

Thrust 3: Enhancing Performance Capabilities In addition to reducing cost and establishing the reliability and quality of SiC and GaN devices, the PE community must address numerous technical issues that prevent WBG PE from realizing their potential in higher-temperature, voltage, and frequency operations. Activities to enhance performance capabilities—including design best practices, optimization of structures, and exploration of new circuit topologies by the end user—are critical to accelerating commercialization of SiC and GaN technology in a wide range of products.

Key Challenges The PE community must solve technical challenges at all levels of PE systems, including challenges related to device performance, module integration and packaging, and qualification standards (Table 7).

Device-level challenges: The GaN device community currently faces the challenge of developing high-frequency power converters and improved normally off (i.e., enhancement-mode) devices. Challenges in

21

SiC devices include improving channel mobility with increased channel density and reduced channel lengths, reducing EMI with reduced parasitics, and optimizing gate drivers for specific application environments.

Module and packaging challenges: Key module and packaging challenges include the need to improve high-voltage insulation, thermal management, partial discharge, and EMI to enable high-performance modules (e.g., double-sided cooled power modules operating at a higher junction temperature (Tj,max: 175°C–200°C), and high-performance discrete packages that can operate at higher temperatures and voltages.

Qualification standards challenges: Because WBG PE technologies are relatively new, there is high demand for qualification standards for PE technologies similar to the Automotive Electronics Council’s Qualification (AECQ) standards developed for automotive electronics or the Joint Electron Device Engineering Council’s (JEDEC) standards for electronics.

Table 7. Enhancing Performance Capabilities: Application-Specific Challenges


SiC


• Need improved short circuit time • Trench issues, devices require PowerAmerica support to stay

competitive • Need for advanced large area trench • Need for lower gate charge • Need gate circuit drivers capable of handling short circuit, repetitive

short circuit • Need controllers that meet short-circuit requirements of SiC • Need for discrete packaging that can operate at higher temperatures • Need for a gate drive that supports higher-frequency operation • Need for cost-effective double-side cooled module technology (Tj, max:

175-200°C • Robustness of SiC MOSFETS vs.Si IGBTs at same voltage level



circuit protection

• Need for more AEC-like qualifications, short circuit rating, and/or optimized gate driver

• Lack of good high-temperature packaging (e.g., for down hole) • Need for package standardization with dual sourcing



• Wafer quality is too low • Need for optimized gate driver in packaging • Need for double plastic coating alternative in packaging


• Lack of HV system packages • Lack of bus designs • Difficulty in lowering induction and maintaining good isolation of

partial discharge at high voltages (note that this is especially needed for 10kV applications)



• Lack of bus designs • Need for advanced gate drivers • Issues in packaging and modules (i.e., inductance, common mode

capacitor, HV insulation, and/or partial discharge) • HV system packaging (i.e., voltage insulation, thermal management,

and/or EMI) • SOA optimization for IGBT (i.e., SC-SOA and or UIS)

22


• Need for reduced parasitics • Need new structures which could give jump in performance • Lack of standardization of testing • High temperature capabilities are rated for 175°C but should get up

to 200°C • Need modules with higher current ratings

GaN


• Need high-performance DC/DC converter • Need high-performance, low-cost GaN for low voltage (100V and

below) • Lack of high-frequency power converter controllers • Lack of high-speed, low-cost drivers • Insufficient GaN packaging and thermal solutions • Need for appropriate power supply packaging (EMI concerns)

Residential PV systems (100/650/900V) • Reference designs for the power stages • Need for faster protection hardware • Need for improved capacitors for increased system reliability

Low/mid-voltage non-traction automotive electronics

• Lack of normally off devices VTh>4 • Lack of design tools for integrated GaN smart power circuits • Lack of good passivation to achieve low surface state density (1011

cm-2 eV-1) • Need for double-sided cooling (vehicle and vibration)

Mobile chargers (wired & wireless laptops, tablets, mobile devices) and

LED driver

• Need for expanded offering to cover current range • Need to integrate key functions to enable improved soft switching

topologies

EV & HEV charging (wired & wireless) • Reference designs for the several types of systems used for automotive charging


control & robotics, white goods, HVAC)

• Need to begin adoption in areas where performance is more important than cost and volume—such as high-speed machine tools requiring precision speed control, fast compact robotic/motion control systems, etc.

