An Interview with Gene Berdichevsky and Gleb Yushin, Co-Founders of Sila Nanotechnologies

Late last year, Sutter Hill Research sat down with Gleb Yushin and Gene Berdichevsky, two of the founders of our portfolio company Sila Nanotechnologies, to learn about how they created the most exciting innovation in battery chemistry in decades. We found out how they came up with the idea, why it took them so long, and what’s next for Sila.

SHV

How did you meet? How did this journey take shape?

Gleb

I joined Georgia Tech in 2007, and I focused my studies on finding a fundamentally new battery chemistry that could push performance beyond traditional lithium-ion. In a traditional lithium-ion battery, the anode is made of graphite (which is a form of carbon). In theory, a battery with a silicon anode could be made lighter and smaller while storing the same amount of energy. The problem is that silicon stores so much lithium that it tends to swell dramatically, so silicon anodes tend to degrade rapidly.

I realized that creative approaches to materials engineering could mitigate the rapid degradation found in silicon anodes, and so I applied for some Small Business Innovative Research (commonly abbreviated as SBIR) grants and started a company. It proved to be very difficult. I didn’t have any experience building a successful company, and I didn’t have sufficient resources to create high-quality tools and attract the top talent. So I started talking to as many people as I could to learn what it takes to commercialize a product. Since I’m a little bit of an introvert, it wasn’t easy, but I knew I had to push myself to talk to people outside the academic bubble.

Most of the people I met were trying to understand the size of the market, the chance of success, and how long it would take for the company to get acquired, but Gene was clearly much more interested in solving hard problems and building a company that would last for a century. Above all, he was empathetic. It was so much more comfortable for me to go somewhere with the person who was not talking about profit or getting rich, who was thinking about building something big, and who was clearly showing care for the team.

SHV

Is that how you remember it, Gene?

Gene

Mostly, yeah. I was an entrepreneur-in-residence at Sutter Hill Ventures (SHV), and I was looking at a lot of different battery technologies and talking to different professors. Another professor, Dan Steingart, introduced the two of us. We had a phone call on a Wednesday, and we hit it off. We were both Russian immigrants and had other things in common. And so I said, “Why don’t I come to Atlanta on Monday, and you can show me your lab, and we can talk it through?” I flew out there, spent two days, came back the next Wednesday, went into Mike Speiser’s [Managing Director at SHV] office and said, “Here’s the idea.” Mike looked at me and said, “This is the company you’re going to start.” When you know, you know. From that point, we spent a couple months sorting out diligence on everything and getting the Georgia Tech IP secured. By August, we incorporated, and in September, we were funded.

SHV

When you said you knew right away, did you know right away that it was going to take a decade to get to market?

Gene

I knew it was going to take a long time, but in my world a long time was five years. Gleb’s long time was maybe one year. We were both wrong, but by different orders of magnitude. In some ways, though, I always knew in the back of my head that if we were solving a problem that was worth solving, and we were building the best team in the world to do it, the longer it takes, the bigger the moat. You can think of it like swimming out into the ocean until you hit land. As long as you know the direction is exactly correct and you have the best team to move as quickly as possible, then whenever you get there, the harder it is for anyone else to get there. We made some mistakes, and someone following us could go a little bit faster, but probably not by much.

SHV

How did you get involved in battery technology in the first place?

Gleb

I got my bachelor’s and master’s in Physics. However, I soon realized that all the biggest discoveries in physics had been made at least thirty years before I was born. So for my Ph.D., I switched to material science. It’s an overlap of mechanical engineering, chemical engineering, physics, and chemistry, all together. It’s a relatively new field, only fifty to sixty years old, and so I thought it might be very interesting. I worked on electronic devices, materials for electronic-device applications, and before that I worked on photonic crystals and optoelectronics. After I finished my PhD, I found a postdoctoral position at Drexel University to study different types of carbon nanomaterials for various applications in Professor Yury Gogotsi’s lab. I was promoted less than a year later to research professor in the same lab. We worked on very diverse topics, from blood purification to supercapacitors to gas storage, including hydrogen storage for fuel cells, and so forth. I learned a lot about how challenging it is to store and transport hydrogen, and all the safety issues associated with it, and all the infrastructure that has to be built. And I thought, “Oh my God, people should work on batteries! Why are people not working on batteries?” The initial answer was that batteries are a mature technology, there is nowhere to innovate in batteries, and commercial lithium-ion batteries are just too expensive for transportation.

This was in 2005. And so I thought, okay, there must be some innovations that people can do in batteries. So I looked more broadly into the classes of materials that have very high theoretical potential — that are potentially broadly available, are low cost, and have high energy density or high specific energy. There are what are called conversion-type electrodes on both the anode and cathode sides, but they are very unstable, they degrade very quickly — not only silicon for the anodes but also various types of sulfides or fluorides for the cathodes. These were all known in the field, and they were known not to work. To me, starting with the anode made more sense, since the anodes are thicker in lithium-ion batteries. And I knew more about the anode chemistries. I didn’t work on lithium-ion batteries before I joined Georgia Tech. But my proposal to Georgia Tech was to figure out how to make a battery with a silicon anode.

