Thirty-five years ago the silicon power MOSFET was a disruptive technology that displaced the bipolar transistor as the power conversion device of choice for the semiconductor industry – and a $12B market emerged. The dynamics of this transition taught us that there are four key factors controlling the adoption rate of a new power conversion technology:
- Does it enable significant new applications?
- Is it easy to use?
- Is it VERY cost effective to the user?
- Is it reliable?
A lot has happened in the few years since gallium nitride (GaN) technology was introduced as a significantly outperforming displacement technology for the MOSFET. GaN has assumed the position as the presumptive replacement for the aging power MOSFET, but in order for it to take the mantle as the semiconductor technology of choice, it must meet four requirements that all new technologies must meet to this leadership position. So let’s look at the four key attributes and see where GaN stands in addressing them.
Does it enable significant new applications?
GaN transistors and integrated circuits are significantly faster and smaller than the best silicon MOSFETs. Today, commercially available eGaN® FETs and ICs are 5 to 50 times better than the silicon state-of-the-art. This large jump in performance has led to several new applications that were not possible until the availability of GaN technology. But eGaN FETs, and in fact any GaN transistor from any of several manufacturers, are still several orders of magnitude away from GaN’s theoretical performance limits. There is a learning curve ahead that only widens the performance gap between GaN and silicon, and continues to enable new applications and transform entire end markets.
Is it easy to use?
GaN transistors (specifically eGaN FETs) are very similar in behavior to the aging power MOSFETs, and therefore power systems engineers can use their design experience with minimal additional training. To assist design engineers up the learning curve, EPC has established itself as the leader in educating the industry about gallium nitride devices and their applications. EPC published the industry’s first GaN transistor textbook (in English and Chinese) – GaN Transistors for Efficient Power Conversion. The second edition was published in 2015 by J. Wiley and is available through Amazon as well as textbook retailers. More recently, we have published two application-focused handbooks to further assist power designers of DC-DC conversion and wireless power transfer systems in the use of GaN. EPC is working with more than 60 universities around the world in order to lay the groundwork for the next generation of highly skilled power system designers trained in getting the most out of GaN technology
GaN transistors and integrated circuits from EPC are produced using processes similar to silicon power MOSFETs, have many fewer processing steps than MOSFETs, and more devices are produced per manufacturing run because GaN devices are much smaller than their silicon counterparts. In addition, lower voltage (<500 V) GaN transistors do not require the costly packaging needed to protect their silicon predecessors. This packaging advantage alone can cut the cost of manufacture in half and, combined with high manufacturing yields and small device size, has resulted in the cost of a GaN transistor from EPC to be lower in cost than a comparable (but lower performance) silicon power MOSFET. Today the designer does not even need to take advantage of the higher performance of GaN to realize cost savings in their system!
To date, several manufacturers of GaN transistors have reported excellent results from in-house stress testing. EPC continues to publish results of reliability reports and its 9th reliability report shows a combined total of over 8 million device hours under stress with no failures.
Field failure breakdown by root cause category.
Reliability report #7 (published in March 2016) reported results from tracking parts in the field for 17 billion hours over a six-year period. GaN FETs, aided by the fact that they are chipscale, and therefore do not suffer from failure modes common to packaged semiconductors, achieved a remarkable 0.24 failures for every billion device hours. There is no doubt that eGaN FETs are suitable for any application in which MOSFETs are used.
The four requisite attributes for GaN to displace the silicon MOSFET have been achieved. Switching speed, small size, competitive cost, and high reliability give the GaN transistor the “winning edge” to displace the silicon MOSFET in power conversion applications. Similar analysis shows that soon the same will be true for power ICs and analog integrated circuits created with GaN technology. Perhaps in 3-5 years the same will be true for digital integrated circuits. GaN is a relatively new technology and has just begun its journey up the learning curve…and this is “Why GaN?”