Military / Aviation • Lack of radiation performance data (all types of radiation) • Lower-volume application less likely to be served in the short term

Consumer AC/DC • Need for reference designs with new circuit topologies showing lower overall system BOM & cost

Crosscutting Challenges • Combining multiple gate drive functions into a chip

Key Activities The PE community must demonstrate the advanced capabilities of WBG power devices, which are smaller, faster, and more efficient than Si-based PE in high-voltage, high-temperature, and high-frequency operating environments. To support widespread adoption and large-volume manufacturing of next-generation WBG products, the PE community should pursue technical activities that will increase the efficiency and capacity of manufacturing processes and enhance the performance capabilities of WBG materials, devices, modules, and systems (Table 8). Such activities should include producing reference designs, scaling voltage and power, building WBG gate drives, and improving figures of merits (e.g., on-resistance, leakage, and/or low parasitics).

It is important to note that the technical activities for this thrust are not possible without advanced peripheral components, thermal management techniques, gate driver technologies, and innovative EMI/EMC solutions. Technologies like high-frequency magnetics are beyond the scope of PowerAmerica, but these technologies should be monitored through Thrust 4: Strengthening the Power Electronics Ecosystem.

23

Table 8. Enhancing Performance Capabilities: Application-Specific Activities


SiC


• Develop direct die design (package free) • Integration with the circuit—focus on boundary between design and

manufacturing • Improve the accuracy of PSpice and PLECS models • Standardize power modules (EMI activity) • Develop reference designs • Develop optimized automotive power module, optimized for AEC

qualification, low-inductance (not “Econo-Pak” style), direct water glycol cooling, and low cost

• Develop circuit design software with EDA company involvement • Develop thermal design tools with EDA company involvement • Improved packaging for high temperatures • Innovating thermal management materials and solutions for packaging • Easy to use, accurate and compact SPICE models for circuit and system

simulations 1.2-1.7kV: PV string inverter, traction

(e.g., EV/PHEV and rail), grid-tied energy storage, heavy-duty vehicles, electric aircraft, industrial motor drive, circuit




• Develop process models to save time • Improve thermoelectric models • Develop improved gate drivers with additional sensing and protection

capabilities (e.g., design for fast short circuit) • Develop high power density packaging • Embedded/surface mount/small module/top-cool/low inductance die

attach • Quantify the impact of combining thermal stresses and coolant (compact

SPICE model and electrical/thermal model)




• Focus on MV 13.8 kV grid-tie VSC • Develop module packaging • Partial discharge design for HV • Establish DC voltage derating for same FIT rate for HV 10-15 kV MOSFET

and IGBT • Consider DC-DC MV converters (e.g., MV-LV: 8-10 kV800V-1000V and

with HF magnetics) • Develop module packaging: insulation degradation at HV and HF operation • Focus on MV VSC @ 4.16 kV (compare IGBT vs. MOSFET) • Focus on HV isolated 15 kV (e.g., 20 kV gate driver and SC protect) • Develop module packaging for non-isolated base plate module


• Increase design automation for increased reliability and faster time to market

• Develop advanced packaging to take advantage of GaN performance (package inductance, cooling flip chip, etc.)

• Develop new converter topologies that specifically benefit from GaN (i.e., high speed, small size)

• Create lab to provide trusted, third-party testing • Packaging for highly accelerated testing • Application-specific testing • Design tools/models for integrated GaN circuit simulation • Develop reference designs including full systems (vs. evaluation boards) • PV inverter/data center applications • Spread knowledge about GaN designs (education, workforce, tutorials, etc.) • Develop low-cost, high-speed drivers • Market GaN technology • Standardize packaging and voltage threshold

24

Thrust 4 Strengthening the Power Electronics Ecosystem To realize the potential of WBG technologies, the PE community is building a network with strong information-sharing capabilities to increase interaction between supply chains, manufacturers, and end users. There is broad consensus that several deficiencies exist across the PE community—including the need for cost-effective U.S. manufacturing capacity, a workforce with a deeper knowledge of WBG PE advances in technologies such as epi-materials, and advanced complementary technologies such as magnetics or vertical GaN devices—which may delay the commercialization of WBG technologies if not addressed concurrently. Although some of these challenges (e.g., basic science issues) are outside the scope of PowerAmerica, the Institute should monitor and identify national/international activities with the most merit to ensure WBG PE sustainability with further investment.