Gene

Yeah, and I saw the same story play out from the industry side. I started my career at Tesla. One of the responsibilities I had when working on the Roadster battery was to measure and test all the cells in the market over the four years from when we started until we launched. And so I got to see a couple of things. I saw that the performance improvements that had been present for the prior fifteen to twenty years were starting to plateau. Then I saw that the cost declines were starting to level out as well. Basically, in terms of the price–performance curve, lithium-ion was stalling. It was the same across different vendors — everyone basically had the exact same thing and no one was really innovating. For me, the issue was, if battery performance is going to stall out, it’s really going to limit EV adoption. What are some of the battery technologies that you could use to push electric vehicles forward? I was seeing the industry limitations, and it sounds like Gleb saw that the academic pasture was pretty picked over in the old technology. So he came at it from: Where are the new ideas and the new science to be done?

SHV

Can you speak a bit more about the technical problem Sila is trying to solve? I know it involves the swelling and degradation of silicon anodes, and it sounds like you’ve engineered a material that addresses that problem. Can you talk about your approach a bit? Why is a silicon anode better?

Gleb

What we’re aiming to do is greatly accelerate the adoption rate of electric vehicles and renewable-energy technologies, and we want to do that by significantly improving performance and reducing the costs of lithium-ion batteries. At the atomic-chemistry level, a single silicon atom can store over twenty times more lithium ions than a carbon atom does in conventional graphite. This means that silicon-based anodes can be made lighter and thinner than conventional graphite anodes. And so, a lithium-ion battery with silicon anodes could be lighter and smaller while storing the same amount of energy. (Alternatively, it could also store more energy in the same cell size.) From an industrial perspective, higher energy density means you need fewer lithium-ion battery cells for an electric vehicle with the same driving range. This saves substantial costs on manufacturing and requires less of all other materials that are used in the lithium-ion cells (foils, separator, electrolyte, cathode, housing, etc.). Also, thinner silicon anodes enable much faster charging — large EV batteries with our anodes can be charged in 15 minutes, and I believe that 5–10 minute charging could be achieved in the near future, when the electrical charging stations will be able to support very high charging currents.

Those are all reasons why silicon anodes are better. The problems come from the fact that silicon stores so much lithium that it expands by over 300%, which creates lots of mechanical and chemical issues that may lead to rapid degradation. Also, during charge and discharge, every single silicon atom moves. Controlling this motion is critical since if just 1 out of 10,000 silicon atoms loses its way and contributes to an undesirable side reaction during charge or discharge, the cell-level degradation would be too severe for most applications. What we’ve developed over the years is a unique, low-cost manufacturing technology to produce precisely engineered porous composite particles that accommodate silicon swelling and control the atomic motion of silicon during lithium insertion and extraction. Once that unique particle architecture is in place, the dimensions of the composite particle change very little during battery operation, and the mechanical and chemical degradations can be reduced dramatically.

SHV

Can you tell us more about that development process? Did you have a particle architecture from day one, and just needed to figure out how to achieve it? Or was there more?

Gene

First of all, there are probably a dozen different particle architectures that could work; a lot of them can be found sketched out in our patents. The concept is one thing, but then the next piece is the physical embodiment of how you want it to be. What processes and what components and inputs and reactions can you use to create the physical body? It’s really on Gleb and our scientific innovation team to say, “Oh, we could take this synthetic pathway to create that structure.” But for a lot of the synthetic pathways that a scientist could dream up, there isn’t a piece of equipment you can buy that makes it. And so, we vertically integrated the full equipment and process stack to create the reactors, to follow the pathway that the scientists sketched out. As quickly as possible, we tried to ascertain whether there was a “there” there or not. If it was a dead end, we shut it down and move on to the next pathway. But you have to be willing to develop these fundamentally new chemical-engineering pathways, which aren’t used at scale in the battery industry. We like to borrow pathways that are used at scale in different industries so that we’re not starting from total zero. Part of our secret sauce is how, over time, we found that if we stitched together a couple of these steps that are done in other industries for other purposes, we could create something that’s really unique.

Gleb

When you take processes which already exist, you can estimate the cost in the future when they scale. If you want to be in all these different cars, and eventually the majority of cars are going to be electric, there will be certain restrictions on what input materials are commercially viable and what materials are not. That’s a good thing, because otherwise it’s almost an infinite number of possible paths, right? And so we selected chemicals that would allow us to create what we want to create, but that also have been proven to work well in very large factories and have a low cost structure at scale. If you want to make an impact on the industry, you have to develop particles that would be compatible with existing lithium-ion battery-manufacturing facilities and could serve as a drop-in replacement. We learned that we have to develop methodologies that will allow us to scale very rapidly, preferably using very low-cost manufacturing tools.

Gene

The other thing I would just say is, there are a lot of dead ends along the way. Structures and materials and synthesis pathways look promising initially, and you keep working on them, and you figure out, look, this isn’t really going to work. In the industry now, we see people going down those pathways, and we have the years of experience and the scars to prove to ourselves that it didn’t work and won’t work. We spent ten years and 55,000 iterations of synthesis to perfect and scale our technology. That’s where we married what Silicon Valley is really good at, which is building tools and systems and rapidly iterating, with what academia is really good at, which is coming up with really radical ideas. If you can marry those two things, you’ve got something that’s magic.