Kristin Sampayan, CEO of Opcondys, Inc., is guiding the commercialization of low-loss, high-speed, high-voltage switching technology developed at Lawrence Livermore National Laboratory.
Tech Briefs: What made you think that light activated switches could be used for high voltage?
Kristin Sampayan: The technology comes out of Lawrence Livermore National Laboratory. They were looking for a fast way to switch high voltage for a particular application, so they investigated using wide bandgap materials as a photoconductive switch. Photoconductive switching has been used for decades. A lot of the materials that had been previously used, like silicon and gallium arsenide, become conductive when they're hit with intense light, but continue to conduct even when the light goes away until all the current that needs to flow through them is gone.
But what Lab researchers found with certain wide bandgap materials, is that once they were hit with the light, they became conductive, but then when the light went away, the conductivity dropped. And not only that, the conductivity was proportional to the light intensity. They realized that this phenomenon could be used in a device to replace MOSFET transistors. Instead of the voltage at the gate, it was the light intensity in the bulk of the material that controlled the device. The Lab patented the invention, calling it the Optical Transconductance Varistor (OTV). However, they don't commercialize technology. So, through a process, we were able to license the technology and then formed Opcondys to bring this technology to market for high voltage switching.
Tech Briefs: What’s the advantage of controlling power, rather than just switching it on and off?
Sampayan: You often need to regulate the voltage during peak demand and off-peak times to minimize voltage fluctuations to the customer. Also, the time that it takes to switch on and off can be controlled, depending on whether you want to go full power. In addition, you can control the current through the device for some function such as fault limiting.
Tech Briefs: Will this work on both AC and DC?
Sampayan: It will work on either DC or AC and can convert between the two. It essentially works like a light switch and dimmer — It's either conducting or it's not — but also, the conductivity can be varied. Since it’s a bulk conduction switch, when the light hits this material, all of it begins conducting at once. It doesn't have junctions of dissimilar materials like MOSFETs and IGBTs. So as a bulk conduction device, it's truly bidirectional — it doesn't care about polarity.
Tech Briefs: Could you give me a sense of how you envision this operating on the grid?
Sampayan: It would go into, say, inverters for grid-tied energy storage for renewable energy. There are applications, for example wind turbines, where the AC that's generated by the turbine, needs to be converted to DC and then back to AC to get on the grid. It’s anytime where you need to change one type of power, whether AC to DC or one DC level to another DC level.
Tech Briefs: OK, so you're not talking about using it to switch actual grid power like to disconnect or reconnect a branch.
Sampayan: No, probably not there, at least at this point, but there are potential applications in breakers, reclosers — devices like that.
Tech Briefs: What kind of practical operating voltages are you talking about?
Sampayan: Well, right now we are developing a 20-kilovolt switch. We’re currently working on a project for ARPA-E with a goal of producing a 10-amp, 20-kilovolt device that can run up to 100 kHz; that’s about 7 times faster than comparable devices on the market. But because this is a bulk phenomenon, we could go higher or lower in voltage — it’s all based on the size of the switch material.
Tech Briefs: I see. So, when you say 20 KV you mean it can actually switch 20KV?
Sampayan: Yes, we have switched and controlled 20kV with a single device, but also multiple units can be stacked because of the light isolation. This would enable control at voltages talked about for the transcontinental “super grid.”
Tech Briefs: How could this reduce CO2 emissions?
Sampayan: Losses are based on how fast a device can switch voltage. With all the new controls that have to be used on the grid as we add more renewable resources, we have to make sure that everything is coordinated — that the DC produced by solar gets translated to AC with lower loss. In inverters and other equipment, there are a lot of semiconductor devices like IGBTs and MOSFETs. The problem with them is that they need to be thicker to hold off the high voltages that are used on the grid. That also means that they switch more slowly because of their drift velocity — the time it takes for carriers to cross the drift region and have the device become conducting. Since that tends to be fairly long, during the transition period in which you're changing from very high resistance to low resistance and going from zero current to higher current, you have I2R transition losses, which are all based on speed.
Because our Opticondistor (OTV) is a bulk conduction device and operated by light, it can turn on much faster, and so because of that shorter transition period, the transition losses are a lot less. We'd therefore be saving energy that is now being wasted. Losses in the transistor devices that are being used now can amount to 10% of the electricity that flows through them
And as we use more renewable sources, particularly wind and solar, more and more of that energy, as it's produced and distributed, is going through power electronics used in inverters and other types of control and regulating equipment.
Tech Briefs: Would you then have to operate this with a very fast turn-on control pulse?
Sampayan: Yes, it’s basically how fast we can switch our light sources, which use low voltages. We're using laser diodes, which can be turned on in a matter of nanoseconds, compared to about half a microsecond or so with a high voltage MOSFET. And because the laser diodes operate at a lower voltage, they can be turned on very quickly with common electronics.
Tech Briefs: Are there any leakage current losses when the device is off?
Sampayan: That's the thing about these wide bandgap materials, and in particular, the silicon carbide we’re using right now. They're very good insulators when they're not being hit by light, so in their natural state, there's extremely low leakage current
Tech Briefs: Is it difficult to grow the silicon carbide crystals? Is that a manufacturing issue? If you want to make these in any kind of quantity, will you be able to do it at a reasonable cost?
Sampayan: Yes, but it's very much a growing market right now that has been rapidly advancing. It's being particularly driven by a demand in electric vehicles and so manufacturing of silicon carbide has exploded over the last few years. Producers have increased both the quality and the quantity of silicon carbide, although, most of what they're producing is for IGBTs and MOSFETs. But as we develop our markets, it wouldn't be difficult for them to grow the large quantities of the silicon carbide we need. We have looked to companies that are experts in growing silicon carbide to provide the material for us.
Tech Briefs: Where are you at now?
Sampayan: We’re still developing a prototype — we're testing our first actual module. We've done bench scale tests, so now we're looking at a device that we could take out to customers.
Tech Briefs: Who do you see as your first customers?
Sampayan: Our first customers would probably not be manufacturers of equipment for the power grid. They tend to need everything to be well tested, well wrung out, before they will adopt it into their equipment. So, we're mostly looking at markets that are a little more willing to try out something new because they need a high voltage switch that can operate at much faster speeds than a MOSFET or IGBT. So, what we're really looking at as our first customers are applications for pulsed high voltage sterilization of air and liquids such as juices and milk, and also for charged particle accelerators — either for fast kickers that kick the beam into a particular beamline or doing switching for linear induction accelerators. That's where customers have indicated they're a little more open to adopting a technology like this right away.
An edited version of this interview appeared in the August 2021 issue of Tech Briefs.