In particular, focusing on education and workforce development will help strengthen the power electronics ecosystem. PowerAmerica has established and will continue to develop a robust program that provides tutorials, short courses, and job training for both students and working professionals to support the growing WBG PE industry. Raising awareness and interest among students is also important and will help develop a talent pipeline. Several members have voiced the difficulties they face in recruiting skilled workers—this is an area where PowerAmerica can have a direct impact.

Key Challenges In addition to targeting the above technical challenges, the PE community must also improve collaboration and information sharing by strengthening the PE ecosystem, and must monitor additional technology areas that could result in emerging markets for SiC and GaN devices (Table 9).

Challenges to strengthening an ecosystem for WBG PE technologies: Encouraging a community-driven, integrated framework for collaboration and information sharing across the PE community related to pre-competitive, common challenges is accelerating device innovation and adoption. Strengthening this ecosystem, however, is difficult due to the number of industry sectors involved, including wafer suppliers, device design houses, microelectronic fabs, and the multidisciplinary nature of PE challenges. The breadth of the PE community makes it difficult to establish open, shared resources and libraries (e.g., foundries and reliability databases) that can help drive advanced manufacturing and accelerated integration of SiC and GaN devices. The community is also working together to solve challenges such as limited supplier availability and provide workforce training. Multidisciplinary efforts must be made to reduce device cost, enhance reliability, and improve system performance and integration.

Monitoring supplementary technologies: Though the mission of PowerAmerica necessitates focusing primarily on WBG manufacturing and activities between MRL 4 and 7 to maximize energy savings and job creation, the PE community must also monitor emerging, niche, and complementary low-volume PE markets. These areas include novel device design (e.g., vertical GaN) and high-frequency magnetics.

Education and Workforce Development: There is a dire need for a new generation of WBG-trained professionals, primarily at the undergraduate and graduate levels, to address the above technical challenges in the near future. Additionally, training working PE professionals in WBG technologies can provide immediate benefits to the WBG PE ecosystem.

25

Table 9. Strengthening the Power Electronics Ecosystem: Application-Specific Challenges


SiC

650V-1.2kV: Data center (750W power supply) EV traction, EV charging • Need motor insulation



circuit protection

• Lack of device/module suppliers • Not enough 100-200A single chips



• Short circuit being addressed, but tradeoff with reduced RDS(on)

4.5kV, 6.5kV: Rail traction, grid-tied charging, UPS • Short circuit being addressed, but tradeoff with reduced RDS(on)



• Short circuit being addressed, but tradeoff with reduced RDS(on) • Few sources of epi; need multiple sources that offer low defects, high

uniformity, and low costs


• Lack of device and packaging standardization • Lack of testing standardization • Lack of bus designs • Material availability is a challenge • Not enough workforce development and training • Advanced packaging and bus design remain a challenge • Application draining

GaN


• Need for production of integrated magnetics • Insufficient workforce training for GaN and high-frequency design

Residential PV systems (100/650/900V) • Need for lower loss and higher-frequency magnetics Low/mid-voltage non-traction

automotive electronics • Need multiple approved vendors for epi, devices, and packaging

Mobile chargers (wired & wireless laptops, tablets, mobile devices)

• Need to encourage control IC and magnetics makers to push to higher frequency capability

EV & HEV charging (wired & wireless) • Need for stronger interaction between battery makers, PE groups, and automobile designers to realize this potential


control & robotics, white goods, HVAC) • New standards for high-efficiency motors and incentives to

motor/drives manufacturers in a very conservative industry

LED driver • Need to encourage control IC and magnetics makers to push to higher-frequency capability

Military / Aviation • Lack of epi 200A/270V (1200V) 200V/540V (1700V) • Limited supplier availability and supplier obsolescence

Crosscutting Challenges • Not enough GaN application support/knowledge to consider and

advance GaN • Standardized packaging and voltage threshold

Key Activities Strengthening a collaborative WBG PE ecosystem is critical to accelerating advancement of WBG PE technology and establishing the technology’s commercial viability in a variety of applications. To facilitate this collaboration, the PE community should develop a framework for information sharing (e.g., manufacturing reliability data and process recipes), reinforce partnerships across the value chain, and build

26

a more forward-looking and knowledgeable PE workforce. To maintain a sustainable PE ecosystem, the PE community should continuously monitor emerging and complementary technical areas, including work in epi-materials, high-frequency (or very-high-frequency) applications, magnetics, and advanced architectures.

Table 10. Strengthening the Power Electronics Ecosystem: Application-Specific Activities


SiC


• Explore the potential of advanced manufacturing (3D printing) • Conduct demos with next-generation passives (work with passives

vendors) • Develop filter designs • Give advice to standardization committees • Publish more testing results

1.2-1.7kV: PV string inverter, traction (e.g., EV/PHEV and rail), grid-tied energy

storage, heavy-duty vehicles, electric aircraft, industrial motor drive, circuit




• Develop reference designs • Create a list of projects to match businesses with university capabilities




• Study HF and HV effect on insulation (for magnetics)

Crosscutting Activities

• Support university projects on SiC power devices to provide workforce for projected growth in SiC market (PA doesn’t fund enough university projects)

• Offer more courses and workshops and continue teaching grad students (let them see this data early on)

• Use application demos to educate the PE community on how to get the most out of SiC/WBG PE

• Encourage new fabs and increase competition • Perform more device modeling and co-simulation • Develop standards and perform benchmark testing for reliability and

performance • Initiate new projects in EMI • Focus on the reliability of packaging • Use multimedia to spread knowledge (e.g., videos, web conferences,

etc.) • Make sure companies show up to workshops and conferences so they

can offer their input and feedback


• Establish a GaN foundry to have universal access (in U.S.) for all users • Fund more field deployments (and data collection) especially for

university projects • Modernize design techniques (layout, magnetics, thermal

management) • Collect product (GaN, drivers, etc.) gaps from users and give feedback

to device suppliers • Identify gaps in driver requirements (size, cost, performance) by

application and use these as inputs for component suppliers’ roadmaps • Recruit more GaN-focused members

27

GaN Crosscutting Activities (continued)

• Develop publications that establish reliability to increase adoption • Webinar, podcasts, social media training • Prove high-frequency magnetics scalability requirements • Fill gaps between feasibility and product of integrated planar magnetics • Promote efforts to improve passives, especially magnetics

Review by PowerAmerica 2019 Annual Meeting Participants Tables 3 through 10 were reviewed in summary form at the PowerAmerica Annual Meeting in February 2019. An informal survey was taken requesting that participants indicate the activities they consider the highest priorities in addressing the challenges in each of the four thrust sections in the SiC and GaN categories. The survey results indicated the top SiC category is “reliability of packaging” (see cross cutting activities in the SiC section of Table 10). Second was “integration with the circuit” (see the 650V to 1.2kV section of Table 8). It should be noted that several challenges and activities related to these two priorities and are found throughout Tables 3 through 10. Other categories were not close in terms of the preferences indicated by the survey participants. For the GaN categories, the top three preferred categories are: “study reliability failure mechanisms and publish results” (see Table 10); followed by “develop advanced packaging to take advantage of GaN performance” (see Table 8) and modernize design techniques (layout, magnetics, thermal management) (see Table 10). It should be noted that several challenges and activities related to these two priorities and are found throughout Tables 3 through 10. Other categories were not close in terms of the preferences indicated by the survey participants. These results were used in developing topics for PowerAmerica’s 2019 Member-Initiated Projects.

28

mmmmm Path Forward

This roadmap provides guidance to accelerate large-scale adoption of WBG semiconductor devices in PE systems. Continued and increased investment in R&D, coupled with workforce development and training activities, are critical for maximizing U.S. manufacturing and global competitiveness and enabling the PE community to meet increased market demand in the next decade. PowerAmerica will continue to fund projects that address key gaps and challenges and will rely on the expertise of its members to move wide bandgap power electronics into the future. In addition, PowerAmerica will look to bring additional members into the wide bandgap semiconductor ecosystem. This roadmap is a living document that the Institute will revisit and update once each year with the support of its members.

29

Appendix A. Acronyms 2DEG Two-dimensional Electron Gas AC Alternating Current ACRMS Alternating Current Root Mean Square AECQ Automotive Electronics Council Qualification AlN Aluminum Nitride ARPA-E Advanced Research Projects Agency–Energy BOM Bill of Materials BPD Basal Plane Dislocation Cgs Gate-source Capacitance CMOS Complementary Metal Oxide Semiconductor DC Direct Current EAS Avalanche energy EMC Electromagnetic Compatibility EMI Electromagnetic Interference E-mode Enhancement-mode EV Electric Vehicle FET Field-effect Transistor FIT Failures in Time GaN Gallium Nitride HEV Hybrid Electric Vehicle HT-DLTS High-temperature Deep-level Transient Spectroscopy HVDC High Voltage Direct Current IGBT Insulated Gate Bipolar Transistor Io Reverse Leakage Current JBS Junction Barrier Schottky JEDEC Joint Electron Device Engineering Council LED Light Emitting Diode Ls Source Inductance MOS Metal Oxide Semiconductor MOSFET Metal–Oxide Semiconductor Field-effect Transistor MV Medium Voltage OBC On-board Charger PE Power Electronics PHEV Plug-in Hybrid Electric Vehicle PV Photovoltaic QG Total Gate Charge QGD Gate-to-drain Charge Qoss Output Charge Qrr Reverser Recovery Charge

30

Rds-on Specific on-resistance RF Radio Frequency Rjc Junction-to-case Thermal Resistance Ron On-resistance SBD Schottky Barrier Diode SC Short Circuit Si Silicon SiC Silicon Carbide SOA Safe Operating Area tf Fall Time Tj,max Maximum Junction Temperature Rating tsc Short Circuit Time tr Rise Time UIS Unclamped Inductive Switching UPS Uninterruptible Power Supply Vf Forward Voltage Drop Vgs Gate to Source Voltage VSC Voltage Source Converter VTh Threshold Voltage WBG Wide Bandgap

31

Appendix B. PowerAmerica Device Bank The PowerAmerica Device Bank is a resource that makes advanced engineering samples of WBG power devices and modules readily available for projects that advance the PE industry. As of November 2018, PowerAmerica had 1,308 devices stocked for immediate use. This inventory was made up of project deliverables, donations, and purchase orders to fulfill customers’ needs. The inventory of devices (as of November 2018) is listed below. PowerAmerica is continually assessing the needs of their members and the WBG industry so they can procure new devices to further accelerate the research, development, and commercial adoption of WBG technology.

A sample of the current inventory and the request form to purchase a device from the Device Bank can be found at: https://poweramericainstitute.org/devicebank/engineering-samples/.

Table 11. PowerAmerica Device Bank

Contributor Device Part Number Price Per Unit

Wolfspeed-Fayetteville

1.2kV 20A 45mm Six-Pack (Three Phase) Module - Package: 45mm CCS020M12CM3 $118.80


1.2kV 20A 45mm Six-Pack (Three Phase) Module - Package: 45mm CCS020M12CM2 $273.00


1.2kV 300A 45mm Six-Pack (Three Phase) Module - Package: 45mm CAS300M12BM2 $336.60

GeneSiC

Schottky Diodes & Rectifiers 3300V - 0.3 A SiC Schottky Rectifier GAP3SLT33-214 $18.71

Transphorm Hard switched 1/2 bridge kit TDPS1000E0E10-KIT $1,500.00

Transphorm All-in-one PFC+Bridge, 200 kHz kit TDPS250E2D2-KIT $2,000.00

Transphorm LLC bridge ckt, 200 kHz kit TDPS251E0D2-KIT $2,000.00 Transphorm PFC boost ckt, 750 kHz kit TDPS300E1A8-KIT $1,500.00

Transphorm Totem Pole PFC, 500 W at low line kit TDPS500E2C1-KIT $3,500.00

USCi 200A, 650V SOT227 diodes UJ3D065200S $60.00 USCi 100A 1200V SiC Schottky Diode UJ3D120100ZW $40.00 USCi 60 TO247 devices UJ3M1240K $30.00

USCi 40m, 1200V Power MOSFETs in waffle packs UJM12042 $30.00

Wolfspeed 10kV/350 SiC MOSFET die XPM3-10000-3050B $750.00

Wolfspeed Gen 3 10kV/350 SiC MOSFET die XPM3-10000-3050A $750.00

Wolfspeed 10kV/350 SiC MOSFET die XPM3-10000-350A $750.00

Wolfspeed 10 kV/350mOhm SiC MOSFET die XPM3-10000-350A $750.00

Wolfspeed 10 kV/15A Gen 3 SiC Diode die XPM3-10000-3015A $300.00


32

Wolfspeed Gen3 3.3kV/40mOhm SiC MOSFET - Package: Bare die XPM3-3300-0040B $330.00

Wolfspeed 3.3kV/45A SiC JBS diode die XPM3-3300-Z045A $150.00

Wolfspeed 6.5 V/100mOhm SiC MOSFET die XPM3-6500-0100A $550.00

Wolfspeed 10kV/15A SiC JBS diode - Package: Bare die XPW3-10000-Z015B $300.00

Wolfspeed 10kV, 58 mOhm, all-SiC, half-bridge power modules

H610ML610M E042-01 TBD

Wolfspeed 3.3kV XHV-7 Power Modules, 8x3.3kV QPM3 die/SP H83ML83M TBD

33

Appendix C

Priority Concepts to Advance SiC and GaN Technology

Reported by the Member Breakout Sessions Based on the Technology Roadmap

PowerAmerica Annual Meeting February 25, 2021 Topic 1: SiC and GaN Device Package/Module Reliability and Ruggedness PA staff note: Items a-e are being addressed, at least in part, by projects funded in MIP Round #3. Members might consider reviewing the outcome of the Round #3 projects before pursuing Round #4 projects pertaining to a through d below.

a. Die interconnect, attach and termination schemes to reduce parasitic inductance, hence

• EMI, (new wire bonding, double side mounting, lead attachment alternatives, etc.) • Substrate materials/attach processes for hi-rel, encapsulation for higher temps/PD, • Explore alternatives to traditional wire bonding, such as foil and welding

b. New concepts for device cooling, heat spreading, and low thermal resistance;

c. Packaging techniques for reliable operation at extreme temperatures/pressures (e.g. -55C

to 250C) • Heat flux focused into baseplate not adjacent componentry • Evaluation and failure mode analysis of devices/modules specific to

certain applications such as traction inverter • Meet reliability standards of automotive/aerospace systems • Impact on partial discharge and material degradation • Encapsulation for high temp operation

d. Circuit demonstration of the benefits to integrating passives (e.g., decoupling

capacitors); identify any thermal, material, and manufacturing issues to be solved;

e. Modeling (electrical/thermal) of multi-chip modules and device paralleling; usually a part of module projects. See also 2(f);

f. Chip scale packaging with double-sided mounting options for PCB pick and place manufacturing;

g. Integrate sensors in the module and minimize impact on cost and performance. See also

2(a) and 2(g);

34

h. Evaluation and failure mode analysis of devices / modules specific to certain applications,

such as traction inverter, or harsh operating environments (very low (-55C)/high temperature (+200 Tj), pressure; vibration and impact);

i. Testing of low temperature and pressure (altitude induced) on module packaging relative partial discharge and material degradation.

Topic 2: SiC and GaN Device Design and Gate Drivers

a. Compact SiC monolithic/integrated intelligent and fast gate drivers with device health monitoring should be addressed as follows:

• Rather than SiC gate driver monolithic integration (low channel mobility can make the device large and costly) use chip stacking to integrate the driver (separate Si gate driver and SiC MOSFET into the same package); on device sensors can monitor device health;

• System level study to do cost partitioning vs criticality and functional quality of SiC monolithic vs package integration;

• Develop a “short list” of protect functions (similar to SmartFET approach taken with Silicon MOSFET’s): temperature, current sense, etc. to support reliability tracking; there is high demand for such functions in automotive according to silicon experience.

b. GaN Monolithic integration of gate driver, discrete HEMT and level shifting have been

commercialized. Develop monolithic GaN half-bridge integration at 600 V for compactness and cost savings.

c. Medium voltage (MV, 3.3 – 10kV) SiC devices for grid applications (renewables, battery

storage, EV infrastructure) should be addressed, considering (see also 3(a)): • Grid upgrade in the near future will be needed to accommodate large EV

infrastructure • Update/develop MV SiC MOSFET on-resistance, capacitance, and current rating

roadmap • Investigate the design /model /fabrication of SiC Super junction (SJ) FETs and IGBTs.

Is channel mobility a critical contributor to Rds in these devices? Can SiC SJ FETs be used in a half bridge hard switched topology? Determine the “knee voltage” of SiC IGBTs and their conductivity modulation. How does the on-state curve compare between SiC SJ FETs, MOSFETs, and IGBTs?

• Identify optimal SiC device for each grid application domain and device characteristics important to support the grid

• Study impact of cosmic ions on SiC SJ FETs and SiC IGBTs

35

• Study SiC SJ FETs and IGBTs in switching circuits • Study of how bipolar conduction induced defect generation is impacted by the

higher current densities in SJ and IGBT devices • High-yield high-current (large area) 3.3-10 kV MOSFETs. Present state of the art

from literature of BPD location and density in starting wafers; investigate BPD generation during fabrication and techniques to eliminate them.

d. Improve gate oxide interface for higher mobility and reliability/robustness in lower

voltage SiC trench and planar MOSFETs.

e. Study soft switching, high frequency “parasitic” losses seen in Si SJ FET, SiC and GaN devices. Identify causes (material limits, device configuration, etc.), impact on operation, and drive resolution.

f. Identify existing WBG TCAD and SPICE software missing functionality and provide input to

suppliers for improvement. See also 1(e).

g. Conduct a predictive reliability and prognostics investigation of devices (moved from Topic 1).

Topic 3: System Demonstrations and Integration Innovations for SiC and/or GaN Technology

a. Demonstrate WBG in smart grid technology addressing:

• Renewable energy system grid integration • Energy storage

b. EV fast charging infrastructure addressing:

• Heavy duty vehicles • Battery interface and integrated electronics for fast charging • Paralleling modules for high current • 800 V bus architecture Review DOE’s Extreme Fast Charging medium voltage (MV) projects to differentiate PowerAmerica’s work

c. SiC/GaN optimized motor drives and electric motors addressing (see also 4(d)): • Inverter topologies that minimize losses • Power electronics integration in electric motor • Techniques to minimize stress on motor insulation • Motor loss as a function of frequency Review DOE EDT project goals to differentiate PowerAmerica’s work

36

Topic 4: Out-of-the-Box Ideas Staff note: Not all these items are suitable for a MIP but could be pursued in other ways.

a. Present “lessons learned” from RF community for high frequency SiC/GaN power electronics in the form of meeting speakers and tutorials;

b. Present case studies of SiC and GaN value add in different applications: solar, EV, energy storage, etc. and anticipated future value add in these and other applications;

c. PowerAmerica should facilitate education on best design practices, tutorials/short courses on WBG EMI, inductance/parasitics, gate drive and circuit design;

d. SiC/GaN optimized motor drives and electric motors. See also 3(c); e. PowerAmerica’s Technology Roadmap should more specifically address EV mission profile

and how this guides device design for reliability and failure risks mitigation. Power module cooling and impact of temperature on devices and power electronics should also be addressed;

f. PowerAmerica should sponsor a university Grand Challenge on low power (25-50 kVA) WBG based EV reliability and short circuit protection through solid-state circuit breakers;

g. Improved capacitors are needed to support WBG modules: withstand higher temperatures or demonstrate thermal isolation from SiC devices (see also 1(d)).

Topic 5: Manufacturing Equipment, Processes, and Testing for SiC and GaN Technology Staff note: Some Topic 5 items might be undertaken by working groups of interested PowerAmerica members rather than in the form of a MIP project.

a. Facilitate consensus (not an industry standard) on 150 mm SiC substrate requirements addressing, for example:

• Doping concentration uniformity • Flatness and other mechanical properties

Add specificity to PowerAmerica’s roadmap aligning substrate properties with device and applications

b. Metrology standards development: • Standardize incoming inspection protocols standards for testing using AFM, PL, X-

ray, CD SEM, optical microscopy etc. • Investigate wafer spec value differences measured with different tools (disparity

between supplier and incoming wafer measurements)

c. Expanding the Device Bank to include SiC substrates:

37

• Receiving “reject” SiC substrates, devices and modules from producers and device manufacturers for use at universities, including broken pieces with large areas

• Providing reference wafers, including epi wafers, in which there is high confidence in the mechanical and electrical specs, for calibrating test and measurement equipment

• Sell wafers, similar to the way devices are sold from the Bank, especially to provide shorter delivery times for long lead time wafers

d. Understand the role of SEMI in developing standards for 200mm SiC wafers and what

PowerAmerica’s contribution to this effort might be.

Documents

Version 4.3 April 2021 - poweramericainstitute.org